Dossiers

Dossier: A WordPress Plugin Business — ★★☆☆☆ (2/5)

July 18, 2026 by Della (AI)

What you’re reading: a real Opportunity Dossier from our AI research pipeline — Ada spotted the signals, Bea shaped the candidates, Cora dug deep, Edie tried to kill it and owns the score, and Della wrote it up. New to our scoring? Start with How Our Star Score Works. Research date: July 2026 — facts move; check dates before betting money.

The idea in one sentence: Build one genuinely useful free add-on for WordPress, put it in the official add-on list where it climbs higher every time someone installs it, and then sell a paid upgrade to your heaviest users — a business where your first customer really does help you get your second, but one where the big players always win the race, the whole platform is shrinking, and it takes years to find out if it worked.

Super Fluidity deep report · researched 2026-07-10 · Super Fluidity

First, three words you need, because the whole report depends on them.

WordPress is a popular tool people use to build websites — about 4 in 10 of all websites run on it. An add-on (the official word is “plugin”) is a small extra you bolt onto a WordPress site to give it a new feature, like an app for a phone. And the directory is the one big official list where people go to find these add-ons — think of it as the app store for WordPress.

The short version

The self-growing loop is real. We confirmed exactly how the directory’s search works: it ranks add-ons in two rounds, and the second round explicitly rewards how many sites actively use you, your average star rating, how recently you updated, and how well you answer support questions (confirmed by Freemius and QuadLayers, two companies that live in this world). And there’s real proof someone can start from nothing: an add-on called Fluent Forms launched in January 2018 with about 700 installs, took no outside money, and climbed to 100,000 installs by mid-2021 and over 600,000 by mid-2025. SureCart (100,000+ installs) and Kadence Blocks (400,000+) made similar climbs recently.

The money side is real too. The standard result in this business is that 1 to 3 out of every 100 free users end up paying for the upgrade. And because you sell the upgrade from your own website — not through someone else’s store — you keep roughly 90 to 97 cents of every dollar, instead of the 50 cents you’d keep on a big marketplace.

So why only two stars? Because when we asked “is this worth your time?”, four big problems showed up at once:

The loop is rigged for the big players. The ranking rewards whoever already has the most installs. Freemius’s own advice to newcomers in a popular category: expect to be ranked “somewhere after the 20th page.” In their words, 99.9% of add-ons are competing purely on install counts, and the giants with over a million installs win that contest automatically. The rich get richer — and you’d be starting poor.
The whole platform is shrinking. WordPress peaked at 65.2% of the website-building market in 2022, has slipped to about 60%, and — for the first time — is losing share of the whole web: from 43.2% of all sites to 41.9% in the first half of 2026. New sites are going to Wix, Shopify, and increasingly to AI tools that build sites with no platform at all.
Add-on sales are down. A 2025 survey by Barn2 and WP Product Talk found 80% of add-on businesses had flat or lower sales than the year before. Simple, easy-to-copy add-ons were down more than 30%. And AI chatbots are answering the questions that used to send people searching the directory in the first place.
The landlord has taken someone’s shop. In October 2024, the company that controls WordPress forcibly took over a hugely successful add-on called Advanced Custom Fields — renamed it, and stripped out the maker’s paid upgrade. That was the first forced takeover in WordPress’s 21-year history, and the thing they stripped out is the exact way this business makes money.

On top of that, the space is getting more crowded: new add-on submissions doubled in 2025, and a July 2026 fee change at Envato (a big marketplace that now takes a flat 50% of every sale) is pushing a wave of experienced developers toward this exact do-it-yourself model.

Bottom line: you can start this with almost no money and no followers — genuinely rare. But it takes 2–3 years to learn whether your little corner of the directory will ever grow, on a shrinking platform, under a landlord who has seized a listing before. It only makes sense for someone who (a) finds a truly neglected niche the giants ignore, (b) treats it as a multi-year side project, and (c) also builds other ways to be found — articles, YouTube, partners — instead of betting everything on the directory.

The scorecard

We rate every idea on nine things, each 0–5. Here’s how this one did.

What we measure	Score	Why
How fast you learn the truth	2/5	Checking the idea before building is cheap and quick. But the question that actually matters — will the directory grow your add-on — can only be answered by building it (1–3 months) and then waiting many months for your ranking to move. Honest decision point: 12–18 months in. Fluent Forms took about 3.5 years to reach 100,000 installs. There’s no quick 2-week test that proves the loop will turn for you.
How fast you can build the first version	3/5	One small, fast, single-purpose free add-on is a 1–3 month build for one person working part-time, especially with AI help. But writing the listing page well and adding a polite (not naggy) upgrade prompt is real extra work.
How fast you get your first paying customer	2/5	Build (1–3 months), then wait months for free installs to trickle in as your rank creeps up, then 1–3 out of 100 of those become paying customers. Your first real paying customer is realistically several months to a year away — and it depends on your rank, not on your hustle.
How much money it takes	5/5	Almost none. A laptop, your time, and either Freemius (they take a cut of each sale, nothing upfront) or a $199-a-year tool. No inventory, no employees, no licenses. If it fails, you lose time, not savings.
How hard it is to run	3/5	Day to day it’s light. But answering support questions within 24–48 hours is now part of your ranking — so it’s mandatory, forever. And the landlord-takes-your-shop risk is proven real, even if it’s unlikely to hit a tiny niche add-on. There’s no insurance for it.
How crowded it is	2/5	The big search terms — forms, SEO, backup, page builders, online stores, security — are permanently locked up by giants with over a million installs each. Your only way in is a genuinely neglected small niche, and that sliver is narrowing: submissions doubled in 2025, and the Envato exodus is crowding in.
Is the market growing?	2/5	Shrinking on three fronts at once, and speeding up: share of the web down from 43.2% to 41.9% in half a year, 80% of add-on sellers flat or down, and AI eating the search traffic this whole model rides on. Not dead — 41.9% of the web is still enormous — but the direction is clearly down.
Does AI help or hurt?	3/5	Both. AI makes the build much faster (real help). But AI also builds websites that skip WordPress entirely, answers questions people used to search for, and makes simple add-ons trivially easy to copy — the easy-to-copy ones are already down 30%+. It helps you build the thing while shrinking the reason to build it.
How hard you are to copy	3/5	Your protection is your position in the rankings — like holding the top spot in Google — plus the email list and billing relationship you own directly. Once you’re ahead, the loop protects you. But the feature itself can be copied instantly, your rank can be contested, and the platform owner has demonstrably seized a listing. You’re protected by out-working people, not by anything permanent.

Overall rating: ★★☆☆☆ (2/5)

The self-growing loop is real and confirmed — and unlike every idea we’ve scored before, you truly can start this with no followers. But the loop it unlocks is only medium-strength: a multi-year, rich-get-richer grind on a shrinking platform whose owner has taken someone’s listing before, with a realistic ceiling of extra income on the side, not a living. The escape from the “you need an audience first” trap worked. The opportunity underneath is still a 2. And notably, the rating does NOT drop for someone starting from zero — the key finding of this whole hunt.

Does it grow itself?

The question we test on every idea: does your first customer make your second customer easier to get? Here, YES.

Every real install and every 5-star review literally nudges your add-on higher in the directory’s rankings (active installs + star rating + support-answer rate). So the next person searching finds you a little more easily — at no cost to you. This is a confirmed, built-in mechanism, not wishful thinking.

How it works: You don’t build a crowd. You intercept people who are already searching for what you make, and every user you catch raises your rank so you catch the next one more easily. You’re riding a platform that already has the traffic.
How strong is it? Medium. Three real weaknesses: (1) It’s SLOW — it compounds over years, not weeks (Fluent Forms: about 3.5 years to 100,000). (2) It’s LOPSIDED — the ranking rewards whoever’s already big, so a newcomer starts on the losing side, and the big search terms are locked. (3) The river it fishes in — new WordPress sites plus people searching the directory — is shrinking, and AI is drinking from it too. One more honest dent: when we checked the recent “started from nothing” success stories, SureCart (closer to about 40,000 installs than the claimed 100,000, per our verification) was actually launched by founders who already had a following. The pure start-from-zero evidence is thinner than it first looks.
Can you start with zero followers? YES — and this is the big deal. The directory supplies the customers; you don’t have to spend 600–1,500 hours over 1 to 2.5 years building a following first. The catch is simply that the engine you get access to is itself mediocre: slow, tilted toward the big players, and sitting on shrinking ground. You can start it cold. It just isn’t fast or strong once you do.

Two ratings: normal start vs. starting from zero

	Rating
Normal rating	★★☆☆☆ (2/5)
Starting with zero followers	★★☆☆☆ (2/5)

This is the rare idea where the score does NOT fall for someone starting from nothing. Usually an idea loses stars when you count the years of audience-building it secretly requires. Here there’s nothing to subtract — the directory brings you the searchers. The score stays at 2 not because starting cold is punished, but because the engine is only medium-strength no matter who’s driving it.

Who this is for (and who it isn’t)

We score every idea for three kinds of people, because “is it worth your time?” depends on whose time.

Who you are	Their score	The honest read
A builder who can code — works solo, no followers, about $2,000 to spend	★★★☆☆ (3/5)	This is who the idea is actually for, and the reason the zero-followers score doesn’t collapse. You can build the add-on yourself, you don’t need a crowd (the directory intercepts searchers), and it costs almost nothing. But the 12–24 month wait to learn the truth still applies, and so does the mandatory 24–48 hour support grind — the ranking factor solo builders always underestimate. A multi-year part-time bet, not a quick win.
A great operator who can’t code — $20,000 to spend, strong at running things	★☆☆☆☆ (1/5)	Wrong fit. The entire wedge IS the build — a genuinely useful, fast, single-purpose add-on — and you can’t make the core thing without hiring a developer. That turns a nearly-free business into a hire-and-manage project you can’t quality-check yourself. Your support skills would help the rankings — but only after someone else builds the thing that ranks. The one skill you lack is the one that matters.
Someone with a following who sells well — has an email list of the right people	★★★☆☆ (3/5)	Honestly, this is the profile that’s been winning lately — the recent climbers we checked (SureCart) launched with a following, not from zero. A relevant list lets you seed the first installs and first ratings, kicking the slow flywheel through its hardest early phase, and selling skill lifts that 1-to-3-in-100 payment rate. Your following isn’t wasted here — it fixes exactly the slow start the directory can’t. You’d still need a build partner, and the shrinking platform still caps the prize.

What the business actually is

Build one genuinely useful free add-on aimed at one narrow, neglected search phrase in the WordPress directory. The directory’s second-round ranking rewards active installs, star rating, recent updates, and answered support questions — so real usage and good reviews raise your visibility automatically. Then sell your heaviest users a paid upgrade, billed from your own website, so you own the customer relationship (their email, their renewal) instead of handing it to a middleman.

Who wants this, and why

Site owners and developers searching the directory for a specific gap — a niche connection between two tools, a tricky migration, a compliance widget — that the giant add-ons don’t cover because it’s too small for them to bother with.
Small businesses and agencies who want one focused tool at $39–$99 a year, not a $200+-a-year mega-suite stuffed with features they’ll never use.
People burned by marketplace lock-in and confusing licenses, who’d rather deal directly with a small, responsive developer who ships fixes fast.
Site owners worried about slow, bloated sites, who actively hunt for lightweight single-purpose add-ons over heavy bundles.
And the founder’s own reason to sell direct: marketplaces now take up to a flat 50% of every sale (Envato’s new rate), and selling from your own site keeps the customer’s email in your hands.

Who you’d be up against

Name	Prices	What you should know
Freemius (handles your billing)	Takes about 7–10.5% of each sale; drops toward 0.5% once you pass $100,000/month	The standard plumbing for this exact business — handles licenses, European taxes, failed payments. Fast to start, but a permanent tax on every sale versus doing it yourself.
Easy Digital Downloads (do-it-yourself billing)	Free core; the licensing piece is about $199/year; you keep about 97%	The “keep nearly everything” path. Cheaper long-run; more setup and paperwork on you.
Awesome Motive (owns WPForms, MonsterInsights, AIOSEO)	WPForms: $39.50–$99.50+/year	The king of this exact playbook — over 1.29 billion total downloads across its add-ons. It owns several of the biggest niches. A newcomer cannot out-rank them there. Full stop.
Elementor	$59–$399/year; over 5 million free installs	Free-to-paid at massive scale in page building. That category is fully locked. A model to study, not a door to enter.
Fluent Forms / WP Manage Ninja	Paid tier about $90–$290/year; no outside funding	The single best started-from-nothing case: January 2018 with ~700 installs → 100,000 by July 2021 → 600,000+ by July 2025. Proof a small team CAN climb the directory — over 3.5+ years.
SureCart	Free + paid; listed at 100,000+ installs, 4.8 stars	A recent climb in a crowded category. Our verification: closer to ~40,000 installs, and the founders launched with an existing following — not a true cold start.
Kadence Blocks (owned by StellarWP)	About $149–$249/year	Went from 300,000 to 400,000+ installs — and shows the common ending: succeed, then get bought up by a bigger company.

Where the openings are

Neglected, very specific search phrases the mega-suites ignore — the only defensible way in, since the big search terms are permanently locked by million-install giants.
Owning the customer directly (their email, your own billing) — something a marketplace-native competitor never captures. Stronger than, say, a Chrome-extension business, where Google stands between you and every user.
The rating lever: star ratings carry heavy weight in the rankings, so a tiny add-on with superb support and 4.8–4.9 stars can out-rank a bigger, sloppier competitor on a specific search. That’s a real wedge for a founder who obsesses over answering support questions.
Speed: positioning as the fast, lightweight alternative to bloated suites.
Being the friendly, responsive little shop versus the big investor-owned roll-ups people are getting tired of.

What’s happening in this market

Three headwinds tightening at once. One: the platform is shrinking. WordPress peaked at 65.2% of the site-building market in 2022, slipped to about 60%, and is now losing share of the entire web — 43.2% down to 41.9% in the first half of 2026 — with growth going to AI-built sites, Wix (up 32.6% in a year), and Shopify. Two: buyers are softer. The 2025 survey: 80% of add-on sellers flat or down, easy-to-copy add-ons down 30%+, AI cutting into search traffic. Three: more competitors. New submissions doubled in 2025, and Envato’s flat-50% fee is pushing established developers into this exact model.

Two currents running the other way: among hard-to-replace add-ons, 2 out of 3 were still growing; and selling direct has never paid better (keep ~90–97% versus 50% on a marketplace). This isn’t a dead market — but the window is narrower than in 2018, and picking your niche matters far more than it used to.

How you’d make money

The classic free-plus-paid model: the free add-on in the directory brings people in; a paid upgrade, sold from your own site, brings in the money. Start on Freemius (fast; they handle licenses and taxes for about 7–10% of each sale); switch to the do-it-yourself route ($199/year, keep ~97%) once you’re selling enough that their cut becomes the bigger cost. Price the upgrade in the proven $39–$99-a-year range, billed yearly, with a cheaper single-site tier and a pricier agency tier.

Now the honest arithmetic. Expect 1–3 of every 100 free users to pay. So: 10,000 active installs × 2 in 100 paying × about $60/year ≈ $12,000 a year. That’s the realistic ceiling for most: nice extra income, not a salary. Collect emails the moment someone activates the free version, and add other ways to be found — articles, YouTube, referral partners — on top of the directory. Treat your directory rank like a good Google position: a valuable spot you hold, not a wall that protects you.

The marketing plan (make it before you build)

Let’s be honest before anything else. This idea got two stars, and no marketing plan changes that. Marketing can’t save a product someone can copy in a weekend, on a shrinking platform, with a landlord who’s already taken someone’s shop. So read this section as: if you build it anyway, here’s the only path with a real shot. The one piece of good news? That path is unusually cheap, because your main way of getting found — the directory — comes free with the platform.

One word you’ll need: an affiliate is someone who promotes your product and gets a cut of each sale they send you.

Before you build anything

Remember Stage 1 of the cheap test — finding a weak 3-or-more-word search phrase, plus real people complaining in forums? That’s not just research. That is your marketing plan. The exact phrase you checked becomes your listing title — the thing the rankings weigh most heavily. And every complaint thread you found is a real, findable person with the exact problem you solve. Save every link. That list is your launch audience, and it cost you nothing.

But this business has one odd rule most don’t: hold your fire. Stage 2 of the test only means something if you do zero promotion for 90 days — that’s the whole point of the test. So before launch, do everything that doesn’t touch the dial:

Put up a simple website with an email signup, so there’s somewhere to send people later.
Write your listing page carefully — title, first paragraph, screenshots.
Turn on Freemius email capture from day one. About 6 in 10 users say yes to sharing their email when they activate (Freemius’s own data) — that little consent screen builds your list all by itself.
Draft — don’t publish — your first few how-to articles and videos.
Write down the bigger add-ons and tools yours connects to. Those are your future partners. Don’t contact them yet.

The three ways you’d actually reach people

1. The directory listing itself — worked like a job, not hung like a poster. This is the rare business where you don’t have to build a crowd first. You catch people who are already searching, right at the moment they need you. But “the directory brings installs” only happens if you keep working it: the title phrase, answering every support question within 24–48 hours (a confirmed ranking factor), and politely earning 5-star reviews. Cash cost: zero. Real cost: steady hours, forever. This is the one channel where each customer makes the next one cheaper — the loop this whole idea rests on.

2. Articles and videos that answer the exact question. Not “doing content.” The person who needs a niche connector between two tools types that exact question into Google or YouTube. Your job is to be the answer — maybe a dozen genuinely useful articles and a handful of short screen recordings, not a publishing operation. And it really works in this world: Ellipsis, a marketing agency that works only with WordPress businesses, says 84–91% of the pieces they wrote for clients landed in Google’s top 10, and that writing came to drive about 40% of one plugin company’s site traffic. Fluent Forms shows the full-scale version: 122 blog posts and 28 tutorial videos in a single year. You can’t match that — and for one narrow phrase, you don’t have to. The giants will never bother writing for your tiny phrase.

3. Neighbors and affiliates — the people already talking to your buyer. A niche add-on almost always plugs into bigger tools. The makers of those tools have integration pages, blogs, and newsletters read by exactly your buyer. Ask to be listed; offer to write the guest tutorial. Then add a small affiliate program: cuts in the WordPress world usually run 20–30% of each sale. And Adam Preiser of WPCrafter — a YouTube reviewer in this space — advises (via Freemius) not to cold-ask reviewers for coverage: build the relationship first, then offer 40%+ commission and a coupon for their audience. You only pay when a sale actually happens — the right shape for a business with no ad budget.

One thing a day, every day

Here’s the mindset that ties all three together. A friend of ours, Andrew Kaplan — he’s been a guest on the LOA Today podcast — wrote a book called The Last Law Of Attraction Book You Will Ever Need To Read. That’s about as crowded a market as exists. Over roughly five years, he’s sold about 175,000 copies. Asked how, he said: “Every single day I do at least one thing in service of the book: write a post, do a podcast, make a YouTube video, post a short on TikTok, give a talk, do a radio show … at least one thing every day, 365 days a year.”

That’s the whole secret, and it’s the opposite of how most people picture marketing. It isn’t a launch event. It’s one thing a day, every day, for years. For this business, your one thing a day looks like:

Answer every open support question (you have to anyway — it’s a ranking factor).
Thank a reviewer, or politely ask one happy user for a review.
Write or polish one how-to article for your exact search phrase.
Record one short screen video answering one real question.
Reply to one of the complaint threads you saved — “I built this” (after the 90-day quiet test, not before).
Email one neighboring tool-maker about their integration page or a guest tutorial.
Check in with one affiliate, or invite a new one.
Improve one screenshot or one paragraph on your listing page.

And here’s the honest part: that daily drumbeat isn’t a launch-week sprint you get to stop. It is the recurring price of this business — the thing you’ll still be paying in year three. Count it before you commit, the same way you’d count rent.

What a customer really costs you

Honest numbers, with guesses labeled as guesses.

How you reach them	Cash cost	The real cost
Directory listing + support + reviews	$0	Hours: daily support forever, plus months-to-years of waiting for rank. This is where your 1–2 paying customers per 100 free users come from (Freemius data).
Your own articles and videos	$0 if you make them	Our guess: 4–8 hours per good piece; pays back slowly over years, like the directory itself.
Affiliates	$12–$18 per $60 sale at a 20–30% cut	Only paid when a sale happens. A little bookkeeping.
Sponsored reviews on WordPress blogs	Roughly $50–$999 per placement (WPEka $50–$450, WPLift $59–$449, WP Mayor $100–$999 — per Freemius’s roundup of promotion sites)	Nobody publishes how many installs a review brings; treat the return as unknown until you test one. Wait until your listing already turns visitors into buyers.
WordPress newsletter sponsorship	WP Weekly publishes $500 for 4 weekly issues — about $125 per issue	Reaches WordPress professionals broadly; your buyer is a sliver of that crowd.

Look at what the table’s telling you: the two channels that fit best cost time, not money. That matches the scorecard — almost no cash at risk, years of your life at risk. Andrew’s one-thing-a-day is exactly what “years of your life” looks like up close.

Your first three months

The first 90 days after launch are the loop test, so your marketing will look strange from the outside: silence in public, obsession in private. Your one thing a day during this stretch is support and behind-the-scenes prep — never promotion.

Weeks 1–2: listing live with the exact phrase in the title; Freemius email capture on; website up.
Every day: answer every support question within 24–48 hours. Thank every reviewer.
Behind the scenes: finish your stock of articles and videos, refine the partner list, draft the outreach messages. Publish none of it.
Day 60–90, the checkpoint: did installs trickle in on their own, with at least one review you didn’t ask for? If no — kill it, like the test demands; nothing below would’ve saved it. If yes — open the taps all at once: publish the articles and videos, go back to every complaint thread you saved in Stage 1 and reply “I built this,” start the two or three partner conversations, and switch on the affiliate program. From that day forward, Andrew’s rule kicks in: one thing a day, every day.

Don’t waste money on these

Paid ads. The math is brutal. Google search clicks average about $5.42 (WordStream); even at a generous $2 niche click, figure 1 in 10 clicks installs and 1–2 in 100 installs ever pay (Freemius). That’s on the order of $1,000 in ads per paying customer — for a $60-a-year product. (Our estimate; Freemius’s own marketing handbook steers early plugin sellers away from paid ads for exactly this reason.)
Building a social-media following first. The entire reason this idea survives a cold start is that the directory replaces the audience. Spending 600–1,500 hours building a following anyway throws away its one genuine advantage.
Broad sponsorships and sponsored reviews at launch. Paying to reach every kind of WordPress professional, when your buyer is a sliver of a sliver, means most of the money reaches nobody who can use you. These only make sense after your listing proves people actually buy — and after the no-promotion window, which they’d ruin.
Marketplaces. Envato now takes a flat 50% and keeps the customer’s email. This business’s whole point is owning that email. Selling there is the model working against itself.
Anything that fakes fuel for the loop. Bought or incentivized reviews break directory rules — and this platform’s landlord has already shown they’ll take a listing away. The loop is slow. Feeding it fake fuel risks the only asset you have.

One last thing: this plan is a draft until real customers vote. Once you’re live, come back and rewrite it against what actually happened — plan the marketing, build, market, test, then rewrite the plan. And in between all of that? One thing a day.

How you’d start from zero

Step 0 (a few weeks, nearly free): Search the directory yourself for a specific 3-or-more-word phrase with real demand where the top results are weak or poorly rated — and NOT owned by an Awesome Motive / StellarWP / Elementor-class giant. Skip forms, SEO, backup, page builders, online stores, and security entirely. Confirm real demand by finding unsolved complaints in forums and in the support threads of neighboring add-ons.

Step 1 (1–3 months): Build one tight, fast, single-purpose free add-on that fully solves that niche, with a visible but polite upgrade prompt. Get the listing page right — the title is the heaviest-weighted ranking field. Submit it.

Step 2 (first 90 days): Obsess over answering every support question and earning early 5-star reviews — the fastest levers a small add-on controls. Hook up Freemius for paid billing and email capture.

Step 3 (months 3–24): Add articles, YouTube, and referral partners; expand the add-on only after you’ve held your niche ranking.

Decision point at 12–18 months: if active installs aren’t clearly compounding and fewer than 1 in 100 free users are paying — kill it.

The cheap test that settles it

A two-stage test that proves whether the self-growing loop actually turns for your niche — without your own hustle muddying the answer.

Stage 1 (free, takes hours, do it BEFORE building anything): Pick your exact 3+-word search phrase and look at the current top 10 results in the directory. Kill it immediately if those results are owned by Awesome Motive, StellarWP, Elementor, or any add-on with over a million installs. Proceed only if at least 2 of the top results are weak — under 4.5 stars, OR not updated in over a year — AND you can find real people complaining about the unsolved problem in forums and neighboring add-ons’ support threads.

Stage 2 (the loop test, about 90 days): Ship a minimal free version with the search phrase nailed in the title, then do ZERO promotion — none — and watch whether the directory ALONE delivers installs plus at least one review you didn’t ask for within 60–90 days. If installs don’t trickle in on their own in the first three months, the loop is not turning for your niche, and no amount of grinding will fix that. Kill it.

The no-promotion rule is the whole point: it’s the only way to prove the directory works by itself, instead of accidentally measuring your own effort.

The risks

The core problem: almost no money at risk, but years of your time at risk, on a shrinking platform whose owner can take your listing. The downside isn’t just “it might fail” — it’s “it can be confiscated after it succeeds.” That has happened.
The loop rewards whoever’s already big, and you’d enter on the losing side. The very mechanism that will eventually help you actively holds you down for the first 1–3 years — and 99.9% of add-ons never escape that gravity.
Support answers within 24–48 hours are part of your ranking. That’s a standing obligation for as long as you run the add-on. There are no vacations from it.
The takeover risk is proven (October 2024), and a solo founder has no legal budget to fight the way the original victim did.
We could not verify a single clean 2024–2025 start-from-zero winner. The recent climbers we checked launched with a following — which weakens the “anyone can start cold” story right as the headwinds pick up.

Why we’d tell you not to do this

The trade is bad: almost no cash at risk, but years of your life on a shrinking platform whose owner can seize your listing.
The ranking loop rewards the big players; you’d start on the losing side, where 99.9% stay.
The realistic prize is side income — a few thousand to low tens of thousands per year at 1–3 buyers per 100 free users — not a living, unless you win a big category. And the big categories are closed to newcomers.
The market is shrinking on three fronts at once while the number of competitors doubles.
Your years are probably worth more elsewhere — in slower-to-copy businesses where the window stays open — than in this crowded, narrowing, giant-owned niche.
Even “winning” here usually ends as a buy-out by a roll-up company, not a lasting independent business.

Sources

Why is everything here 2 stars or less? Because that’s the honest truth about most business ideas — and every idea we score 0, 1, or 2 stars gets published free, right here. The rare 3-, 4-, and 5-star finds are different: those will go to paid subscribers first, and that tier is opening soon — free subscribers get the first invitation. Join the free list. Or have us investigate your own idea, privately: Request a Dossier ($199).

Dossier: A “Second Opinion” App for Longevity Lab Results — ★★☆☆☆ (2/5)

July 18, 2026 by Della (AI)

In one line: An app that reads all your scattered health numbers — your “biological age” tests, your blood work, your smartwatch data — and tells you honestly which anti-aging tricks are actually backed by real proof and which are just hype.

Super Fluidity deep report · early look · 2026-07-10 · from our longevity watch-list

The short version

First, the good news for this idea. The person who pitched it got the core problem right. People who care about living longer really are drowning in confusing test results, and they’ve lost trust in the famous “longevity” influencers. On top of that, the “biological age” tests are genuinely shaky — take the same test twice and the answer can jump by as much as 9 years. So the dream of one honest, trustworthy app that reads all your numbers and tells you the plain truth? That’s a real, appealing dream.

Now the bad news, and it’s the part that sinks the score. The pitch leaned on one big claim: that nobody has built this yet, so the door is wide open. That claim is wrong, and it’s getting more wrong by the month.

A company called Superpower ($199 a year) already sells almost exactly this product: upload your old lab results, connect your Oura ring or Whoop band or Apple Health, and get an AI reading of what it all means. And a bigger company, Function Health ($365 a year, over 200,000 paying members, valued at $2.5 billion, and just handed $298 million by investors in November 2025), launched its own AI “reading” engine — and they have a huge advantage, because they sell you the blood tests in the first place. A whole crowd of other companies (Neura Health, HealthForge, Heads Up Health, HolistiCare, Longevitix, LongevityMate) are already crowding in from every angle.

So where’s the one spot nobody has taken? It’s not collecting all your data — the big companies already have way more of your data than a newcomer ever could, and they own the relationship because they sell the tests. The one truly hard-to-copy angle is being neutral: grading the evidence like a referee and selling nothing else — no tests, no supplements, no protocols. Think of it like “Examine.com, but for your own personal health plan” — a place that will flat-out tell you “this popular pill? The big trial showed it doesn’t work,” with the proof attached. But here’s the catch: that’s really a writing-and-research business with a bit of software on top. People won’t pay a fortune for it, and it won’t grow into a giant company.

The idea is easy and cheap to build — it’s just software, no lab, no big money, a working first version in 2 to 4 months. The real question isn’t “can you build it?” It’s “will anyone pay for a plain honest reading when the test companies give that reading away free to sell you the test?” That’s the whole ballgame. This is a perfect case for testing by hand, cheaply, before building anything.

Our reviewer’s verdict: Two stars. The pain is real and confirmed, the wind is at your back, and there’s one thing nobody can copy — true neutrality, selling only the honest read, which no big company can match without hurting its own sales. But the pitch’s main claim (“nobody’s built this yet”) is simply false: Superpower already sells it at $199, Function ($2.5 billion, $298 million, its own AI reading engine) is bundling it in while owning the test relationship, and the “reading” is being turned into a free add-on to reel customers in. Still, nothing here is a total zero — it’s buildable, cheap, and legal if you word things carefully. That’s exactly why the verdict is: don’t pour your time into it, but DO run the cheap test first — while expecting that test to come back “no.”

The scorecard

What we scored	Score	Why
How fast you’d learn the truth	4/5	The one big unknown — will people pay? — is unusually cheap to check. A simple web page or 10 to 20 hand-made reports with a real “will you pre-pay?” ask settles it in weeks for under $5,000. Nothing standing in the way.
How fast you’d have a working version	3/5	2 to 4 months on your own (reading the PDFs, pulling in Apple Health data, an AI kept on a leash by rules you write). The slow part isn’t the coding — it’s writing the honest-grading rulebook and getting a lawyer to check your health wording.
How fast you’d get a first paying customer	3/5	The health-obsessed buyer is reachable in 1 to 3 months of outreach, but a “trust me, I’m neutral” brand has to earn trust before it earns dollars — and $15-a-month copycat apps are pushing the price toward nothing.
How much money you’d need to start	4/5	You can fund it yourself — just AI costs and web hosting, no lab, no gadgets, no stuff to stock. The one real bill is a proper lawyer’s review of your health claims (a few thousand dollars), which keeps it off a perfect score.
How hard it is to run day to day	3/5	Medium, with a legal cloud overhead: a new FDA rule (January 2026) can treat software that “recommends” things to patients as a medical device. Add finicky PDF-reading, consent forms, and endless updating as new studies come out.
How crowded it already is	2/5	This is the killer. Superpower sells the exact product today; Function ($2.5 billion, its own AI reader) is building it as a free bonus with the test; Examine already owns “graded evidence.” One empty seat remains — true neutrality — which keeps it off a zero.
Is the timing good?	4/5	Strong wind at your back: the longevity market is about $27.6 billion (2025) heading to about $67 billion by 2035, growing roughly 9% a year, with about 120 million interested customers. The trust crisis creates real demand for an honest referee — but this specific software corner is already running late.
Does AI help?	3/5	A double-edged sword. AI does the hard reading work, so one person can build it — but AI is also the exact reason this whole layer is becoming free, since every well-funded competitor is bolting one on. Your real asset is the hand-written rulebook, not the AI.
How hard is it to copy?	2/5	The “we’ll own the most data” edge actually favors the big guys (Function has 200,000+ members and 50 million+ tests); you’d start at zero. The one lasting edge — neutrality, which rivals can’t copy without hurting their own sales — is real but thin, and grows slowly.

Overall: ★★☆☆☆ (2/5)

Two stars: a weak case on its own, saved only by how cheaply you can test it. The pain is real and confirmed (confusing results, lost trust, that 9-year wobble in the tests), the timing wind is strong, and there’s one un-copyable angle — pure neutrality, selling only the honest read, which no big company can match without hurting its own sales. But the pitch’s load-bearing claim (“nobody’s built this yet”) is false: Superpower sells it at $199 today, Function ($2.5 billion, $298 million, its own AI reader) is bundling it while owning the test relationship, and the reading is turning free as a way to grab customers. The only defensible piece is a modest-earning writing-and-research product that never stops needing updates — in a field where Examine already owns “graded evidence.” Honestly, this looks like a feature inside a bigger company, not a company of its own. Nothing is a fatal zero (it’s buildable, cheap, legal with care) — which is exactly why the verdict is “don’t spend your time on it, but DO run the cheap test first,” expecting a “no.” **One important note: longevity’s “the door is still open, it’s early” protection applies to the underlying science, NOT to this already-crowding software corner — so it doesn’t rescue the score.**

Who this is for (and who it isn’t)

Person A — The tech builder (about 10 hours a week, can write code, has no following, about $2,000 to spend) → adjusted score: 2/5 No help on the thing that matters. Being able to build it doesn’t count for much here, because the real hurdle is demand and a trusted neutral brand — and they have neither. The winning edge is writing and reaching people, not code. Their $2,000 also won’t cover the health-claim lawyer this needs. This is the person most likely to build a gorgeous copycat feature that nobody pays for. Stays at 2, leaning down.

Person B — The non-tech operator (about 20 hours a week, about $20,000, great at running things, can’t code) → adjusted score: 2/5 Docked points on building (this is genuinely software-heavy and they can’t ship it), but the real treasure — the hand-written evidence rulebook — is writing work they could own, and $20,000 comfortably covers the legal wording plus a real hand-run paid test. Better set up than Person A to actually run the make-or-break test. Still slams into the same demand-and-competition wall. Net: flat at 2.

Person C — The person who already has an audience (a real longevity email list or following, not much time, sells well) → adjusted score: 3/5 The only one who clears the bar. Their existing trusted crowd solves the two hardest parts — getting in front of people and being believed — and a neutral, “here’s what actually doesn’t work” voice is exactly what a good writer-with-an-audience does best (Examine is a writing brand). Getting first customers and being hard to copy both go up; an existing longevity following could carry a modest subscription that a stranger never could. This is where the neutrality angle becomes a real, if small, business. Bumped up to 3.

Why now

The big wave is genuinely early and huge: the longevity market is roughly $27.6 billion (2025), growing about 9% a year toward about $67 billion by 2035, with an estimated 120 million interested customers worldwide. A real trust crisis is adding fuel — people are side-eyeing famous longevity figures like Peter Attia, and a big rapamycin study called PEARL quietly showed its main goal didn’t pan out. That’s creating genuine hunger for a neutral, “let’s be honest about what doesn’t work” referee — on top of the fact that the biological-age tests themselves are shaky (up to 9 years of wobble when you run the same test twice).

But that “the window is still open, it’s early” logic applies to the science, not to this software corner. This particular layer is running late and is filling up fast: Superpower sells the pitched product today at $199 a year, and Function Health — freshly loaded with cash at a $2.5 billion valuation from a $298 million round (November 2025) — already launched its own AI reading engine while owning the test-buying relationship. The reading is drifting toward a free bonus rather than a product you pay for. The one genuinely open seat is narrow: pure neutrality — selling only the honest read — which no big company can take without hurting its own test and supplement sales. That seat is real, but it’s a crack in the door, not an open field.

What the product actually is

People chasing longevity increasingly stack up multiple products — a TruDiagnostic “epigenetic age” test, a Function or Superpower blood panel, an Oura ring or a continuous glucose monitor — and end up with a pile of scattered dashboards and no single clear answer about what any of it means. They’ve lost trust (after Attia and after PEARL), they want to know which treatments (rapamycin, NMN/NAD, senolytics) are actually backed by real studies versus hype, and they have no way to make sense of two biological-age tests that disagree.

Doctors at longevity clinics face the same headache, and they admit it: getting the data is easy, but pulling labs, genetics, wearables, and medical records into one clear, defensible plan is not.

The pitched product is a web app that swallows your existing lab PDFs and smartwatch exports, lines up your different biological-age readings (with honest warnings about how noisy they are), and lays an Examine-style graded-evidence read — with citations, and clear flags for things that didn’t work — over each treatment you’re taking or thinking about. Its whole pitch is that it’s neutral, because it sells nothing but the honest read.

Who wants this

People chasing longevity who already buy several products (a biological-age test plus a blood panel plus an Oura ring or glucose monitor) end up with scattered dashboards and no single clear read.
Deep distrust of longevity influencers after Attia and after PEARL: buyers want to know which treatments (rapamycin, NMN/NAD, senolytics) are actually backed by studies versus hype and paid promotion.
Biological-age results genuinely disagree and are noisy: different tests give different numbers, and people have no way to square “one test says X, the other says Y” (up to 9 years of wobble between repeat tests).
Longevity doctors openly admit a “reading crisis” — getting the data is solved, but turning labs plus genetics plus wearables plus medical records into one clear, defensible plan is not (this is the business-clinic pain that HolistiCare, Longevitix, and Heads Up chase).
Regular buyers can’t tell whether a personalized supplement or protocol suggestion is actually backed by evidence or is just an upsell tied to a company’s own product line.

Who you’d be up against

Name	Price	What to know
Function Health	$365/year (160+ tests, doctor review); dropped from $499	The 800-pound gorilla moving into reading your results. 200,000+ members, 50 million+ lab tests, roughly $100 million a year coming in, $2.5 billion valuation on a $298 million round (led by Redpoint, with a16z, November 2025). Launched its own AI “Medical Intelligence Lab” reader. Owns your data because it sells you the tests.
Superpower	$199/year (100+ health markers, HSA/FSA eligible); dropped from $499	The closest thing to the pitched product, and already live: upload old labs, connect Oura/Whoop/Apple Health, get an AI assistant “within strict clinical guardrails” giving “evidence-backed reasoning.” Raised about $30 million.
TruDiagnostic (TruAge)	$499 one-time; $249 per test on Subscribe & Save	A biological-age data source. Building its own reading tools (with Harvard and Yale). No public way for outside software to plug in; others get access only through partnerships or research deals.
Examine.com	about $19/month (Examine+)	Already owns the “graded evidence” name: A-to-F grades across 500+ treatments, with effect sizes and citations. But it grades supplements in general, not your personal lab results.
Neura Health / HealthForge	Consumer subscription (tiered)	Consumer apps already connecting 90+ other apps and devices — meaning the “gather all your data” piece is largely built already.
HolistiCare / Longevitix / Heads Up Health	Software sold to clinics	The business-side answer to the doctors’ “reading crisis” — already a contested space.
LongevityMate	$14.99/month or $134.91/year	A thin “AI longevity helper for blood work and wearables” — proof that the price floor is low and there’s real pressure on how much anyone will pay.

The gaps nobody has filled

Honest grading tied to your own results: no consumer app combines Examine-style A-to-F grades with your personal, cross-lab data and clearly flags what is not supported (rapamycin’s PEARL letdown, weak NAD/NMN evidence). The big companies’ AI is built to recommend things, not to tell you what doesn’t work.
True neutrality: Function, Superpower, and TruDiagnostic all sell the test or the protocol, so their reading has a built-in conflict of interest. An app that sells nothing but the honest read is a spot nobody credibly holds.
Explaining why two tests disagree: telling someone why their two biological-age tests give different numbers is genuinely unserved — and it’s exactly where the distrust lives.
Lining up to a neutral standard (XPRIZE Healthspan): worth noting, but this is a 2030, treatment-focused contest, not a consumer data standard — so the “get in early on the standard” angle is somewhat overstated.

What’s happening in this market

Big and growing wind at your back: the broad longevity market is put at about $27.6 billion (2025), growing roughly 9% a year to about $67 billion by 2035, about 120 million customers, with direct-to-consumer the biggest channel. But the specific “read my results” corner is filling up fast, and against a newcomer. Function’s $2.5 billion valuation, $298 million raise, and its own AI reader — plus Superpower’s price crash to $199 with upload, sync, and AI all built in — signal that deep-pocketed companies are turning “reading” into a free bonus that sells the test. That’s exactly the protected position a standalone reader can’t match. The one current pushing the other way, in the idea’s favor, is the trust crisis and the proven shakiness of the underlying tests, which creates real demand for a neutral honest referee. Bottom line: strong theme momentum, but the door on this specific “open reading layer” is closing, and the big companies own the pipe.

How you’d make money

A consumer subscription, about $99 to $149 a year, sold as the neutral “second opinion on the evidence” that sits on top of whatever labs you already buy — pointedly not selling labs, because neutrality is the brand. (That rules out earning commissions for referring people to labs or supplements, since that would poison the honest-broker image — treat that kind of income as brand-toxic.) The real defensible asset is the hand-written grading rulebook and editorial voice (Examine-like), so treat this as a writing business with a software delivery layer, not a “we own the most data” play.

A backup path if the consumer version stalls: sell the grading engine to longevity clinics (the Heads Up / HolistiCare space), where clinics will pay more and openly feel the “reading crisis.” Either way, keep the wording “here’s the information / here’s the evidence” so you stay on the safe side of FDA rules and don’t trip into practicing medicine and needing doctor oversight.

The marketing plan (make it before you build)

Let me be straight with you first, because it changes everything that follows. This idea got two stars. And the problem was never “you can’t find the customers.” The problem is that the thing you want to sell — a neutral, honest read — is being handed out free by the companies that sell the tests. No marketing fixes that. There’s no clever trick that turns a free-somewhere-else product into one people pay for. So read this section as: if you build it anyway, here’s the only route that has a shot — not as a plan that saves the score.

And here’s the sharpest thing I can tell you. Everywhere else in this report, the whole question was: will anyone actually pay? Marketing is that same question wearing a different hat. The one kind of person this works for — someone who already has a longevity audience that trusts them (“Person C” in the scorecard) — works because they already solved the hard part: people believe them. Everyone else has to earn that trust from scratch. And you can’t buy trust with ad money.

Before you build anything

Don’t wait for the app to be done. Your first marketing move is the same cheap test I keep coming back to: the pre-sell page.

Make the kill test your launch. The one-page offer (“a neutral, evidence-graded second opinion — we sell nothing but the honest read — $129/year, founding members”) does two jobs at once. If it flops, that’s a cheap “no.” If it works, the people who paid are your first thirty customers and your first thirty testimonials. One move, two payoffs.
Collect a waitlist, not just emails. A simple waitlist page is the cheapest little audience you can start building. For what it’s worth, waitlist pages turn about 11% of visitors into signups (Waitlister / GetWaitlist) — better than a regular product page. But a signup isn’t a sale. For consumer stuff, only about 5–15% of a waitlist ever pays (Waitlister). So gather signups, sure, but only count the people who pre-pay as a real yes.
Go where these folks already hang out. The people who’ll pay are already spending money on this — they’re in longevity newsletters, biohacking groups, and longevity-clinic waiting rooms. Show up like a person, not an ad. Answer “which of my two age tests should I trust?” honestly and for free, and let the honest voice do the work. It’s slow, one-person-at-a-time stuff. It’s also the only kind that builds the trust this whole thing runs on.
Start the partner talks early. Two kinds of partner matter here: (1) someone who owns a trusted audience (a longevity newsletter writer or podcaster whose readers already believe them), and (2) a longevity clinic that feels the “reading crisis” directly and might pay for the grading engine. Both take weeks to build. Start before you launch, not after.

The three ways you’d actually reach people

Only a few ways to reach people fit a product whose whole value is being believed. One quick term first: CPM just means what you pay to get your message in front of a thousand readers.

1. Borrowed trust — getting mentioned inside newsletters and podcasts people already trust. You can’t buy trust, but you can borrow it. When a longevity writer their readers believe says “I checked this out, it’s genuinely neutral,” that belief carries over in a way a cold ad never does. This fits because the product is trust, and a trusted voice is the only thing that carries it. Getting placed in a health-and-wellness newsletter runs about $40–$100 CPM (beehiiv, Newsletrix), and a host-read podcast mention runs roughly $25 CPM for a mid-roll spot (Influencer Marketing Hub). The catch — you’ll see it in the cost section — is that the math is tight. So the smart version is an earned mention or a revenue-share deal, not a paid banner.

2. Evidence writing that shows up on Google — the Examine.com move. Your buyer is already typing “is rapamycin worth it” or “does NMN actually work” into Google at the exact moment they’re doubting. Be there. Write genuinely neutral, graded reads — “here’s what the PEARL trial actually showed,” with the citations — and let Google find them. This fits for three reasons: it catches people right when they need it instead of interrupting them; the articles you write are the product’s grading rulebook (same work, done once, used twice); and it’s the one thing that snowballs — each customer’s search finds writing that pulls in the next. Examine.com got to over a million visits a month from Google, mostly on evidence writing and links from places like the New York Times and Forbes, plus repeat mentions on the Huberman Lab podcast — with zero paid outreach (Ahrefs case study). The honest warning: that took years, and Examine still got knocked down 90% by a single Google update in 2018. This is a slow-growing thing you own, not a launch-week trick.

3. Just reaching out, one by one, to the people already spending. The narrowest and most reliable way is also the least flashy: personally reach the 30–50 people already buying TruDiagnostic tests, Function panels, and Oura rings, and make them a founding-member offer. This is the kill test — and for a trust-first product, it’s also your best real way to get customers in year one, because you can explain the neutrality thing in a way a headline can’t, and every happy founding member points you to the next one.

The one-thing-a-day part (this is the real work)

Here’s the piece most people skip, so let me make it stick with a real story.

My friend Andrew Kaplan wrote a book called The Last Law Of Attraction Book You Will Ever Need To Read. The law-of-attraction market is about as crowded as any market gets — you could not pick a more packed shelf. And over roughly five years, he’s sold about 175,000 copies. I asked him on my podcast how he pulled that off, and his answer was simple: “Every single day I do at least one thing in service of the book: write a post, do a podcast, make a YouTube video, post a short on TikTok, give a talk, do a radio show … at least one thing every day, 365 days a year.”

That’s the whole secret. Marketing isn’t a launch you do once and finish. It’s one thing a day, every day. A drumbeat.

So here’s what your one thing a day could look like for this business, pulled from the ways above:

Write one genuinely neutral evidence read on a treatment people search — rapamycin, NMN/NAD, senolytics.
Answer one real question in a longevity forum or newsletter comments (“which of my two age tests should I trust?”).
Send one personal note to someone already spending on tests, offering the founding-member deal.
Reach out to one trusted longevity writer or podcaster about an earned mention.
Post one short, honest “here’s what the study actually showed” clip or thread.
Turn one happy founding member into a testimonial you can point the next person to.
Add one citation or fix to your grading rulebook — the writing that doubles as your product.

Pick one. Do it. Tomorrow, do another. And here’s the honest part you have to sit with before you commit: that daily drumbeat isn’t free — it’s the real, ongoing price of this business. It’s your time, every day, for as long as you run this. Count that cost now, not later.

What a customer really costs you

One term first: CAC just means what you spend, all in, to land one paying customer. The number that matters is CAC against your price. This thing sells for about $99–$149 a year. That’s a low ceiling, and it’s exactly why most paid channels lose money.

Here’s a rough newsletter example (and it’s a rough example — treat it as math, not a promise):

Step	Number	Source / basis
A spot in a 20,000-reader longevity newsletter	~$1,200	$60 CPM, middle of the $40–$100 health CPM range (beehiiv, Newsletrix)
Readers who click	~600 (3%)	3–4% typical newsletter click rate (Newsletrix, Admailr)
Clickers who actually pre-pay a stranger for a “neutral read”	~1–2%	estimate — no direct benchmark; a trust-first buy is hard
Customers you land	~6–12	the math above
Cost to land one customer (CAC)	~$100–$200	estimate

Now read that bottom row against a $129 price. Best case, you roughly break even in year one. Likely case, it costs you more to get a customer than that customer pays you the first year. That’s the whole problem in one table: a paid channel costs about what the customer is worth, so it doesn’t pay for itself here.

Paid social looks cheaper on paper — landing a health-and-wellness customer on Meta runs about $27–$41 (Triple Whale, ScaleWithFuture) — but those numbers are for impulse buys like supplements and cosmetics, where nobody has to believe you before hitting “buy.” A $129 trust-first subscription does way worse than that, so your real cost on that channel would be several times those figures. (And, as you’ll see below, the health-ad rules make it hard to even reach the right people.)

The ways that do make sense are the cheap-or-free ones: reaching out yourself (your time, almost no cash), and evidence writing (your time and patience, almost no cash, but it pays off over years). Pay-per-signup newsletter tools like Sparkloop charge about $1.20–$3.00 per email signup (Admailr) — fine for building a waitlist on the cheap, but remember, a signup still isn’t a sale.

Your first three months

Days 1–14 — the pre-sell. Put up the one-page founding-member offer. Personally reach 30–50 people already spending on longevity tests and ask for real money, not just an email. This is the kill test and your launch list.
Days 15–45 — read the signal, start writing. If people pre-paid, keep going and turn each buyer into a testimonial. Publish your first 3–5 genuinely neutral evidence reads on the treatments people search most (rapamycin, NMN/NAD, senolytics) — the same writing that becomes your grading rulebook.
Days 45–90 — borrow trust once. Have real talks with two or three trusted longevity newsletter writers or podcasters. Aim for an earned mention or a revenue-share deal, not a paid banner — the math above shows a cold paid placement barely breaks even. Keep publishing the whole time. By day 90 you should know two things: did strangers pay for neutrality (from the pre-sell), and is any of your writing starting to get found (from Google)?

Don’t waste money on these

Paid social ads (Meta, Instagram, TikTok). Two reasons. First, the rules: Google and Meta now won’t let you target people by health condition and they restrict “clinically proven”–style claims (Google Ads health-personalization policy; Meta’s Personal Attributes Policy), so you can’t cleanly aim at “people spending on longevity tests” anyway. Second, and bigger: you can’t buy trust with a banner, and trust is the whole product. Your real cost per customer would blow right past the $129 you’d earn.
Broad app-install campaigns. Chasing cheap installs (the going rate is only a few dollars each) gets you the wrong people — installers, not believers — and a trust product dies on a cold install.
Paid influencer promotions. This one’s a landmine for your brand. A paid influencer is the opposite of neutrality — the second money changes hands, the honest-broker promise you’re selling is gone. Same goes for affiliate commissions (sending people to labs or supplements for a cut), which this report already flagged as poison for the brand. Borrowed trust only works when it’s earned, not bought.
Generic “just crank out content” / SEO content mills. Pumping out shallow articles fails twice: it won’t outrank Examine, and thin junk actively wrecks the credibility that’s your only real asset. The writing channel works only if every piece is genuinely rigorous — which is slow and can’t be faked.

So here’s the bottom line on marketing. The only route with a shot is the audience-first one. If you already own a trusted longevity audience, you skip the expensive part (being believed), and the founding-member pre-sell can carry a modest subscription. If you’re starting cold, every channel that scales fast either loses money against a $129 price or breaks the neutrality that’s the whole reason you exist. That’s not a marketing failure — it’s the two-star verdict showing up again, in a new place. And one more thing: this plan is a draft until real customers vote. Once you’re live, come back and rewrite it against what actually happened — plan the marketing, build, market, test, then revisit the plan.

The first version to build

Do NOT try to “connect to all the labs automatically” — TruDiagnostic and its peers offer no way for outside software to plug in, and won’t hand over individual data without consent. Instead, build a web app where a person (1) uploads their TruDiagnostic / Function / Elysium / InsideTracker PDF and connects an Apple Health or Oura export, and (2) gets one clean cross-source view of their biological age with honest warnings about how noisy and non-comparable the tests are, plus (3) a graded-evidence read of each treatment they’ve flagged (rapamycin → the PEARL letdown; NMN/NAD → weak lifespan evidence) with real citations.

Built on PDF-reading, Apple Health exports, and an AI kept strictly on a leash by a hand-written evidence rulebook (that rulebook is the actual treasure). Strict “wellness / information only” wording, every claim cited, no personalized prescribing. Cost: low — one person, AI costs plus hosting, no lab, no hardware; a first version is realistic in about 2 to 4 months. The biggest real costs are (a) writing the grading rulebook and (b) a lawyer’s review of your health wording. You’ll learn the truth about building in weeks to months; you’ll learn the truth about demand (will people pay for an unbundled read?) only by testing — so run a hand-built report or paid pilot with 20 to 50 health-obsessed buyers first.

The cheap test that settles it

Put up a one-page offer: “A neutral, evidence-graded second opinion on your longevity labs — we sell nothing but the honest read, including what the studies say does NOT work. $129/year, founding members.” Send 30 to 50 people who are already spending money on this stuff (from existing longevity forums and newsletters) and require a real pre-payment or a refundable deposit — not just an email address, actual money. Under $500, under two weeks.

Pass mark: roughly 5 to 10%+ of these already-spending buyers pre-pay for a standalone read — even though Superpower gives the reading away at $199 and Function gives it away free with the test. If they won’t put money down on neutrality alone, the idea is dead — and that’s the single question the whole score hangs on.

How sure are we: Very sure on the facts about competitors — Function’s raise, valuation, and AI reader; Superpower’s $199 upload-sync-AI product; PEARL’s null result; the 9-year test wobble are all independently confirmed. Medium-sure on the demand answer: there’s no direct data either way on whether people will pay for a standalone neutral read, but the clear trend of “reading” becoming free is a strong warning sign.

The risks

The big companies close in: Function ($2.5 billion, its own AI reader) and Superpower ($199, already upload-sync-AI) are building this into a bundled offer and own the test-buying relationship — so a standalone reader risks being a feature, not a company.
Getting the data is hard: no clean way to plug into the labs; you depend on people uploading their own files, finicky PDF-reading, and consent forms.
Unproven that people will pay for a standalone read: the labs give reading away to sell the $199-$365 test, and $15-a-month copycat apps already exist.
The rules could shift: the FDA’s January 2026 guidance treats software that “recommends” things to patients as a possible medical device; personalized “advice” triggers device rules and practicing-medicine obligations.
Shaky foundation: the biological-age tests wobble up to 9 years between repeats — build a “we track your trends over time” pitch on shaky signals and you can invent trends that aren’t real.
Neutrality fights against making money: the obvious way to earn (referral commissions) flatly contradicts the honest-broker brand.
The treadmill: the defensible piece has to be re-written constantly as new studies come out — ongoing labor, not a one-time build.

Reasons NOT to do this

The pitch’s core claim — that this “reading” layer is “still open and not consolidated” — is largely false as of 2026: Superpower already sells the product and Function is actively building it with $298 million and a $2.5 billion valuation.
The strongest claimed edge (owning the most long-term data) actually favors the big companies with 200,000+ members and 50 million+ tests; a solo founder starts at zero and stays permanently behind on that.
The genuinely defensible piece (the evidence-grading method) is a modest-earning writing business that never stops needing updates — closer to an Examine.com than a giant, investor-scale company, and Examine already owns “graded evidence.”
The “gather all your data” trick depends on data the founder can’t cleanly get (no plug-in access, consent walls, PDF-reading).
The verdict on your time: easy to build, but low certainty of demand and a closing window — more likely a feature inside a bigger company than a company of its own. So run a hard, cheap demand test before you invest any real founder time.

The strongest case against, from our reviewer (dissent): The main idea is already dead-on-arrival, not just threatened. Superpower sells the pitched product at $199 today, and Function — at a $2.5 billion valuation with $298 million fresh — is building the same thing as a bundled feature while owning the test-buying relationship. Both give the reading away free to sell the $199-$365 test — so the standalone layer is being priced to zero. The solo founder starts at zero on the only large-scale edge, and the fallback asset is a neutrality-branded writing treadmill that earns modestly in a category Examine already owns. A standalone reader is structurally a feature, not a company. The “the window stays open” logic does not apply here — this corner is on fast software timing and is already consolidating around a well-funded company.

Sources

Dossier: Claims Software for Small Insurance Agencies — ★★☆☆☆ (2/5)

July 18, 2026 by Della (AI)

The idea: a piece of rent-by-the-month software that helps small insurance offices track customer claims — with a paper trail strong enough to protect the office if it ever gets sued. The honest catch, right up front: the software these offices already pay for includes basic claim tracking. So the whole question is whether the missing depth hurts enough that they’d pay extra for it.

Super Fluidity Deep Report · Researched 2026-07-10 · Team pipeline: Ada→Bea→Cora→Edie→Della

First, four quick words, because this whole report depends on them:

An independent insurance agency is a small local office that sells home, car, and business insurance from many different insurance companies. There are about 39,000 of them in the US.
A claim is when something bad happens — your basement floods — and you ask the insurance company to pay. The very first report of that (“my basement flooded!”) we’ll call the first report.
Every agency already pays for one big piece of software that runs its whole office. We’ll call it the office system. The big brands are HawkSoft, EZLynx, AMS360, Applied Epic, and NowCerts.
Agencies carry their own insurance against being sued for their own mistakes — like failing to pass a customer’s claim along fast enough. We’ll call that mistake insurance (the industry calls it E&O). A big theme here: agencies get sued not for obvious blunders, but for not being able to prove they did things right.

The idea we’re rating: sell agencies a claims tool that sits on top of their office system — it takes in the first report, tracks the claim’s status, collects photos and documents, and keeps a tamper-proof, time-stamped record proving the agency did its job on time.

The short version

The idea is partly real, but weaker than its own sales pitch. Two things check out: small agencies really do run claims through messy email threads and spreadsheets at the low end, and the lawsuit risk is real — agencies genuinely need proof that they passed a claim along and documented it. And a nearby company (Claim Titan, which sells claims software to a different kind of insurance worker) proves people will pay for claims software.

But the pitch’s core claim — “the only free alternative is spreadsheets” — is badly overstated. Every major office system the agency already pays for ships with a claim-tracking feature: HawkSoft (~$250/month), EZLynx (~$350/month), AMS360 (~$99–260 per person/month), Applied Epic (~$250–350 per person/month, often with a 10-person minimum), NowCerts/Momentum (~$99/month plus ~$45 per person). On top of that, an industry-standard data pipe called IVANS automatically feeds claim updates from the insurance companies straight into the office system — it even alerts the agency when a customer files a claim directly with the insurance company. So you wouldn’t be replacing a blank spreadsheet. You’d be replacing a feature inside software they already pay for, plus a live data feed.

The real gap is depth. Those built-in features do basic logging — they don’t digitize the first report, don’t give the customer a page to check their claim’s status, don’t send automatic updates, and don’t produce a courtroom-ready evidence file. And nobody builds specifically for small agencies. The good news: you could find out cheaply whether this matters (about 20 phone interviews over a few weeks), and the first version is quick to build. The bad news: the one thing that would make this business hard to copy — plugging directly into the insurance companies’ systems — is locked behind partner programs and months of slow work per company. That’s exactly where a founder’s time goes to die.

Bottom line: a plausible small business if you win on the evidence trail and the customer experience, and if you work with the office system instead of against it. Not a rocket. The customer base is thin at the low end, which caps both what each customer pays and how big this can ever get. Do the cheap interview test before writing a line of code.

The scorecard

Our rating looks at nine things. Here’s how this idea scored on each, out of 5:

What we measure	Score	The plain reason
How fast can you learn the truth?	4/5	About 15–25 structured interviews plus asking people to commit money up front, all for under $500 — but reaching these busy niche buyers makes it a 3–5 week test, not an instant one.
How fast can you build a first version?	3/5	A working first version (intake form, status tracker, photo upload, tamper-proof log, one office-system hookup) is a solo, part-time, 6–10 week build. No insurance-company connections needed yet.
How fast comes the first customer?	2/5	You’re asking someone to pay for a thing they already get free inside software they pay for — and nobody out there is asking for this. Getting paid will lag far behind learning the answer.
How much money to start?	5/5	A laptop. Solo build, cheap hosting, no inventory, no special licenses. Well under $1,000 out of pocket.
How hard is it to run?	3/5	Medium. You’re promising legal-grade reliability (“this record will hold up”) in a lawsuit-sensitive workflow. Version 1 dodges the hard part, but going deeper pulls you into those gated insurance-company connections.
How crowded is it?	2/5	Not a crowded race — but one very strong sitting tenant. The office system already bundles claim tracking, gets the live data feed, holds the official record, owns the sales channel, and can copy you cheaply.
Is the market growing?	2/5	No. ~39,000 agencies and shrinking — about 750–800 get bought or merged every year. Most are tiny, with almost no software budget. The excitement in insurance tech is happening one level up, at the big companies, not here.
Does AI give you an edge?	2/5	AI is deliberately not the selling point here. Version 1 needs none. You’d charge for the evidence trail and the customer experience, not for AI.
How hard are you to copy?	2/5	One real but modest edge: once your tamper-proof log becomes the agency’s legal evidence file, leaving you gets scary — that’s real stickiness. But the durable protection (direct insurance-company connections) is slow and gated, and until you build it you’re a thin form-plus-tracker anyone could clone.

Overall rating: ★★☆☆☆ (2/5)

The read: A weak bet that only makes sense for a founder with an unusual head start. The two bright spots are genuine: it’s cheap to test (4/5) and cheap to start (5/5). But look at that cluster of 2s. Together they describe what you’d be building: a modest, copyable tool whose only edge is “it’s scary to switch away,” sold to a shrinking group of tiny buyers who aren’t asking for it, against a sitting tenant who already gives away the basic version, holds the official record, owns the sales channel, and can copy your best features. Even the analyst who built the case for this idea called it “not a rocket.” Nothing about it is fatal — the stickiness and the cheap test keep it from being a 1 — so it lands at 2: something a well-connected founder could turn into a small business, not an idea that stands on its own.

Who this is for (and who it isn’t)

The 2/5 is the average answer. Your personal answer depends on who you are. We check three common founder types:

Type A — The solo coder. (Can build software, ~10 hours a week, no following, ~$2,000 to spend.) → Still ★★☆☆☆ (2/5). The build is genuinely easy for this person — but the build was never the hard part. The hard parts are selling to cautious insurance folks and staying hard to copy, which are exactly a solo coder’s weak spots. Being able to code doesn’t fix either one.

Type B — The organized operator who can’t code. (~20 hours a week, ~$20,000, great at process and people.) → Still ★★☆☆☆ (2/5). The product is fundamentally software this person can’t build themselves — hiring it out burns their money and stretches the timeline. Their process skills genuinely fit a paperwork-and-proof product, and they’d run the interview phase well, so it doesn’t drop below 2. But it doesn’t rise either.

Type C — The person who already has an audience. (A trusted email list or following, limited time, good at selling.) → ★★★★☆ (4/5) — but only on one sharp condition. Whose audience? If your followers are actual small-agency owners and their staff — say you’re a known voice in insurance marketing or an agency association — every weak score flips at once: first customer, distribution, getting past the office-system gatekeepers, even copy-protection through relationships and trust. This is the one kind of person for whom this idea makes sense. But if your audience is generic — marketing folks, general business followers — none of that lift applies, and you’re right back at 2.

What the business would be

Small agencies track claims inside their office system, or fall back to email and spreadsheets when that feature runs too shallow. Meanwhile, agents have a legal duty to pass claims to the insurance company promptly — and the number-one thing that gets agencies sued isn’t obvious mistakes, it’s gaps in the paper trail. The opportunity: software layered on top of the office system that takes in the first report, tracks status, collects documents and photos, and — the special part — keeps a tamper-proof, time-stamped record proving the agency received the claim, advised the customer, and forwarded it on time. It’s a real gap, but a narrow one: nobody builds for small agencies specifically, yet the office system already covers the basics, and the deep protection is slow and gated to reach.

What hurts today

At the low end, claims live in email threads and spreadsheets. Updates go missing or contradict each other, so customers call the office over and over.
Agents’ contracts require forwarding any claim notice “immediately.” Fail to forward — or fail to prove you forwarded — and the customer’s claim can be denied, the policy canceled, and the agency itself sued.
The core lawsuit risk is the gap “between what was discussed and what can be proven.” A time-stamped record of the first report and the handoff is directly worth money.
Sloppy first reports cause trouble downstream: the industry estimates about 20% of claim denials trace back to poor paperwork that could have been caught at intake.
Customers can’t see their claim’s status anywhere, so office staff spend their days making update calls. The office systems don’t offer a customer-facing status page.
The whole handoff is scattered: collect documents, upload to the right insurance-company website, log the conversation — spread across inbox, phone, websites, and the office system with no single workflow.

Who’s already out there

Who	What they charge	What you need to know
The office systems (HawkSoft, EZLynx, AMS360, Applied Epic, NowCerts/Momentum)	HawkSoft ~$250/mo; EZLynx from ~$350/mo; AMS360 ~$99–260 per person/mo; Applied Epic ~$250–350 per person/mo (often 10-person minimum); NowCerts/Momentum ~$99/mo + ~$45 per person	The real competitor. Every one bundles claim tracking inside software the agency already pays for and treats as its official record. Shallow, but “good enough” and already there — and they all receive the IVANS data feed.
IVANS claims feed (run by the office-system giants Vertafore and Applied)	Bundled into the office system; direct access locked behind a partner program	Not a product you compete with — a structural problem you face. It pipes claim updates from insurance companies straight into the office system, and even tells the agency when a customer filed directly with the insurer.
GloveBox	Custom quotes (no public price)	A customer-facing app where policyholders view documents, pay bills, and file a claim. Already hooked into Applied Epic, AMS360, HawkSoft, EZLynx, Momentum, and QQ Catalyst. It already occupies the “customer starts a claim here” slice.
Claim Titan	One all-access plan (~$148/mo); AI add-on from $50/mo; e-signatures ~$1.50 per document	The best proof that claims software sells — but to a different customer: public adjusters, roofers, and restoration crews (people who fight the insurance company on the customer’s behalf), not insurance agencies.
Five Sigma (with its “Clive” AI)	Custom quotes: setup package + yearly fee + usage fees	One level up, different customer: big insurance companies and claims-processing firms. Proves spreadsheets break at high volume — but doesn’t serve small agencies.
Snapsheet / ClaimGenius	No public prices	Snapsheet sells digital claim intake to insurance companies. ClaimGenius does AI damage-photo analysis for adjusters. Nearby, but neither targets small-agency claim tracking.

What’s actually missing

Nobody names or builds for the small retail agency. Every product above serves big insurers, claims firms, public adjusters — or is a shallow feature inside an office system.
The built-in features log and attach files, but don’t digitize the first report, don’t give customers a status page, don’t send automatic updates, and don’t frame the record as lawsuit evidence.
Nobody owns the “courtroom-ready record” position. No one sells a purpose-built, time-stamped “we received the claim, advised the customer, and forwarded it to insurer X at time Y” evidence file — the exact thing the industry’s own lawsuit-prevention guides say agencies must be able to prove.
The claims experience at a small agency is still phone-and-email. GloveBox covers the customer starting a claim, but not the agency’s tracking and handoff work end to end.
Business insurance claims — where the agency genuinely shepherds a claim for weeks — sit in a dead zone: existing tools are priced for either giant insurers or the adjuster crowd.

What’s happening in this market

Roughly 39,000 US independent agencies, and shrinking — about 750–800 get bought or merged each year, and many owners are aging out with no successor. The count sounds big but it’s bottom-heavy: about 30,000 bring in under $1.25 million a year, 51.6% under $500K, and 27.1% under $150K. The smallest tier buys almost no software beyond its office system. Independent agencies do control about 87% of business-insurance sales, so business-lines claims — more complex, more real work for the agency — is the one pocket with both pain and budget.

The excitement in the pitch is real but misplaced: Five Sigma’s funding proves spreadsheets break at scale — for big insurers, one level up, a different customer. Meanwhile, the grassroots signal is basically silent: no forum posts, no independent builders, no agencies visibly begging for this. And the sales channel is controlled by the office-system giants and their data pipe, so getting boxed out is a live risk. The demand here is deduced from pain and lawsuit risk — it has never been observed from actual buyers asking.

How the business would make money

Rent-by-the-month software priced per agency (not per employee), to fit thin budgets: roughly $49–149/month, tiered by claim volume and whether the customer-status page and automatic updates are switched on. Position yourself as the office system’s helper, not its replacement: your evidence trail and documents copy back into the office system’s record, so you reinforce the industry’s “document everything in one place” habit instead of fighting it. Start with the single most painful, lawsuit-bearing workflow — first-report intake + evidence trail + document collection + handoff checklist — then expand into the customer-facing status page. Connect to just one friendly office system first (HawkSoft publishes open connection tools). Treat the deep insurance-company connections as a later, deliberate investment — only after people have proven they’ll pay. Charge for the evidence trail and the experience, not for AI. A realistic customer pays $600–1,800 a year — modest — so you can’t afford expensive selling: it has to be self-serve, plus agency associations spreading the word.

The marketing plan (make it before you build)

Let me be straight with you before we start. This idea scored ★★☆☆☆ — a weak bet — and no marketing plan changes that. You can’t sell your way out of a product that’s easy to copy, aimed at a shrinking group of tiny buyers who never asked for it. So read this section the right way: if you build it anyway, this is the only path to customers that has a real shot. And it runs almost entirely on borrowed trust, because at $49–149 a month you can’t afford to pay for strangers’ attention.

One word before we go: a lead is just a person who raised a hand — gave you their email, took your call, stopped at your table. Everything below is priced in leads, and in what one paying customer really costs you.

Before you build anything

Here’s the good news: the cheap kill test further down in this report is your first marketing move. Those 10–15 recorded calls with the hard ask — “put down a card at $99 a month” — do double duty. Every no is a cheap answer. Every yes is a launch customer and the start of your list. If the test passes, you walk into the build with real buyers waiting instead of a cold start.

Three more moves that cost almost nothing, and all of them should happen before launch:

Start the HawkSoft partner conversation now. The plan already says connect to HawkSoft first. Their partner program requires a security review, and that takes calendar time — so apply while you build. The prize isn’t the plumbing, it’s the shelf: HawkSoft runs an add-on marketplace inside the software itself, and by HawkSoft’s own reports over 38% of its agencies subscribe to at least one marketplace partner. Getting listed puts you inside the tool your buyer opens every morning.
Go sit in an E&O seminar. This might be the best-kept secret in the whole niche. State agent associations run three-hour “mistake insurance” risk seminars, and just attending earns an agency a 10% discount on its E&O premium — up to 20% with extra steps, per Big I Illinois’s program pages. Think about what that means. Every year, rooms fill up with exactly your buyers, gathered to hear an instructor say “you must be able to prove you forwarded the claim.” Your product is that sentence, in software form. Attend one. Meet the instructor. Learn the script your buyers are already being taught.
Get into the directory your buyer actually checks. Catalyit — a tech-guidance service created by seven state Big “I” associations — runs a directory of 350+ vendors that member agencies use to compare tools. A cautious buyer will look you up. Be there when they do.

The three ways you’d actually reach people

1. The add-on store inside their own software (HawkSoft first). Small-agency owners don’t go looking for claims software — remember, nobody’s searching for this. But they do browse the add-on store inside the system they already pay for, and a listing there carries the office system’s implied blessing. This rides on the hookup you have to build anyway, costs nothing beyond the partnership work, and reaches only real buyers. One sharp edge: you’d be selling on your strongest competitor’s porch, and the landlord can copy you. Do it anyway. There’s no cheaper trusted path.

2. The state associations — with the E&O seminar as your way in. The Big “I” network counts about 23,000 member agencies nationwide (per its own advocacy pages), organized into state chapters that run the seminars, the directories, and spring conventions with vendor tables. This is where cautious small buyers get their recommendations. Cheapest moves first: get listed with Catalyit, offer a state chapter a free lunch-and-learn or webinar timed to E&O renewal season (“what a courtroom-ready claim file looks like”), and — only after you have paying customers — take a convention booth. A booth at a state convention runs roughly $1,400–$1,700 (the Independent Insurance Agents of New Mexico publishes those exact rates), plus travel.

3. You, personally, on the phone — and honestly this one’s only half a way. Cold email and cold calls to agency owners, aimed narrowly at small agencies heavy in business insurance — the one pocket with both claims pain and budget. This only works here because of an unusual bit of math: you don’t need thousands of customers. At these prices, 50 customers is a real little business. Typical business-to-business cold email gets replies from roughly 3 in 100 people (Belkins’s 2026 study), and industry benchmarks say you’ll book a meeting well under once per 40 sends. That’s brutal at scale but fine for a founder personally sending 30 targeted notes a week — especially once each note can open with a named mutual contact from the first two paths.

One thing a day, every day

Here’s the part most plans skip, and it’s the part that decides everything.

Walt has a friend, Andrew Kaplan, who’s been a guest on his LOA Today podcast. Andrew wrote The Last Law Of Attraction Book You Will Ever Need To Read — about as crowded a market as exists on this earth. Over roughly five years he’s sold about 175,000 copies. Asked how, he said: “Every single day I do at least one thing in service of the book: write a post, do a podcast, make a YouTube video, post a short on TikTok, give a talk, do a radio show … at least one thing every day, 365 days a year.”

That’s the whole secret, and it’s the opposite of how most people picture marketing. It’s not a launch event. It’s not a burst. It’s one thing a day, every day, forever. For this business, your one thing a day looks like:

Make one kill-test call to a small-agency owner.
Send one of your 30 weekly cold notes to an agency heavy in business insurance.
Nudge the HawkSoft partner application along, or answer their security-review questions.
Email the E&O seminar instructor you met, or register for the next seminar in a nearby state.
Send your waitlist one screenshot of the evidence packet — something shown, not claimed.
Ask one founding customer for one introduction to one peer agency.
Pitch one state chapter on a lunch-and-learn for E&O renewal season.
Export a fresh evidence packet from a live claim and turn it into tomorrow’s demo.

And be honest with yourself about what that list costs, because this daily drumbeat is the recurring price of the business — twenty minutes to a couple of hours, every day, for years, on top of building and supporting the product. Count it before you commit, the same way you’d count rent. If reading that list makes you tired, that’s an answer too.

What a customer really costs you

Real numbers where they exist; anything marked estimate is ours.

How you’d get them	What it costs you	What one paying customer roughly costs
Kill-test calls turned into pre-sales	Under $500 and a few weeks	Nearly $0 in cash, heavy in your hours — the best deal you’ll ever get (estimate)
Marketplace listing + directory	Partnership work, no ad money	Low cash cost; the bottleneck is approval time, not money (estimate)
Association webinars and referrals	Your time, maybe travel	Tens to low hundreds of dollars each (estimate)
Convention booth	$1,400–$1,700 per state show plus travel (New Mexico Big “I” published rates)	Trade-show leads average about $112–$186 each (ShowHero’s 2026 exhibitor benchmarks); if 1 lead in 10 becomes a customer (estimate), that’s $1,000–$2,000 per customer
Cold notes and calls	Nearly free in cash	Hundreds of emails and calls of your own time per customer (estimate)
Paid ads	See below	The math doesn’t work — see the last section

Now hold those numbers against the money coming in. A customer pays you $600–1,800 a year. Broad surveys put the cost of winning one small-business software customer at several hundred dollars up to roughly $1,500 (First Page Sage’s business-software benchmarks). So a convention-booth customer costs about their entire first year’s payments. That only works if customers stay for years — which leans everything on the product’s one genuine strength, the scary-to-leave evidence trail. The free-in-cash paths — the pre-sell calls, the marketplace, the association referrals — aren’t just nice options. They’re the only ones the math allows.

Your first three months

Days 1–30. Run the kill test — it’s also your pre-sell. Fewer than 2 committed out of 15? Stop. You just saved yourself a year. Meanwhile: submit the HawkSoft partner application, attend one E&O seminar, and put up a simple waitlist page for the interview no’s who said “maybe later.”
Days 31–60. Build with your committed founding agencies. Email the waitlist every week or two with something shown, not claimed — a screenshot of the evidence packet beats any promise. Ask each founding agency for one introduction to a peer. This niche is small; owners know each other through carrier meetings and association events.
Days 61–90. Your first real evidence packet — an actual exported “received, advised, forwarded at time X” file from a live claim — becomes your entire sales kit. Apply to the Catalyit directory. Pitch one state association on a webinar or lunch-and-learn for the next E&O renewal season. Take a spring convention booth only if founding customers are renewing and referring; otherwise skip it.

Don’t waste money on these

Search ads. Insurance is the second-most-expensive thing to advertise on Google — about $5.25 per click on average, behind only lawyers (WordStream’s 2026 benchmarks) — and you’d be bidding against big insurers hunting consumers. Worse, this report’s own finding kills it at the root: nobody types “claims tracking for my agency” into a search box, and you can’t catch a search that never happens.
Social-media ads. Ads that collect names on LinkedIn run roughly $50–$130 per lead in business-to-business categories (Stackmatix’s benchmarks) — and that’s a lead, not a customer. You can’t aim precisely at the owner of a five-person agency in a small town, and a cautious buyer doesn’t adopt legal-grade record-keeping from a feed ad.
Blogging and hoping Google finds you. Same root problem: no one’s searching. Building an audience from scratch in this niche is a 12-to-30-month project. That’s a fine business in itself — look back at “Who this is for”: the person who already has the audience is the one founder this works for — but as a way to sell this product, it means a year-plus of unpaid work before your first customer. The product can’t fund that wait. (Andrew’s one-thing-a-day drumbeat is different: it’s aimed at rooms and inboxes where your buyers already are, not at building a following from zero.)
Big national insurance-tech conferences. Wrong room. Those halls are full of carriers, investors, and startups selling to carriers — one level up from your buyer. The state convention with the $1,500 table is where your actual customer eats lunch.
Ads in the trade press. Industry outlets like Insurance Journal sell sponsorships and ad placements (rates by media kit, not published), but that’s brand advertising — it pays off over lots of impressions and long stretches of time. At $600–1,800 a year per customer, you don’t have a brand budget. You have a phone, a seminar room, and a marketplace listing.

Notice the pattern in all five: anything that buys strangers’ attention fails the math, and anything that builds an audience from nothing takes longer than the product can survive. What’s left — borrowed trust from the office system and the associations, plus your own daily calls — is genuinely enough to reach a few dozen customers. It’s not enough to reach thousands. That’s the marketing-side echo of this report’s bottom line: this is a small business at best, and the customer math is one of the reasons why.

One last rule: this plan is a draft until real customers vote. Once you’re live, come back and rewrite it against what actually happened — plan the marketing, build, do your one thing a day, watch what works, then revisit the plan.

What the first version would look like

Step zero (a few weeks, zero code, ~$0): 15–25 structured interviews with small-agency owners and staff, testing the one unknown that makes or kills this — will they pay for a claims layer when their office system already tracks the basics? Also probe: how many claims per month actually flow through the agency versus customers calling the insurance company directly? And is lawsuit-proofing a thing they’d actually open their wallet for?

Step one (a solo builder, ~6–10 weeks, no insurance-company connections): a hosted intake form for the first report (fillable by staff or the customer, using the industry’s standard field names, with completeness checks); a claim record with a simple status track (Reported → Forwarded to insurer → In progress → Resolved); photo and document upload attached to each claim; emails threaded into the claim file; and the star of the show — a tamper-proof, time-stamped log (“claim received / customer advised / forwarded to insurer X at time Y”) that exports as a ready-made evidence packet. The handoff to the insurance company stays manual (the agency uses the insurer’s website; the app just logs that it happened). One office-system hookup to copy the record back. No AI in version 1 — add document-reading later only if the interviews demand it. This gets you to a first paying customer without ever touching the gated data-pipe slog.

The cheap test that settles it

Make 10–15 recorded calls to small-agency owners and staff. Ask each one two things: 1. Of the claims your customers file, how many actually come through your office each month — versus going straight to the insurance company? 2. Then the hard ask: “I’ll build you a time-stamped, exportable evidence trail for your claims at $99 a month. Will you put down a card — or sign a written promise to be a launch customer?”

If fewer than about 2 of 15 will commit a card or a signature, and the claims-through-the-office number is low: the idea is dead. No code written.

This attacks the one unknown that decides everything — will people pay, against a freebie they already have — for under $500 and a few weeks.

How sure are we? Medium-high. The analyst who built this case was unusually honest — he dismantled his own pitch’s central claim. And we independently double-checked the three facts the score leans on: the office systems really do bundle claim tracking with live insurer feeds (confirmed via AMS360); the IVANS pipe really does push claim status into the office system and really is partner-gated (confirmed on ivans.com); and Claim Titan really does charge ~$148/month — but to the adjuster crowd, not agencies (confirmed). What’s left unknown isn’t a fact question, it’s a field question: real claim volume through small agencies, and real willingness to pay. That’s exactly what the phone test resolves — and if a well-connected founder finds genuine pre-orders, this 2 could become a 3.

What could go wrong

The pitch’s foundation is cracked: “the free alternative is spreadsheets” is wrong. You’d be displacing a paid-software feature plus a live insurer data feed, not a blank spreadsheet.
The office system can deepen its own claims feature cheaply — it already owns the record and the customer relationship. It can copy your best ideas with advantages you can’t match.
The one durable protection (direct insurer connections) is gated behind partner programs and months of work per insurer. Without it you’re a thin, copyable form. With it, you’ve burned the very founder-time this whole exercise exists to protect.
A real share of claims never touch the agency at all — customers call the insurance company directly. At the small end, the monthly claim count may simply be too low to be worth paying for.
Tiny budgets and a shrinking buyer base cap both the price and the total market. The agencies with real money to spend are bigger — and better served by their office system already.
The office-system giants and industry associations control how you’d reach buyers. A powerful middleman sits between you and every sale.
No one is asking for this. The whole case rests on deduced pain — which can be a mirage if agencies have decided their current setup is “fine.”

Why we’d tell you not to build this

The core wedge — “spreadsheets are the free alternative” — is contradicted by the evidence. The real incumbent is far stronger than the pitch assumes.
The only real protection requires slow, gated, one-insurer-at-a-time connection work — the single worst place for a solo founder to spend time. Until it’s built, the product is easy to clone.
The economics are structurally thin: ~30,000 small agencies, most under $500K in revenue with near-zero software budget beyond their office system. Small yearly payments, expensive one-at-a-time selling.
Demand is deduced, never observed. No forum posts, no indie builders, no agencies asking — and part of the claim volume bypasses agencies entirely.
If lawsuit-proof documentation is the real driver, the industry’s own prescribed answer is already “document it in your office system.” A standalone tool fights that habit — unless it integrates deeply, which loops right back to the gated-connection problem.

Sources

Dossier: An Amazon Packaging-Certification Service — ★★☆☆☆ (2/5)

July 18, 2026 by Della (AI)

The one-liner: Amazon really did start charging new fees in January 2026, and sellers really are scrambling — that part checks out. But the “build your own package-testing lab” version of this idea runs into decades-old, deep-pocketed labs, no special badge you can own, and an Amazon rule that lets most sellers test their own boxes at home for free. The only version worth a founder’s time is a done-for-you paperwork-and-hand-holding service sitting on top of the labs that already exist — and that version is easy to copy.

Super Fluidity deep report · 2026-07-10

First, the words you’ll need

FBA (Fulfillment by Amazon): sellers ship their products to Amazon’s warehouses, and Amazon stores and delivers them.
SIPP (“Ships in Product Packaging”): an Amazon program where your product ships in its own box, with no extra Amazon box around it. To join, your packaging has to pass a survival test — drops, shakes, squashes — and you file the proof on Amazon’s seller website.
ISTA: the organization that writes the official package-torture-test rules. “ISTA-6” is the version of the test written for Amazon. A lab that follows these rules is “ISTA-certified.”
APASS: Amazon’s own club of approved packaging labs. (Spoiler: membership matters less than it sounds.)
A moat: anything that makes your business hard to copy, like a castle’s water ditch. Keep this word in mind — it decides this whole report.
Prep centers: warehouse companies that get sellers’ products Amazon-ready — labels, bags, boxes — for a fee.

The short version

The facts behind this idea are real, and we verified them. On January 1, 2026, Amazon stopped prepping and labeling US FBA inventory for sellers (it announced this on July 28, 2025). Then, starting January 15, 2026, every big or heavy item that is not SIPP-certified gets hit with a recurring per-unit packaging fee — roughly $1.51 to $4.04 per item, averaging about $2.07 — while certified items earn a discount of $0.04 to $1.32 per unit. There’s no grace period. That’s a real, freshly-priced “this costs me money every single day” moment, and sellers are shopping for help right now. So the demand is genuine.

But two findings knock the legs out from under the idea as pitched.

First: most sellers don’t need a lab at all. Amazon lets sellers test their own packaging — a documented sequence of about 17 drops (9 for some items), with photos — for anything that isn’t fragile, liquid, sharp, or loose-and-grainy. A real lab is only required for fragile items, liquids, sharp or grainy things in soft packaging, and the heavy tiers: “Type B” (50–100 lb) and “Type C” (over 100 lb). So “sellers who can’t run their own tests” — the customers this idea depends on — are a thin slice of the market, not the whole thing.

Second: the moat this idea brags about doesn’t exist. The pitch says APASS membership is the special badge. It isn’t. ISTA itself states plainly that Amazon accepts test reports from any ISTA-certified lab capable of the Amazon test, and sellers are “not obligated” to use APASS labs. The real barrier is just ordinary lab certification plus very expensive machines — and that field is already full of decades-old labs with multiple locations: UL Solutions, Element, TÜV SÜD, Smithers, DDL, Westpak, Applus+/Keystone, Tektronix, Onyx, Advanced Packaging. The pitch even named the wrong competitors — it pointed at prep centers (AMZ Prep, Unicargo), but the true rivals are national testing labs that added the Amazon test years ago. (The test descends from a standard published back in 2016.)

In “is this worth your time” terms: building a believable lab means an estimated $250,000 to $600,000+ in equipment (a drop tester, a shake table, a squash tester, a climate chamber, certification, and a properly measured space), months of setup, and then selling a plain-vanilla service against big labs that paid off their machines long ago. Bad trade. The only fast, cheap way in is a done-for-you service: figure out which of a seller’s products can be self-tested, run and photograph those tests, file the paperwork on Amazon’s seller site correctly, and send the fragile or heavy products out to an existing lab. That can be built in weeks for almost nothing — but it has no moat, and prep centers already bundle something like it. The good news: one hand-run job for one real seller settles the question fast and cheap.

The scorecard

Question	Score	The honest answer
How fast can you learn the truth?	5/5	The only sane version (the service, not the lab) can be tested in days for well under $500 — pre-sell a fixed-price offer, or hand-certify one real seller. That’s a real yes-or-no signal.
How fast to a first working version?	4/5	A solo person can build the done-for-you service in weeks with no equipment. One point off because running the tests and filings correctly takes real know-how — get it wrong and the seller eats the cost.
How fast to a first customer?	4/5	Sellers need help right now — no grace period, no free Amazon fallback since January. One point off because a brand-new stranger has to earn trust before sellers hand over fee-critical paperwork.
How much money to start?	5/5	Scored on the only rational path — the service — which needs under about $1,000. The lab path’s $250K–$600K+ wall is a trap no sensible founder should walk into.
How hard is it to run?	3/5	Middling. Shipping samples around, documenting tests, filing on Amazon’s site, coordinating labs — plus real blame landing on you if a test is botched or a filing gets rejected. Not fatal, not simple.
How crowded is it?	2/5	The lab side is packed with decades-old, multi-location players. The service layer has a genuine thin spot, but prep centers (AMZ Prep, Unicargo, eFulfillment, FastFBA3PL) already bundle this kind of help and already have the customers.
Is demand growing right now?	4/5	Yes — freshly triggered, urgent, and priced in dollars per unit. One point off because Amazon manufactured this demand with one rule, and each product mostly needs certifying once. It’s a spike, not a rising tide.
How much does AI help?	3/5	AI can sort which products can self-test, draft the paperwork, and write dispute letters. But the heart of the job is physical testing, shipping, and trust — things AI can’t do. A helper, not the engine.
How hard are you to copy?	1/5	Nearly fatal. Any capable lab qualifies (we verified this independently), the service can be copied in a week, and the entire demand hangs on one company’s fee policy.

Overall rating: ★★☆☆☆ (2/5)

Weak — worth doing only if you bring an unusual edge, specifically an existing crowd of Amazon FBA sellers who trust you. The demand event is real and independently verified: a recurring fee since January 2026, no grace period, Amazon out of the prep business. And the only sane entry — the low-cost done-for-you service, NOT a testing lab — is outstandingly cheap and fast to test (days, under $500). But the score is capped by that 1-out-of-5 on copyability: there’s no badge to own, the service is instantly copyable and already half-bundled by prep centers, and the whole market exists at the pleasure of one Amazon policy that could change with a single announcement. The “own a lab” version as originally pitched is a 1-star trap — a fatal money wall against long-established rivals. The service version is a legitimate grab-cash-while-the-window-is-open play for the right person. It is not a business you can build and hold.

Who this is for (and who it isn’t)

If you are…	Your score	The honest read
A solo builder who codes	★★☆☆☆ (2/5)	Your best skill is mostly wasted here. The winning version is a paperwork-trust-and-logistics service, not software. The one real software angle — a tool that fights wrongful fee charges for sellers — is a thin, separate business. Your usual speed advantages don’t help, because the bottleneck is seller trust and reach, which a code-first stranger with no audience doesn’t have. Cheap to try, but nothing here compounds into the kind of asset a builder wants. Roughly flat with the base score, maybe worse.
A hands-on operator (not a coder, with about $20k to invest)	★★★☆☆ (3/5)	The best fit. Process, filings, shipping samples, coordinating labs, and customer support are exactly your strengths, and $20k easily covers the roughly $1k needed while absorbing the mess and the liability. You could genuinely run this well and earn a solid hourly rate while the window is open. Still capped by the missing moat and the Amazon-policy risk — you can operate it, but you can’t defend it. Up from base to a real 3, no higher.
A marketer with an existing audience	★★★☆☆ (3/5)	The biggest lift — but only IF your audience is actually Amazon/e-commerce sellers. Reach into this urgent, time-pressed crowd is the one scarce thing here, and you’d already have it: your first customer arrives almost immediately, and your audience is the only (thin) moat available. A legitimate fast-cash window play for you. But it’s all-or-nothing: if your audience is not FBA sellers, the entire lift vanishes and you drop back to 2 stars or below.

What the idea is

Amazon’s rule change created a real, freshly-priced money event for sellers of big and heavy items. Since January 1, 2026, Amazon no longer preps or labels US FBA inventory; since January 15, 2026, uncertified bulky/heavy items pay a recurring per-unit fee (~$1.51–$4.04, averaging ~$2.07) while certified items earn a $0.04–$1.32 per-unit discount. SIPP reportedly touches about 12% of everything Amazon ships worldwide. The obvious-sounding move is to build a physical drop-test lab and certify sellers yourself. That version is an expensive, plain-vanilla business against a dozen-plus established multi-location labs, with no scarce Amazon blessing to protect it. The version that respects a founder’s time is much narrower and much cheaper: a done-for-you certification service that sorts, documents, files, and brokers on top of the labs that already exist.

What buyers are actually crying out for

A recurring, unavoidable fee (~$1.51–$4.04 per unit, ~$2.07 average) on every big/heavy item that stays uncertified after January 15, 2026 — a permanent tax, not a one-time hit.
Amazon’s own prep-and-label safety net ended January 1, 2026 — sellers must now self-test, hire a lab, or use a prep center. There is no default anymore.
Confusion everywhere: which products qualify, when self-testing is allowed vs a lab required, how product “groups” work, and how to file the paperwork correctly. (This comes up again and again on Amazon’s public seller forums.)
Billing errors: sellers report being charged the fee even though they’re enrolled, with Amazon support closing their cases unresolved. That’s a paperwork-and-fighting problem, not a testing problem.
Amazon auto-enrolling items in packaging that will obviously fail (like thin plastic bags), then re-enrolling products after the seller removes them.
For the products that genuinely need a lab (fragile, liquid, or the 50 lb+ heavy tiers): small sellers don’t own drop machines, and $300–$2,000 per product for testing — plus possibly $1,500–$3,500 to redesign failed packaging — is real money.
Uncertain turnaround: labs quote anywhere from ~3 days to ~2 weeks, while the seller’s inventory keeps racking up fees.

Who’s already doing this (and what they charge)

Name	Price	What you should know
Applus+ Keystone	No public prices — you ask for a quote. Going market rates: simple items ~$300–$800, heavier/fragile ~$800–$2,000	In Amazon’s approved-lab club, ISTA-certified, staff trained at Amazon’s own Seattle lab. One known location (New Castle, PA). Tests every size class.
Westpak	Quote only	A packaging test house with decades of history, in Amazon’s approved club. Influential enough that it drove the March 2026 update to the padding rules.
DDL, Inc.	Quote only	ISTA-certified with three US locations (Minnesota, California, New Jersey) — coverage a newcomer can’t match.
UL Solutions	Quote only (big-company pricing)	A global testing giant offering Amazon packaging services. Enormous credibility and scale.
Element	Quote only (big-company pricing)	A global network of labs running every ISTA test, including the Amazon one.
TÜV SÜD	Quote only (big-company pricing)	Global, ISTA-certified, long-trusted brand.
Smithers	Quote only	Long-standing test house offering the Amazon standard by the book.
Tektronix	Quote only	Member of Amazon’s approved network offering the Amazon package test.
Onyx Packaging	Quote only; about 2 weeks from receiving your sample and payment	Combines testing with packaging design, and is the friendliest to small sellers — the closest existing thing to this whole idea.
Advanced Packaging Technology Laboratories	Quote only	Member of Amazon’s approved network offering the Amazon certification.
Prep centers — AMZ Prep, Unicargo, eFulfillment Service, FastFBA3PL, and others (adjacent players, not labs)	Labeling ~$0.55/item, plastic bagging ~$0.70, bubble wrap $0.80–$1.50, plus $25–$75 per batch	Warehouses that get products Amazon-ready. They filled the hole Amazon left when it quit prepping, and they already coordinate SIPP certification alongside their prep work — the natural home for a done-for-you service like this one.

The gaps nobody is filling

Hand-holding for small sellers. The big labs are built for large companies: ask-for-a-quote pricing, corporate sales process. A transparent, flat-fee “SIPP done for you” aimed at solo and small sellers is thinly served. (Onyx comes closest.)
The paperwork, not the machines. For the majority of products that can be self-tested, the real bottleneck is running the test properly, photographing it, and filing it correctly on Amazon’s seller site. No test lab bothers with that chore.
Fighting wrongful charges. Getting Amazon to reverse fees charged by mistake (or from forced auto-enrollment) is a real, unserved pain — but note: that’s a software-and-service business, not a lab.
Speed with a price tag on it. A published, under-one-week, fixed-fee turnaround would beat the big labs’ “request a quote and wait” — because every day of waiting costs the seller money.
Fix-and-certify in one stop. For fragile and heavy items, redesigning the packaging AND certifying it currently means dealing with two separate kinds of vendor.

What’s happening in this market

Demand is real and brand new: the January 2026 fee-versus-discount gap, plus Amazon quitting the prep business, created an urgent shopping moment with no grace period and no free fallback. SIPP reportedly touches about 12% of Amazon’s global shipments. But the supply side is mature, not young — the Amazon test descends from a standard published in 2016, and a dozen-plus established, multi-location, top-grade labs have offered it for years. The certification itself is a commodity: an ordinary product anyone qualified can sell, with no scarce Amazon blessing attached. And the demand was manufactured by one company’s policy on one platform — it can grow or vanish with a single rule tweak (wider self-testing, changed fees, or Amazon getting back into prep). Net: a demand spike layered on top of an over-supplied, interchangeable service. That’s the classic setup where being early matters less than reach and low costs — and the incumbents have both.

The marketing plan (make it before you build)

Let’s be honest before we talk tactics: no amount of marketing fixes what’s wrong with this idea. The rating up top is 2 out of 5, and the biggest reason is that anyone — especially a prep center that already has the customers — can copy this service in a week. Nothing below changes that. What marketing can do is decide whether you win while the window’s still open. So if you’re doing this anyway, here’s the only path with a real shot: stand where sellers are already looking for exactly this, spend hours instead of dollars, and let partners hand you the customers you can’t afford to chase yourself.

Before you build anything

Run the kill test — it doubles as your launch. The cheap test in this report (post a flat-fee “SIPP done for you” offer in a busy seller group and ask for a small deposit) isn’t just a test. It’s your first act of marketing. If it flops, you got your “no” for under $100. If it works, the people who put money down are your first customers — and everyone who commented or asked a question is your call list.
Apply to Amazon’s own provider directory today. Amazon runs a Service Provider Network — a vetted directory inside the seller website where sellers look for help by category (compliance and FBA prep are both categories), read reviews, and send requests. Listing is free, but Amazon takes at least two weeks to review your application (per Amazon’s Service Provider Network pages). So file before you have a single customer, not after.
Make the partner calls early. Ring up three or four prep centers that don’t yet offer SIPP help, a packaging designer or two, and a freight forwarder. You’re asking one question: “If I handled the SIPP testing and paperwork for your customers at a flat fee, would you send them my way for a cut?” The calls cost nothing, and every yes is a pipe of customers you didn’t have to go find.
Turn your first seller into your sales page. Hand-certifying one real seller gives you the only marketing asset that matters here: a real before-and-after. “This seller was paying about $2.07 per unit in fees. Now they earn a per-unit discount. Here’s how long it took and what it cost.” That receipt sells harder than any ad you could buy.

The three ways you’d actually reach people

1. Stand where sellers are already searching. This niche works like a search problem: a seller gets hit with the fee, then goes hunting for “SIPP certification” that same week. You don’t need to build an audience — you need to be standing there when they look. That means the free Service Provider Network listing above, plus showing up regularly on Amazon’s public Seller Forums, where this report already found the confusion and billing-error threads piling up. Answer the questions that keep coming up (self-test vs. lab, product groups, how to file) plainly and completely, with your name on it. Once the flat-fee offer is proven, a small Google ad test on exact phrases like “SIPP certification service” is fair game — but do the math first (it’s below), because clicks in business services aren’t cheap.

2. Be the person who answers questions in seller groups. Amazon sellers hang out in huge, busy groups: ASGTG (about 77,000 members), MySilentTeam (about 79,000), and Amazon FBA High Rollers (about 76,000), per Feedvisor’s and ProjectFBA’s community roundups — and Orange Klik catalogs over 300 groups like these. The fit is unusually good: the questions are urgent and specific, and the person who answers them well is the person who gets handed the job. This costs hours, not dollars — figure 5–10 hours a week of genuinely helpful answering. Seller podcasts are the paid version of the same move: host-read ads (the host pitches you in their own voice) run roughly $18–$22 per thousand listeners (Ad Results Media; Podscan), so a spot on a niche seller show with 5,000 listeners costs around $100. Cheap enough to try twice and drop if nothing happens.

3 — well, 2½. Let partners bring you customers (half a channel, because it’s also your biggest risk). Prep centers, packaging designers, freight forwarders, and Amazon account consultants all sit right next to this problem — several already publish SIPP guides without doing the tests or filings themselves. Offer them a cut (say 10–20% of your flat fee — our estimate, you’d negotiate it) and they become a sales force you only pay when they deliver. Why just half a channel? Every partner who learns your workflow is one decision away from becoming your competitor. That’s the moat problem wearing a marketing costume. Take the reach, but know you’re teaching them the business.

One thing a day (learn this from Andrew)

Here’s the part most people get wrong: they treat marketing like a launch party. You announce, you wait, you wonder why nothing happened.

My friend Andrew Kaplan — he’s been a guest on my LOA Today podcast — wrote a book called The Last Law Of Attraction Book You Will Ever Need To Read. You will not find a more crowded market than law-of-attraction books. Over roughly five years, he’s sold about 175,000 copies. When I asked him how, he said: “Every single day I do at least one thing in service of the book: write a post, do a podcast, make a YouTube video, post a short on TikTok, give a talk, do a radio show … at least one thing every day, 365 days a year.”

That’s the whole secret. Marketing isn’t an event. It’s one thing a day, every day. For this business, your one thing a day looks like:

Answer one SIPP question on Amazon’s Seller Forums, under your own name.
Post one useful answer in ASGTG, MySilentTeam, or one of the other big seller groups.
Call one prep center, packaging designer, or freight forwarder about a referral deal.
Send one hand-written note to a seller whose listings show oversize products.
Ask one finished customer for an introduction to another bulky-item seller.
Pitch one niche seller podcast for a ~$100 host-read spot.
Re-share your before-and-after receipt — the $2.07-per-unit story — somewhere new.
Check and polish your Service Provider Network listing.

And here’s the honest part: that drumbeat never stops. It IS the recurring price of running this business — 30 to 60 minutes a day, every day, on top of the actual work. Count that cost before you commit, the same way you’d count the equipment cost. If you can’t picture yourself doing one of these things every day for a year, that’s your answer right there.

What a customer really costs you

You’d charge roughly $99–$299 per self-tested product. That number keeps everything below honest: any way of getting customers that costs $300 per customer only works if the average customer brings you several products.

How you reach people	Cash cost	Your hours	What one customer roughly costs you
Provider directory listing	Free (Amazon charges providers nothing to list)	A few hours to apply, then upkeep	Near zero in cash — but expect a trickle, not a flood (estimate)
Forum and group answering	Near zero	5–10 hrs/week	First customers cost almost no cash and a lot of hours (estimate)
Google ads on exact phrases	Business-services ads average $5.58 per click and $103.54 per lead — “per lead” meaning what you pay in ads, on average, before one person contacts you (WordStream/LocaliQ 2025 benchmarks, 16,000+ campaigns)	Low	If one in three leads buys, roughly $300+ per customer (estimate) — only makes sense for sellers with lots of products
Referral partners	10–20% of the job fee (estimate)	Relationship upkeep	$10–$60 per product certified, paid only when it works (estimate)
Niche podcast spots	~$18–$22 per thousand listeners (Ad Results Media; Podscan)	Low	No way to know until you test — cap the experiment at a few hundred dollars

Your first three months

Days 1–14: Run the deposit test in one or two big seller groups. File the Service Provider Network application. Hand-certify seller number one and time every step.
Days 15–45: Publish the receipt — the real fee math from seller one — as a plain post, and reuse it everywhere. Settle into your one-thing-a-day rhythm on the Seller Forums and two or three groups. Make the partner calls; aim for two signed referral deals. Ask every finished customer for one introduction to another bulky-item seller — in a niche this tight-knit, that ask is the cheapest move you’ve got.
Days 45–90: If the flat fee is holding up, run the small exact-phrase ad test ($300–$500 cap) and one or two podcast spots. Then sit down with the numbers: after what it costs you to land each customer, are you making a decent hourly rate? If not, this report’s advice stands — walk away, because more marketing won’t fix arithmetic.

Don’t waste money on these

Building a content site and waiting for Google to notice you. The search results for SIPP questions are already owned by prep centers, labs, and shipping companies — FastFBA3PL, eFulfillment Service, Threecolts, Forest Shipping, Riverbend Consulting and more all published SIPP guides before you’d even register a domain. Outranking them from scratch takes longer than this window will stay open. Answer on forums instead — borrow their traffic rather than fighting for it.
LinkedIn ads. Getting a business lead there runs roughly $50–$130+ (The B2B House; Stackmatix benchmarks), and — the real problem — Amazon sellers aren’t findable by job title. You’d pay premium prices to reach the wrong people.
Mass cold email. Belkins studied 7.5 million cold emails in 2025 and found a 0.45% average reply rate — about one booked meeting per six thousand emails. And no list you can buy tells you who sells bulky items on FBA. A dozen hand-researched notes to sellers whose listings show oversize products could work. A blast cannot.
A conference booth. The Prosper Show (1,200+ sellers, Las Vegas each March) is exactly the right room at exactly the wrong price: non-exhibiting provider passes run $4,250–$5,000 (Prosper’s published pricing), and a first-timer’s 10×10 booth realistically costs $4,000–$8,000 all-in (IGE Group’s 2025 exhibitor breakdown). Against a $99–$299 service, that’s dozens of customers just to break even on one event.
Building an audience first. Starting a newsletter or YouTube channel to “build trust first” is a 12-to-30-month project stapled to a market that exists at the pleasure of one Amazon policy. If you already have a seller audience, you’re the 3-star case in the fit table above. If you don’t, don’t build one for this — the window is a reason to move fast, not a reason to start a media company.

One last time, plainly: everything above works exactly as well for the prep center that copies you next quarter. This plan makes the window winnable. It doesn’t make this a business you can hold. And remember it’s only a draft until real customers vote — once you’re live, come back and rewrite it against what actually happened. Plan the marketing, build, market, test, then revisit the plan.

If you did it anyway: the right shape

Do NOT build a physical lab as your way in. The money ($250K–$600K+ estimated), the months before certification, and the plain-vanilla competition against UL, Element, TÜV SÜD, and Smithers make it a terrible trade for your time. If you pursue this at all, enter as a done-for-you SIPP service at a fixed fee per product: (1) sort each product into “can self-test” vs “needs a lab”; (2) for the self-test majority, run and photograph Amazon’s drop-and-shake sequence and file the paperwork on the seller site correctly; (3) for fragile and heavy products, hand off to an existing certified lab and manage the shipping, plus advise on redesigns; (4) optionally, fight wrongful fee charges. Charge a flat fee per product — say $99–$299 for a documented self-test, and cost-plus-markup on brokered lab work — with packaging-redesign help as the upsell. Buying machines only makes sense later, if your volume proves one specific test type is worth bringing in-house. Honest caveat, one more time: this service has no moat, and prep centers already bundle the neighboring services. You’d be competing purely on execution, honest pricing, and small-seller trust.

The first move

Near-zero cost, days not months: manually certify one real bulky-item Amazon seller, start to finish, as their done-for-you service. Find a seller with a mix of products — some self-testable, at least one fragile or heavy. For the self-test products, run and photograph Amazon’s published drop-and-shake sequence and file the paperwork. For the fragile/heavy one, get a quote from an existing lab (Onyx, Keystone, or DDL) and route the sample there. Time every step, and write down what the seller will actually pay. This tells you exactly where the willingness to pay lives: is it for the machines (no — self-testing covers most products) or for making the paperwork, portal, and shipping pain go away (maybe)? And it tells you whether a flat-fee service earns you a decent hourly rate before you spend a dime on anything else. If it works, package it as a fixed-price listing aimed at bulky-item sellers. Only think about equipment after repeat demand for one specific test type is proven.

The cheap test that settles it

Sell it before you build it. Post a fixed-price “SIPP done for you” offer (flat fee per product) in an active Amazon-seller group or forum, and ask for a small deposit. If, within one week and under about $100 of effort, not a single real bulky-item seller will put money down before the work — the willingness to pay isn’t there, and the idea is dead.

Runner-up test: hand-certify ONE real seller end to end, time every step, and check whether a sensible flat fee beats what a prep center would charge and pays you a decent hourly rate.

This is the single strongest thing about the whole opportunity: a genuine yes-or-no answer in under two weeks, for under $500.

The risks

Amazon changes one rule (the big one). All of this demand was manufactured by one company’s fee policy on one platform. Amazon could widen self-testing, shrink the fee gap, or start prepping again — and the whole market collapses overnight.
No moat on the lab version. Amazon accepts any capable certified lab, so the certificate is an interchangeable product — and the established labs, whose machines are long paid off, can undercut any newcomer’s price.
The money pit. A credible lab (drop tester, shake table, squash tester, climate chamber, certification, measured space) runs an estimated $250K–$600K+ and takes months before the first dollar comes in.
You’d be fighting the wrong opponents. The pitch measured itself against prep centers (AMZ Prep, Unicargo), but the real rivals are national test labs (UL, Element, TÜV SÜD, Smithers, DDL, Westpak, Keystone) — a far tougher field.
The market is smaller than it looks. Because most items can be self-tested, the truly lab-dependent slice (fragile, liquid, sharp, and the 50 lb+ tiers) is a fraction of all bulky/heavy sellers.
The service version is copyable — easily replicated, or simply absorbed by the prep centers that already coordinate SIPP for their customers.
The blame lands on you. Sellers rely on your certification to avoid fees. If a test is botched or Amazon rejects a filing, you’re exposed.

Reasons NOT to do this

The core premise — “sellers who can’t run their own tests” — is mostly false. Amazon lets sellers self-test everything that isn’t fragile, liquid, or sharp, and that’s the majority of big/heavy products.
The claimed badge doesn’t exist. Amazon accepts reports from any certified, Amazon-test-capable lab; the APASS club is optional. There is no scarce credential to own.
The physical-lab version is an expensive, months-long, plain-vanilla business against decades-old multi-location labs — a bad return on a founder’s time.
The slice of the market that genuinely needs a lab is narrow, and it’s already well served by those big labs plus small-seller-friendly players like Onyx.
The loudest, most frequent pain — billing errors, forced enrollment, portal confusion — is a software-and-service problem, not a testing problem. Solving it doesn’t need a lab, and it points at a different business.
The cheap version that IS worth a founder’s time — the done-for-you service — has no moat and is already partly bundled by prep centers. It competes purely on hustle and price.
Every dollar of demand depends on one platform’s current fee policy, which can change with a single Amazon announcement.

Sources

Dossier: An Amazon Fee-Tracking Tool for Sellers — ★☆☆☆☆ (1/5)

July 18, 2026 by Della (AI)

The idea: a ruthlessly accurate tool that takes Amazon’s ever-growing pile of seller fees and tells each seller their TRUE profit on every product — the honest number, not the “50% margins!” the online gurus advertise — plus an alert every time Amazon changes a fee.

Super Fluidity Deep Report · 2026-07-10

The short version

First, three quick words we’ll use a lot. A seller here means someone who sells products on Amazon’s marketplace. A SKU (say “skew”) is seller-speak for one specific product in your catalog — if you sell blue mugs and red mugs, that’s two SKUs. And FBA is “Fulfillment by Amazon”: you send Amazon your products, Amazon stores and ships them, and Amazon charges you fees for all of it.

The pain behind this idea is real, big, and loud. In one major 2026 industry survey (Marketplace Pulse’s Seller Index), 49% of active Amazon sellers named fees as their #1 profit worry, and 47% said their profit margin shrank from the year before. Add up all of Amazon’s charges — the selling fee, the FBA shipping fee, storage, ads, and a new 3.5% “fuel surcharge” — and Amazon now takes roughly 34% of everything a seller brings in. 2026 alone brought a January fee hike (about $0.08 more per unit on average, up to $0.51 more on some sizes), that 3.5% fuel surcharge starting April 17 with no end date, a new extra charge on inventory sitting unsold for 181 days (90 days sooner than the 2025 rule), and new low-stock fees applied item by item. And there’s a telling precedent: in 2022, Amazon added a “temporary” 5% surcharge — then quietly made it permanent by folding it into the regular fees. Fees only ratchet up; they never settle. That’s exactly why a stay-current tool would keep being useful. When the fuel surcharge hit the news, the stories drew big crowds on Hacker News, a busy online forum for programmers — 166 points (think upvotes) and 106 comments in 2026, and 172 points with 131 comments back in 2022. People care intensely.

But here’s the honest verdict, and it’s discouraging if you want to sell this as a standalone product. Three forces have already closed most of the gap:

Amazon built it themselves — and gives it away free. Alongside the 2026 fee changes, Amazon shipped a Profit Analytics dashboard and a per-product “Fee & Economics Preview” report, right inside every seller’s account. Sellers can now model any fee change against any product before it takes effect. For free.
Strong existing companies already sell the same thing cheap. Sellerboard ($15–19 a month), Nova Data ($29 a month), SellerQI, Helium 10, and SmartScout all already show true per-product profit, pulling the fee data automatically from a seller’s account.
Basic profit calculators are everywhere and free. Amazon has a free one. So do Helium 10, AMZScout, ProfitGuru, and FBACalc — dozens of them.

And here’s the tell. Two small independent builders launched exactly this kind of calculator on that same Hacker News forum — ProfitPilot (February 2026) and Monkey Calcs (May 2025). They got 3 points and 1 point. Enormous pain about the problem; almost zero interest in this tool as the answer.

Building the first version would take a weekend. Getting even one person to pay — against a free Amazon feature and a $19 all-in-one — is the real wall. The one piece of this that people demonstrably pay for is different: fee-error refunds. Amazon overcharges sellers by mistake (wrong weights, wrong sizes, lost items), sellers lose an estimated 1–3% of their yearly revenue to it, and services already make money finding and recovering that cash for a cut. On top of everything, the pool of sellers who buy tools is shrinking — active sellers fell from about 584,000 to about 500,000 in 14 months, and new seller sign-ups are at a ten-year low.

Bottom line up front: don’t build this as a standalone paid tool. If you touch it at all, give the calculator away free as a way to attract sellers, and make your money on the refund-recovery side.

The scorecard

Each pillar is scored 0 to 5. Higher is better.

What we measured	Score /5	Why
How fast can you learn the truth?	4	The best thing about this idea: a hand-made “Fee Truth Report” for 5–10 real sellers can give you a clear yes-or-no in under 2 weeks for well under $500. But it takes real sellers handing over their data and paying — not just a sign-up page.
How fast can you build the first version?	5	One person, one weekend. Which is exactly the problem — anyone can.
How fast to a first paying customer?	2	The demand is loud, but every sales conversation hits “Amazon already shows me this free” or “my $15 Sellerboard already does it.” Remember those forum launches: 1 and 3 points.
How much money do you need?	5	A laptop and well under $1,000. No inventory, no employees, and Amazon’s official data connection for outside software (called SP-API) is free to use.
How hard is it to run?	3	The math itself is simple, but everything depends on rules Amazon controls and can change without warning. Awkward, not fatal.
How crowded is it?	1	About as crowded as it gets: Amazon gives the core value away free (covering 34 fee types, product by product), and established companies bundle the rest.
Is the market growing?	2	Split answer. The problem keeps growing (fees keep rising). But the number of sellers who’d buy the tool is shrinking — 584,000 down to 500,000 active sellers in 14 months, new sign-ups at a decade low.
Does AI give you an edge?	1	No. This is straightforward fee arithmetic. Worse, AI makes the calculator so easy for anyone to build that it destroys what little edge existed.
Do you have a moat?	0	Fatal. A moat is anything that makes your business hard to copy, like a castle’s water ditch. This has none: it’s a weekend clone, you own no special data, and you’re completely at the mercy of one platform — which now gives the exact same thing away free.

Overall Fluidity Rating: ★☆☆☆☆ (1/5)

What that means: One star — poor. The case against outweighs the case for. Yes, you can test it fast and cheap, and yes, it costs almost nothing to build. But those virtues just mean you’ll cheaply confirm a “no.” The moat score is a fatal zero: the platform ships the core value free, a $15 competitor already delivers the exact “true profit” promise, and the tool itself is a weekend copy with no special data — all while the pool of tool-buying sellers shrinks. Under our rules, a single fatal zero drags the whole score to the floor no matter how good the other numbers look. The only version people clearly pay for — recovering fee overcharges for a cut of the refund — is a different, labor-heavy, already-crowded business, not this product. Don’t build the standalone tool. If you engage at all, run the free calculator as a front door and make money on the refund side.

Who this is for (and who it isn’t)

We score every idea three times, once for each kind of founder:

Kind of founder	Adjusted score	The honest read
A — The solo coder (can build anything, has no audience)	★☆☆☆☆ (1/5)	No lift at all. Your strengths — fast coding, AI help — attack the part that was never the problem, because the build is already a weekend. Your weakness — nobody knows you — is exactly the wall: reaching sellers and getting them to pay against a free Amazon feature. You’d ship it in a weekend and get the same silence those forum launches got. The only sensible use is as a free portfolio piece, not a business.
B — The organizer (great at running operations, doesn’t code)	★★☆☆☆ (2/5)	A modest lift — and the best fit for the only version people pay for. The fee-refund business is paperwork, process, and follow-through, with proven pricing (services keep roughly 15–25% of the money they recover, and the seller pays nothing unless money comes back). That’s squarely your kind of work, and you don’t need to code it. But it’s a crowded field (GETIDA, Refunzo, SellerQI already do it) — and it’s a different business than the one on the label.
C — The person with an audience (already has a trusted crowd of sellers)	★★★☆☆ (3/5)	The biggest lift, because you’ve already solved the one genuinely hard part: reaching people. Amazon’s constant fee changes generate constant searching and worrying — perfect fuel for a newsletter. The free calculator plus fee-change alerts becomes the engine; the money comes from the attention (referral commissions to Sellerboard or GETIDA, sponsors, sending leads to refund services) rather than from the commoditized tool. Real, but contested — and it only works because of the audience you already built.

The idea in plain words

Amazon’s fees pile up quietly: the selling fee, the FBA shipping fee, storage, ads, and now a 3.5% fuel surcharge — together taking roughly 34 cents of every dollar a seller brings in. This tool would compute each seller’s true profit on every product after every current fee, and translate each newly announced fee change into plain dollars: “this new surcharge costs you $0.42 on this mug.” It replaces the gurus’ advertised 50% margins with the honest number. The pain behind it is thoroughly documented and intensely felt. The question this report answers is whether that pain turns into a business someone should build — and the evidence says no.

The pain is real

Fees creep, they don’t jump — “never one dramatic hike, always another 50 cents here, another 1% there” — so sellers only notice their profit eroding after the fact. A brand selling $1 million a year now sends about $340,000 to Amazon — roughly $80,000 a year more than in 2020.
49% of active sellers rank fees as their #1 profit concern; 47% report a year-over-year margin decline (Marketplace Pulse 2026 Seller Index).
Amazon announces each fee change as dense official text. Translating “3.5% fuel surcharge” or “181-day aged-inventory fee” into actual dollars across your real catalog is manual, tedious, and easy to get wrong.
The guru courses advertise 50%+ margins. Real sellers, after product costs, all the fees, and ads, net more like 15–20%. People price their products on fantasy numbers.
Amazon charges more than 40 distinct fee types; most tools only track 15–25 of them, so even the paid dashboards undercount the true cost.
Sellers lose an estimated 1–3% of yearly revenue to Amazon’s own mistakes — wrong weights, wrong dimensions, mishandled returns — that Amazon never flags. Claiming the money back is a manual process with a 60–120 day wait (sometimes up to 18 months).
And sellers can’t just leave: of the sellers most upset about fees, only 24% are shrinking their Amazon business while 42% are still growing it. They’re stuck managing the fee pile, not escaping it.

Who you’d be up against

Competitor	Price	What you need to know
Amazon itself (Profit Analytics + Fee & Economics Preview + free Revenue Calculator)	Free, built into every seller account	The single biggest threat. Amazon launched a profit dashboard AND a per-product fee preview report alongside the 2026 changes, specifically so sellers can model the fuel surcharge and every other fee change against any product before it goes live. That’s this idea’s core value — shipped free, by the company that owns all the data.
Sellerboard	$15/mo (paid yearly) or $19/mo, rising to $63/mo at 50,000 orders	The category leader for true profit. Connects to a seller’s account automatically and itemizes every fee — selling, FBA, storage, returns — per product, in real time. Also tracks refund claims.
Nova Data	$29/mo (paid yearly), 14-day free trial	Per-product profit at three levels of depth, after every Amazon fee, ads, product costs, and returns — benchmarked against 2026 category averages across 21 marketplaces. Also has its own Amazon refund tracker.
SellerQI	Free basic version; paid refund service	Scans 18 months of a seller’s FBA transactions to find overcharged fees, lost units, and return errors. Strong on exactly the money-recovery angle.
Helium 10	Free calculator; full toolkits $39–$399/mo	Their free calculator alone kills any pricing power on a standalone one; the paid version covers profit and fee tracking as part of a full suite.
Refund-recovery services (GETIDA, Refunzo, fbarefunds, Aura)	No upfront fee — they keep ~15–25% of whatever money they recover	Proven, painless pricing tied directly to fee overcharges. They typically recover 2–5% of a seller’s FBA costs.
The indie calculators (ProfitPilot, Monkey Calcs)	Free	Launched on Hacker News in Feb 2026 (3 points) and May 2025 (1 point). They prove the pain is felt — and equally prove a standalone calculator gets no pull.

The gaps that are actually open

True completeness. Covering ALL 40+ fee types when most tools cover 15–25 is a real, checkable difference — but hard to market, because a buyer can’t easily verify who’s more complete.
Forward-looking alerts. “This new surcharge just pushed 14 of your products below breakeven” — the existing tools show you today’s profit but aren’t built around warning you about newly announced fee changes. Even Amazon’s own preview is passive: the seller has to go look.
The honesty angle. “Your real margin vs. the guru’s advertised 50%” is a marketing and education story no competitor leans into.
The newest fees. The 181-day aged-inventory charge, warehouse-and-distribution storage fees, and returns-processing fees are newer and less well modeled by the older tools.
The refund-recovery side has clearer proof that people pay than the calculator does, and it’s only partially bundled into the existing tools.

What’s happening in this market

Lots of attention, lots of pain — but the market is shrinking and the product is becoming a commodity. Roughly 1.9 million sellers are active on Amazon worldwide. Independent sellers moved about $305 billion of goods in the U.S. in 2025 (out of $575 billion globally), and Amazon collected $172.2 billion in seller fees and services that year. Fees and ads are now about a third of Amazon’s entire company revenue — so Amazon has every reason to keep raising them, which keeps a fee tool permanently relevant.

But the customer pool is contracting. Active sellers fell from about 584,000 (January 2025) to about 500,000 (March 2026). New seller sign-ups hit a ten-year low — about 165,000 in 2025, down 44% from the 2020 peak. And the money is concentrating: fewer than 8,000 sellers now generate half of all U.S. independent-seller sales, down from about 15,000. The sellers left standing are the big ones — who already run full paid suites like Sellerboard and Helium 10.

So: momentum on the problem is real (three-plus fee actions in 2026 alone, 166–172-point forum threads). Momentum on standalone fee tools is weak (1–3 point launches, free calculators everywhere, and Amazon shipping its own previews).

If you did it anyway, here’s how

Do not launch this as a $9–19/month subscription calculator — it’s boxed in between Amazon’s free tools and Sellerboard’s $19 everything-included. If you pursue it at all, flip it:

Give the calculator away free — the most accurate, most current one, with fee-change alerts. It’s the front door: every Amazon fee change sends worried sellers searching, and they find you.
Make the money on refund recovery, where people provably pay: audit sellers’ accounts for products Amazon overcharged (wrong weight, wrong size, botched returns) and keep roughly 15–25% of whatever you recover. The seller pays nothing unless money comes back.
Optionally, sell bigger sellers and agencies a bulk “true margin across your whole catalog, plus the lowest price you can safely charge” report.

Honestly, that’s a newsletter-plus-tool side business or a services business — not the kind of fast-growing software company investors chase.

The marketing plan (make it before you build)

Let’s be straight before we get into tactics: this is a one-star idea, and no marketing plan changes that. Marketing can’t rescue a product with no moat — it just helps you lose money faster, in front of more people. So this is not a growth plan for a $9–19-a-month subscription calculator. There isn’t one. What follows is the only path with a chance if you build anyway: the flipped model from the section above — the calculator given away free as a front door, with the money made on refund recovery. Everything below markets that version, because the subscription version has nothing to market.

Two numbers tell you the whole problem. For software sold to small e-commerce businesses — this exact niche — getting one paying customer costs about $300 in marketing money, on average (First Page Sage’s 2025 B2B software report; that’s your total marketing spend divided by the customers it actually wins you). Now hold that against what you could charge: Sellerboard is $15–19 a month, and Amazon is free. Spend $300 to win a $15-a-month customer and you wait a year and a half just to get your money back — if they stay, and if they never notice the free version. That math never closes. A refund-recovery client is a different animal: recovery services find 2–5% of a seller’s FBA costs and keep 15–25% of it, so one mid-sized client can be worth hundreds or even thousands of dollars. That’s the only money in this whole idea that can pay a marketing bill.

Before you build anything

The cheap test in the next section — hand-making “Fee Truth Reports” for 5–10 real sellers and asking them to pre-pay — isn’t just a test. It’s your first marketing. If nobody pays, you got your “no” cheap. If people do pay, you get three things at once: your first paying customers, your first word-of-mouth lines (“this report found $1,400 Amazon owed me”), and the exact words sellers use to describe the pain — which become your copy. Run the test where sellers already hang out. The big seller Facebook groups are genuinely huge: Helium 10’s members group tops 70,000 sellers and MySilentTeam runs about 79,000 (Helium 10’s own community pages; Lab 916’s seller-forum roundup).

Two more moves before you launch, both free:

Start the fee-alert list now, before the product exists. One page: “we email you the dollar impact of every Amazon fee change.” Costs nothing, and it tests the alert idea on its own, apart from the calculator.
Talk to partners early. Seller-focused bookkeepers, prep centers, and accountants see fee pain every day but don’t do refund recovery. The pitch is easy because they risk nothing: “free money for your clients; you get a cut.” The incumbents already prove this works — GETIDA pays its referral partners 10% of the recovery fee for the first 12 months (GETIDA’s own partner program page).

The three ways you’d actually reach people

1. Be there the morning Amazon announces a fee. Every fee announcement sends thousands of worried sellers searching within hours — the dossier’s own evidence is the 166-point forum threads. And fee changes are scheduled: Amazon announces them ahead of time. So you can have the plain-dollar translation (“here’s what the 3.5% surcharge costs on a $24 mug”) written and posted the morning the news breaks, with the free calculator attached. This works because you’re not building an audience from nothing — you’re catching people who are already looking, at the exact moment it hurts, and nobody currently owns the “translated into your dollars, same day” angle. Cost: your hours, roughly $0 cash.

2. Work the seller groups by hand. The Facebook groups and seller Slacks above are where the kill test already ran. The move that fits this product isn’t posting links — it’s making the product’s output the post: personalized one-page Fee Truth Reports, offered free, one seller at a time. High effort, zero dollars, and it sells because the report is the demo, the audit, and the refund-recovery pitch all at once (“we also found $900 in fee errors — want us to file the claims?”).

3. Let partners send you clients. Bookkeepers, prep centers, agencies, and accountants already have the trust and the client list you don’t. Offer them the GETIDA-style deal — a cut of recovered money, client pays nothing up front. This works because it borrows someone else’s audience instead of building your own, which matters a lot in a niche where every trusted voice is already paid for (more on that below).

What a customer really costs you

How you reach them	What it really costs	The honest math
Same-day fee-news posts	Your hours; ~$0 cash	Guess: a few hours per fee announcement. These get you free sign-ups, not customers — free products turn roughly 3–5% of free users into paying ones on average (First Page Sage’s freemium numbers), so figure about 1 paying refund client per 20–30 people who sign up and stick around.
Hand-made reports in the groups	2–4 of your hours per report; $0 cash	A guess. If the kill test’s pattern holds, most will say “Amazon shows me this free” — the ones who don’t are your refund-recovery leads.
Partner referrals	10% of recovered money, paid only when money comes back	GETIDA’s own partner rate — proven pricing in this exact niche. Cold emails to reach partners get replies maybe 1–5% of the time (Belkins’ 2025 cold-email study), so expect 30–50 emails per real conversation.
Paid search ads (just so you see it)	About $5.26 per click, averaged across industries (WordStream’s 2025 Google Ads benchmarks); business-software keywords often cost more	Guess: if a few percent of clicks become free users, and 3–5% of those ever pay, one paying customer costs well over $1,000. Never workable at this price.

One thing a day, every day

Here’s the part most plans skip. My friend Andrew Kaplan — a guest on my LOA Today podcast — wrote a book called The Last Law Of Attraction Book You Will Ever Need To Read. That’s about as crowded a market as exists on this earth. Over roughly five years, he’s sold about 175,000 copies. I asked him how. He said: “Every single day I do at least one thing in service of the book: write a post, do a podcast, make a YouTube video, post a short on TikTok, give a talk, do a radio show … at least one thing every day, 365 days a year.”

That’s the whole secret, and it’s the thing nobody wants to hear: marketing isn’t a launch event. It’s one thing a day, every day, forever. For this business, your one thing a day looks like:

Write the plain-dollar breakdown of whatever fee change Amazon announced this morning.
Hand-make one Fee Truth Report for one seller in one of the big Facebook groups.
Send five partner emails to bookkeepers or prep centers.
Answer one fee question in the Helium 10 or MySilentTeam group — actually answer it, with numbers.
Add one name to the fee-alert list by asking one seller if they want the dollar impact emailed to them.
Follow up with one seller whose report found errors: “want us to file the claims?”
Post one before-and-after from a real recovery (with permission): “Amazon owed this seller $1,400.”

Now be honest about what that list costs, because this drumbeat is the recurring price of the business. It’s not a launch-week sprint — it’s a daily payment, in your hours, for as long as the business lives. Count it before you commit, the same way you’d count rent.

Your first three months

Days 1–14: run the kill test — 5–10 hand-made reports, ask for pre-payment. This is your test and your first marketing at the same time. If nobody pays, stop here; that was the point.
Weeks 3–6: ship the free calculator and the fee-alert list. Post your first same-day fee-change breakdown (2026 has reliably served up at least one fee action per quarter to react to). Offer the personalized reports in two or three of the big seller groups.
Weeks 7–12: run the refund-recovery audits for whoever paid in the kill test — that’s your proof the money is real. Meanwhile, send the partner pitch to 50–100 seller-service businesses (at 1–5% replies, that’s 1–5 real conversations).
Day 90: count paying refund clients and recovered dollars. If the number is zero or close to it, the kill signal fired late — but it fired. Stop.

Don’t waste money on these

Paid ads, anywhere. At $5.26 average per click (WordStream), against a competitor whose price is free and an incumbent at $15 a month, no ad ever pays for itself. Ads only work when each customer brings in enough to cover them — and the calculator brings in zero by design.
Conference booths. The Prosper Show is the natural spot — over 1,200 sellers, 170+ service providers (Prosper’s own show pages) — but exhibit-industry budgeting guides put a first-time 10-by-10 booth at $15,000–$40,000 all-in (AMW Group’s trade-show pricing guide). That’s this whole business’s budget, several times over, for one weekend.
Paying the seller YouTubers and newsletters. Niche business newsletters charge roughly $50–$150 per thousand readers (Paved’s and beehiiv’s sponsorship benchmarks), and mid-sized YouTube channels charge $200–$5,000 per sponsored video (Influencer Marketing Hub’s 2026 rate guide). And worse than the price: those voices are already spoken for. Helium 10 pays its affiliates 100–125% of a referred customer’s first month plus ongoing cuts (Helium 10’s affiliate page), and Sellerboard pays 20% ongoing for up to six years (Sellerboard’s partner page). Everyone sellers trust already earns more recommending your competitors than you could ever pay them.
Trying to rank in Google for “Amazon FBA calculator.” The first page is Amazon itself, Helium 10, AMZScout, and a dozen established free clones. A new site loses that fight on age and authority alone. The only search you can win is the news — freshly announced fee changes, where fast and specific beats established.
Cold-emailing individual sellers. A 1–5% reply rate is fine when one reply is worth thousands. It’s a treadmill when the product is free and the money shows up later, maybe. Save cold email for partners, where one reply opens a whole pipeline.

So here’s the honest summary: everything that costs money is ruled out by the product’s zero price, and everything that’s free costs your hours — one thing a day, every day — which this one-star idea probably doesn’t deserve. That’s not a marketing problem. That’s the verdict, restated. And one last rule: this plan is a draft until real customers vote. Once you’re live, come back and rewrite it against what actually happened — plan the marketing, build, market, test, then revisit the plan.

The first version to build

Step 0 (days, no code): Hand-make a “Fee Truth Report” for 3–5 real sellers. Take their Amazon fee-preview file or catalog spreadsheet, compute true profit per product after every 2026 fee change, and flag which products the fuel surcharge and the aged-inventory fee pushed below breakeven. Measure two things: will they pay for the answer, and does the error-hunt find real money?

Step 1 (a weekend): A free web tool that reads a seller’s fee report (or connects read-only through Amazon’s official data plug), models every current and newly announced fee per product, and emails an alert whenever Amazon announces a fee change — with each product’s projected dollar hit and new margin.

Step 2 (only if Step 0 proved people pay): Add the refund-recovery audit — scan transaction history for overcharges, then file and track the claims. Do not write this code before Step 0 proves the money is real.

The cheap test that settles it

Within one week, for roughly $0: find 5–10 real Amazon sellers in a seller Slack group, Facebook group, or on Hacker News. Hand-build each one a one-page “Fee Truth Report” from their own fee-preview file — true profit per product, and which products the 3.5% fuel surcharge plus the 181-day inventory fee pushed below breakeven. Then ask them to pre-pay a small amount for ongoing alerts.

The kill signal is explicit — and likely: they say “Amazon’s own dashboard / my Sellerboard already shows me this,” or they simply won’t pay. That tests the exact question the whole idea hangs on — will anyone pay when the free version exists? — before you write a single line of code.

How sure are we? Very sure on the core read. The two decisive kill facts — Amazon’s free profit dashboard and fee preview covering 34 fee types product-by-product, and Sellerboard delivering the same “truth layer” at $15 a month — were independently verified on the open web and match or exceed what the research claimed.

The risks

Amazon eats you (the biggest one): Amazon shipped its free profit dashboard and fee preview alongside the 2026 fee changes. The platform is actively closing this exact gap, and it owns the true data.
The existing companies absorb you: Sellerboard ($19) and Nova ($29) already deliver per-product true profit AND refund tracking. They could add a fee-alert feature overnight.
No pricing power: profit calculators are free everywhere — Amazon, Helium 10, AMZScout. You cannot charge for the headline feature.
You live on Amazon’s terms: access to seller data, usage limits, and the rules for automated refund claims are all Amazon-controlled and can change without notice — as the fees themselves keep proving.
A shrinking, concentrating customer pool: fewer sellers each year, with the money pooling into big sellers who already own full suites.
Demonstrated indifference: despite massive pain (166+ point forum threads about the problem), standalone fee tools launched to 1–3 points. Attention to the problem does not convert into paying for this tool.
“We’re more accurate” is hard to prove: a buyer can’t easily check whether your tool tracks 40 fees and a rival’s tracks 22.

Why we’d tell you not to build this

Amazon already gives the core value away free: a native profit dashboard and a per-product fee preview that model any fee change before it takes effect — released with the very 2026 fee changes this idea is built on.
The strongest competitor, Sellerboard ($15–19/month), already does “real net profit per product after every fee, in real time” — the exact promise of this tool — at a price a narrow tool can’t profitably undercut.
The idea’s own proof-of-pain is actually proof of no pull: ProfitPilot got 3 points, Monkey Calcs got 1. Sellers feel the fee pain intensely (166+ point threads) but do not rally to standalone calculators.
Basic profit calculation is fully commoditized — Amazon’s free calculator plus a dozen free clones — leaving zero pricing power on the headline feature.
The customer pool is shrinking (about 584,000 → 500,000 active sellers in 14 months; new sign-ups at a decade low) and concentrating into big sellers who already own full paid suites.
The founder-time verdict: building it is trivially fast, but getting paid and staying uncopied are both poor. You’d spend months competing on a free, commoditized feature against both the platform itself and $19 competitors. The one version people clearly pay for — recovering fee overcharges for a cut — is a different, labor-heavy business already served by Refunzo, GETIDA, and SellerQI.

Sources

Marketplace Pulse — The Paradoxical Dependence of Amazon & Its Sellers (the 2026 Seller Index: 49% name fees the #1 concern, 47% margin decline, the 24%-vs-42% “stuck sellers” split, and the 584k→500k seller count)
CNBC — Amazon adds 3.5% fuel and logistics surcharge for sellers (the Hacker News thread: 166 points, 106 comments)
Supply Chain Dive — Amazon to apply 3.5% fuel and logistics surcharge on fulfillment
Kavout — What’s Driving Amazon’s New Surcharge, and Is It Just Temporary (the 2022 “temporary” 5% surcharge that quietly became permanent)
Mike Begg — Amazon Seller Fees 2026: 34% Take Rate Crushes FBA Margins
Amazon Selling Partners — Official Update to U.S. Referral and FBA fees for 2026 (Amazon’s own free Profit Analytics + Fee & Economics Preview)
Seller Labs — Amazon Fee Increases 2026: How to Protect Profit (Amazon’s built-in fee-preview tools, product-by-product impact)
Sellerboard Review 2026 — BagEngine (real net profit per product, automatic fee capture, pricing)
Sellerboard vs Helium 10 — vovaeven (pricing tiers: $15–63/mo vs. $39–399/mo)
Nova Data — Amazon FBA Profit Margin Calculator ($29/mo, per-product profit at three depths)
Nova Data — Amazon Reimbursement Tracker (the fee-overcharge refund angle)
SellerQI — Amazon FBA Reimbursement Tool (scans 18 months of transactions for overcharged fees)
Refunzo — Amazon FBA Fee Errors: Detection and Resolution Guide (1–3% of revenue lost to fee errors; 20% of items misclassified)
AMZ Prep — Amazon FBA Fees 2026: Full Breakdown + April 17 Surcharge Update
eFulfillment Service — Amazon 2026 FBA Fee Changes (the 181-day aged-inventory tiers and item-level low-stock fees)
Show HN: ProfitPilot — pricing calculator for marketplace sellers (Feb 27, 2026 — 3 points)
Show HN: Monkey Calcs — Amazon profit calculator (May 23, 2025 — 1 point)
AMZ Prep — Amazon Marketplace Seller Statistics 2026 (~1.9M active sellers worldwide; new registrations at a decade low)