Paid Idea Landmine Report
A self-serve web service where a would-be founder pastes a business idea and buys, for about $29, an AI-generated report that lists the idea's hidden assumptions, checks each against cited evidence, and prescribes the cheapest next test.
Selling it one report at a time to novices means fighting 20 tools for the same searches at a 2 to 3 percent conversion rate, and the subscription will not work because people validate one idea and leave. Keep the prototype, but change three things before spending on marketing: drop the consumer subscription for packs or business seats, make 'the way around each landmine plus the cheapest test' the visible difference, and pick a distribution channel or language market the incumbents do not cover. Run the four cheap tests below in the next month; if novices will not pay after reading a free report and partners will not take it, stop.
A5 together with A1: acquiring a $29 buyer in a category where 21 tools chase the same queries and the incumbents already rank. Even a better report loses if nobody finds it, and the incumbents' own conversion data implies roughly 40 free users per sale.
Upward: a blind side-by-side where target users pick your report over Preuve and free ChatGPT output for its assumption-and-test structure, plus one partner (an SBDC, university program or accelerator) committing to buy reports in bulk. Downward: free-report recipients declining to pay anything afterwards, cost per purchase above $29 in the channel test, or citation error rates above 5 percent.
The assumption chain
For this idea to work, these things need to be true. Most important first; the ones the whole idea rests on are marked critical.
People do pay for this: DimeADozen reports 3,100+ paying customers against 134,000+ ideas submitted, roughly a 2 to 3 percent free-to-paid rate, and it made $58,000 in its first six months back in 2023. But Preuve AI already sells an 18-section, source-linked report with pivot directions at exactly $29, and its own comparison page counts 21 validators, six of which link sources. A me-too report at the same price has no reason to win the buyer's search.
Do not sell 'a report'. Sell the piece none of them lead with: an assumption-by-assumption verdict with the cheapest real-world test attached to each, and prove the difference in a blind side-by-side. Or change the market (non-English) or the channel (partners) so you are not fighting for the same search result.
What to look for：Confirmed if at least 5 of 8 pick yours and name the assumption-and-test structure as the reason. Refuted if they cannot tell yours apart or prefer the free ChatGPT output.
The founder's own doubt is well founded. In a survey of 2,994 new business owners, 81 percent put their odds of success at 7 in 10 or better and the poorly prepared were as optimistic as the well prepared. Confirmation bias research says people discount data that contradicts their beliefs, and entrepreneurs' high optimism intensifies it. ValidatorAI's data says only 29.3 percent of its users act after validation. Yet scrutiny products do exist and sell: DontBuildThis markets 'brutally honest feedback' and Preuve reports only 17.5 percent of ideas earn a go while still charging $29. QuickMVP's lesson is the sharpest: it was good at invalidating ideas but did not help users pivot, and retention collapsed to three months.
Frame every landmine with a way around it, so the report reads as 'how to make this work' not 'why this fails'. Never deliver a bare kill verdict; always ship a pivot path and a next test. Measure refund requests and 'did you change your plan' as your core quality metrics.
What to look for：Confirmed if the 'way around' version gets more voluntary payment and at least half report a concrete plan change. Refuted if most pay nothing and describe the report as discouraging or 'just opinions'.
ChatGPT's free plan includes limited deep research with citations, rolled out to free users in April 2025, and Claude's free plan includes web search. Back in 2023 an Indie Hackers commenter already noted one could do DimeADozen's report for free in ChatGPT and that the product simply packaged the work to save time. Incumbents survive by pointing out that a chat tells you what you want to hear and by adding live-source verification. Your edge has to be structure and verification the user would not prompt for themselves.
Make the report do things a chat will not do unprompted: enumerate assumptions, cite and verify every claim against the fetched page, record searches that found nothing, and attach a costed test to each assumption. Publish sample reports beside a ChatGPT deep-research output on the same idea so the difference is visible before purchase.
What to look for：Confirmed if yours wins on all three counts for at least 4 of 5 ideas. Refuted if ChatGPT matches you on assumptions and citations, because then only packaging is left.
DimeADozen's own facts page shows 3,100+ paying customers across 3,200+ purchases, about 1.03 purchases per customer, and it sells 'no subscription' as a feature. Preuve sells 5- and 10-report packs and a $19 per month re-scan add-on for people already building, not a subscription to validate more ideas. QuickMVP, a validation tool, lost a third of its users every month. Validation is a once-per-idea purchase and most people have one idea at a time.
Drop the consumer subscription. Monetize repeat use through multi-report packs, a post-launch monitoring add-on, or by selling seats to consultants, incubators and educators who validate many ideas a month.
What to look for：Confirmed if over 20 percent choose the subscription and generate a second report within 60 days. Refuted if under 10 percent subscribe or if subscribers churn after one report.
DimeADozen grew in 2023 by getting listed on AI tool directories and ranking first for 'validate your business idea AI', a window that has since closed: Preuve counts 21 tools in the category in 2026, all chasing the same queries. I found no published search-volume figure for 'validate business idea', so top-of-funnel demand is unmeasured. Category norms of 2 to 3 percent free-to-paid mean you need roughly 40 free users per sale.
Pick a channel the incumbents do not own: partnerships with Small Business Development Centers, university entrepreneurship programs, lenders and accelerators that hand the report to every applicant; or a non-English market where the incumbents say they lack depth.
What to look for：Confirmed if paid cost per purchase lands under $15 or communities produce sales at zero cost. Refuted if cost per purchase exceeds $29 or no community post produces a single sale.
In 2023 DimeADozen was built in days, hit $16,700 in a month, and was listed for sale at $160,000. By 2026 the same capability sits in ChatGPT's free plan and in 21 competing tools. The tech makes your product possible but no longer scarce. Timing favors whoever owns a channel or a market, not whoever ships another report.
Treat the report as table stakes and put effort into distribution partnerships or a language market the incumbents skip.
What to look for：Confirmed if your target market's queries return few or no dedicated tools. Refuted if the incumbents already own the first page.
GummySearch, a founder-research tool used by 135,000 people, closed to new customers in November 2025 because it could not get a commercial Reddit API license. Reports that lean on Reddit, Product Hunt or Crunchbase for demand signals inherit that platform risk.
Prefer public primary sources such as government statistics, company pricing pages and court or regulator texts. Use community data only through licensed search APIs and label it as secondary.
What to look for：Confirmed if under 10 percent of citations rely on platforms that forbid commercial scraping. Refuted if Reddit-style sources dominate the evidence base.
I found no licensing requirement and every incumbent operates openly with a disclaimer and a 14-day refund policy. I did not check advertising rules on claims like 'verified' or 'accurate', which is where consumer-protection exposure would sit.
Avoid accuracy percentages in marketing, label estimates as estimates, and offer a refund guarantee like the incumbents do.
What to look for：Confirmed if no claim needs substantiation you cannot provide. Refuted if your copy promises outcomes or accuracy figures you cannot back.
Suggested next steps
6 steps, cheapest and most telling first.
Cheapest way to learn whether you have a product difference at all before spending on anything else.
Tests the encouragement-versus-scrutiny doubt with real novices and produces the cost and citation-accuracy logs for free.
Measures cost per purchase and whether anyone chooses a subscription, the two numbers that decide whether the consumer model can work.
The most likely way around both the acquisition-cost and the no-repeat-purchase problems; one committed partner changes the business model.
Checks whether an uncontested market exists where the same prototype faces no incumbents.
Low-cost protection against the platform-shutdown risk that killed GummySearch and against unprovable accuracy claims.
Directions worth considering
Keep your intent, avoid the landmines found above.
Trade-off：Slower sales cycles, procurement paperwork, and Preuve and DimeADozen already offer white-label tiers, so you must win on the assumption-and-test format or on price.
Trade-off：Different evidence sources and legal landscape, a separate distribution effort, and the English prototype's source pipeline will need rebuilding.
Trade-off：More product to build, a narrower audience of people willing to actually run tests, and a longer time to first revenue.
Trade-off：Labor-bound rather than software-scalable, needs a bench of reviewers, and the free report becomes a cost center that must convert.
Existing players and precedents
Who is doing this, who tried and stopped, and what each one can teach you.
This is your idea at your price already on sale. It markets against 'ChatGPT telling you what you wanted to hear', admits real per-report AI costs, and monetizes repeat use through packs and monitoring rather than a validation subscription.
Built on GPT prompts in 2023, it made $58,000 in six months via AI directories and SEO and was listed for sale at $160,000. Early users called it a template filled by ChatGPT, so it moved upmarket to 100-page cited reports. Its 1.03 purchases per customer shows validation is a one-time buy.
The category's largest audience chose to give the validation away and monetize attention. Its own data says only 29.3 percent of users act after validation, which shows most visitors are explorers, not buyers.
The free substitute the founder worried about is real. Anyone can get a cited research report on their idea at no cost, so a paid product must win on structure, verification and prescription rather than on 'has citations'.
Proof that a scrutiny-first positioning attracts users, and also proof that the scrutiny itself is being given away. The paid layer is the 'recovery' path, which matches the lesson that users pay for the way around, not the kill.
People pay far more to be handed ideas than to have their own idea checked. That says the willingness to pay sits with serial builders, not with one-idea novices.
Closed to new customers in November 2025 because it could not obtain a commercial Reddit API license, with full shutdown set for December 2026. A warning for any report that leans on Reddit for demand signals.
Usage was strong but retention was three months because most users got no signups, and the tool invalidated ideas without helping them pivot. The company ran out of money and laid everyone off. The lesson: a tool that only says no gets abandoned.
This is a crowded, low-moat category. At least 21 AI idea validators are live, two of them sell almost exactly this product at $29 and at $9 to $129, the largest one is free, and ChatGPT's free plan produces cited research reports. The incumbents' own numbers say the buyer pool converts at around 2 to 3 percent from free to paid and buys about once. Two dead products teach the same lesson from different eras: users abandon tools that only invalidate, and tools that depend on one platform's data can be shut off. The room left is not another report; it is a channel the incumbents do not own, a market they explicitly skip, or a product that turns each landmine into a costed next step.
The idea as we understand it
The choices the idea leaves unsaid, made explicit. Correct us where we read it wrong.
A service where a person types in their business idea and gets back a polished report that surfaces the idea's hidden assumptions, checks each one against real evidence with cited sources, and tells them the cheapest thing to test before spending money. Sold per report at around $29, with a subscription for repeat users. Target customer: people who want to start a business but don't really understand business logic and want to find the landmines early.
We have a working prototype. Our own doubts: people with a new idea may want encouragement rather than scrutiny, and ChatGPT already gives opinions on ideas for free.
First-time, mostly non-technical would-be founders in English-speaking markets who have an idea but little business training and have not yet spent real money on it.
The user pays out of pocket: roughly $29 per report, with an intended monthly subscription for people who keep generating ideas. No third party pays.
Web app reached by search, Reddit and Indie Hackers posts, AI tool directories, and word of mouth; card checkout via Stripe or similar; report delivered on screen and as a PDF.
Large language models with web search can now assemble a cited research report for a few dollars of compute in minutes, which made this product buildable by a solo team in 2023 and made it a crowded category by 2026. The same shift put free cited research into ChatGPT's free tier.
AI startup-idea validation and pre-launch market research tools, a sub-segment of tools for aspiring entrepreneurs that also includes business-plan generators and idea catalogs.
The idea does not name a market or language. I assumed a US and English-first web product because the prototype is described in English and every incumbent is English-first. I read the subscription as monthly repeat access for serial idea generators rather than a monitoring service. I read 'people who don't understand business logic' as main-street and side-hustle founders, not venture-backed tech founders. If you intend a Chinese-language market, several verdicts below improve, see the alternatives.
Answer these and the analysis gets sharper.
- Which market and language do you intend to launch in first? Every verdict above assumes US English; a Chinese-language launch changes A1, A5 and A8 materially.
- What does one prototype report currently cost in model and search spend, and how long does it take to generate? I estimated $3 to $6 from list prices but did not see your logs.
- Have any prototype users paid or been asked to pay yet? Even five real payment decisions would outweigh most of the category proxies used here.
- Who is 'we', and does your team or Authright have an existing audience or partner network of aspiring founders that could serve as the channel?
- Is the target really business novices, or tech founders? Incumbents report their best fit is tech-startup ideas, and novices are the segment most prone to wanting encouragement.
- I could not find search-volume data for the core buying queries; if you have Google Keyword Planner access, the monthly volume for 'validate business idea' terms would sharpen A5.
Evidence
21 pieces of evidence (19 sources, 2 search records). 19 quotes were matched word for word against the archived text.
“Founder Report: $29 one-time for the full 18-section analysis with up to 15 competitors, go-to-market plan, demand signals, source-linked evidence, 3 pivot directions”
Preuve AI's pricing page: a direct competitor selling a source-linked report with pivots at exactly $29, plus packs, a $19/month re-scan add-on and white-label; solo founder citing real per-report costs.
“Paid reports start at $9 (the Starter, a 7-section report of typically 20+ pages, with 60+ URL citations); the full Entrepreneur report is $129 (typically 100+ pages); the idea score is free. One-time prices, no subscription.”
DimeADozen's pricing page: $9 to $129 one-time tiers, 'no subscription' sold as a feature, and an explicit statement that non-English markets are not its fit.
“Only 29.3% build after validation — 87.2% of those who act move within 10 minutes.”
ValidatorAI's homepage: the largest validator by reach (300K sessions) is entirely free and reports that under a third of users act after validation.
“Business Applications for December 2025, adjusted for seasonal variation, were 497,046”
US Census Bureau Business Formation Statistics, primary source for the monthly volume of new business applications, sizing the would-be founder pool.
“favorable, with 81% seeing odds of 7 out of IO or better and a remarkable 33% seeing odds of success”
Cooper, Woo and Dunkelberg (1988), Journal of Business Venturing, primary study of 2,994 new owners: 81% rated their odds at 7 in 10 or better and the poorly prepared were as optimistic as the well prepared. Supports the founder's doubt about wanting encouragement.
“The more telling causes — poor product-market fit (43%), bad timing (29%), and unsustainable unit economics (19%) — reveal why the capital dried up in the first place.”
CB Insights 2026 analysis of 431 VC-backed shutdowns: poor product-market fit cited in 43% of failures, the problem a pre-spend check targets.
“$10 / 1K searches”
Anthropic's pricing page: API web search at $10 per 1,000 searches and Opus 5.5 at $4/$20 per million input/output tokens, used to estimate a $3 to $6 cost per report; also shows Claude's free tier has web search but Research is Pro-only.
“Preuve AI leads with a fast, live-data verdict and a $29 one-time report, with optional monthly re-scans (per preuve.ai/pricing, September 2026); ValidatorAI leads with a free AI score and advice (as of September 2026).”
DimeADozen's own 2026 comparison page (interested source) confirming Preuve's $29 price, ValidatorAI's free model, and describing IdeaProof, WorthBuild and FounderPal.
“3,100+ paying customers across 3,200+ purchases — bootstrapped, no outside funding.”
DimeADozen's facts page: 134,000+ ideas submitted versus 3,100+ paying customers and 3,200+ purchases, implying roughly 2 to 3 percent free-to-paid and about one purchase per customer.
“Research findings confirm that individuals seek confirmation for their decisions and discount”
Von Bergen and Bressler, academic paper on confirmation bias in entrepreneurship: entrepreneurs' high optimism intensifies the tendency to ignore negative information.
“Only DimeADozen, IdeaProof, Verdikt, IdeaBrowser, Gaplyze and FindMeIdea also link any sources; the rest rely on AI-generated analysis or planning workflows without external data verification.”
Preuve's comparison of 21 validators (interested source): prices from free to $1,499/year, only six link sources, Preuve reports only 17.5% of ideas earn a go, IdeaBrowser charges $499+/year for pre-researched ideas.
“Total Entrepreneurial Activity (TEA) returned to the historic high of 19% reached two years earlier.”
Babson College summary of the GEM 2024-2025 US report: 19% of adults starting or running a new business, over two-thirds citing job scarcity as a motive.
“The product will stay online for one final year, but the business itself will close.”
GummySearch founder's shutdown notice, November 2025: a founder-research tool with 135,000 users closed because it could not get a commercial Reddit API license; platform-dependency risk.
“The new lightweight deep research, which will also come to free ChatGPT users starting today”
TechCrunch, April 24 2025: OpenAI rolled a lightweight cited deep-research tool out to free ChatGPT users; secondary source reporting OpenAI's announcement.
“Survival rates for new business establishments vary from year to year due to several factors, including the business cycle, industry, and location.”
US Bureau of Labor Statistics: primary data on 1-year survival of new establishments; the 2022 cohort ranged from 74.4% to 78.6% by census division, so roughly a fifth fail in year one.
“Limited deep research”
ChatGPT's current pricing page: the Free plan lists limited deep research and search, confirming the free substitute exists today.
“Most validation tools tell you what you want to hear. DontBuildThis gives you brutally honest feedback designed to save you time and money before you build the wrong thing.”
DontBuildThis's product page: a free tool positioned explicitly on scrutiny over encouragement, with paid tiers for recovery reports.
Search record: no published monthly search volume for 'validate my business idea' or 'validate business idea' was found, so top-of-funnel demand remains unmeasured.
Search record: looking for AI idea-validation tools that shut down surfaced none; the only closure found was the adjacent Reddit-research tool GummySearch, while about 20 validators are live in 2026.
“Since its launch in March of 2023, Dime A Dozen has been able to attain a total of 900+ paid users and a total of $32,000 in profit.”
Indie Hackers case study (secondary, citing an Acquire.com listing): DimeADozen made $58,000 revenue in six months via AI directories and SEO, was listed for $160,000; commenters called it a ChatGPT wrapper and noted the same could be done free in ChatGPT.
“The average customer retention was three months, meaning every month we lost 1/3 of our user base.”
Javelin founder Trevor Owens' post-mortem: QuickMVP, a pre-AI validation tool, was good at invalidating ideas but did not help users pivot, retention collapsed, and the company ran out of money.