I Checked 50 Product Ideas. Here Is What Survived
Real complaints, real sources, and a look at the cheap tools that already exist.
Everyone wants a list of product ideas. I wanted one that held up. So I collected 50 ideas from threads where people complained about their software, paperwork or tools. I checked each one before I wrote a word about it.
None came out clean. Eighteen stayed on the maybe list after a second pass over the sources, and 32 did not survive. That is a short list, and I think it is more useful than a long one.
Below, I show how I checked each idea and the eight I would test. I also show one idea I ran through a full validation, and the ten I cut. For the ones I cut, I name the free or cheap tool that already does the job.
How I checked each one
The ideas came from five areas. They were general wish-list threads, local and trade businesses, freelancers and creators, back-office work for small teams, and solo builders who sell small digital products. I read threads on Hacker News and dev.to, plus vendor pages and review sites. For each idea I asked four questions.
Is there a real complaint I can link to, with a date and a size? Does a second, independent source say the same thing? Does a free tool, or one under $50 a month, already cover the core of the idea? And what is one cheap action that would show within a week if anyone cares?
A complaint counted only if I could read it. Reddit, Quora and Indie Hackers blocked my reads, so none of them is evidence in this piece. A second pass re-read the sources behind every idea that stayed on the maybe list, and it changed a lot. It found wrong comment counts, a misquoted phrase, and a tool whose website now redirects to an unrelated site. Prices below were checked on November 3, 2026, and prices change.
What the 50 had in common
The threads were small. Most had fewer than 30 points. A few ran higher, such as 41 points for a thread on decision records and 112 for a price-hike story. A loud thread is not the same as a buyer.
In the local business reviews I read, the complaints were about billing, support and cancelling, not about missing features. People were not asking for a new app. They complained about fees and about how hard it was to leave.
Posters often solved the problem themselves. One builder of a bank-statement converter wrote that the business “hovered around break even.” He shut it down because he was grinding out a new integration for each bank’s PDF format. And for most ideas, a free or cheap tool was already there.
Three worth testing first
1. An exit kit for owners leaving a point-of-sale or booking tool.
On October 29, 2026, the FTC, seven state attorneys general and two California local government offices sued Toast, Inc. and four affiliates. They filed in federal court in San Francisco’s district. The complaint alleges hidden fees and hard-to-cancel subscriptions, including an early termination charge of $150 for each remaining month. It is allegations only, and nothing has been decided.
Trustpilot reviews of a field-service app named Jobber point the same way, though they are a minority view. Of 20 one and two star reviews from March to October 2026, at least six were about leaving. One cited data lock-in, and two said they were charged after cancelling. Jobber also shows 234 reviews and a 4.2 score overall.
What people do today is lean on the next vendor to help move the data in. A booking vendor’s pricing page I read promises a one to two day migration in, and it says nothing about exporting out. The gap is knowledge, not software. A one-page “before you give notice” checklist for one product could cover what to export and in what order. It could also say what to ask the next vendor and which charges to expect.
One-week test: write the checklist for one product, offer it free to five owners who complained about it, and ask what they would pay. My doubt is that I found no price evidence, and the next vendor’s onboarding team may do this for free.
2. A “what do I paywall” teardown.
A Hacker News thread from October 2025 drew 13 points and 26 comments. A solo developer of a habit tracker said users told him they would cry if the app disappeared, and almost none paid. He reported 35,238 installs over three years and $86 a month from 20 subscribers, about 0.8 percent.
The replies already offered fixes. One suggested limiting the free core, the way a paid search engine caps free searches. Advice is free and plentiful. What is missing is someone who looks at your own free and paid lines and tells you where to draw the boundary.
One-week test: post one free teardown of a real app in a founder community, and count how many people ask for a second. My doubt is that nobody in the thread asked to pay for this.
3. A decision log that gets filled in.
A March 2026 thread asked how to capture why decisions were made, not just what was decided. It drew 41 points and 64 comments. The thread’s author wrote, “Nobody wants to document themselves out of a job.” A second thread from November 2025, at 16 points, had a commenter who wanted every leader’s decisions tracked over time. A reply said leaders would not buy a product that puts them under a microscope.
The thread’s author says they are building a tool that drafts the reasoning from pull requests, Slack and tickets. I did not price a substitute here. The problem looks like behavior, not a missing template, which is also my doubt.
One-week test: take one real thread from a small team you know and turn it into a one-page decision record. Then ask if they would run it every week.
Five more, with the doubt stated
These five have thinner evidence. Treat each as a test, not a build.
A handover pack for solo software vendors. In the same November 2025 thread, a commenter said none of the companies they worked for would trust a one-person software shop. The vendor was made to offer a self-hosted version and to escrow the code. A code escrow service I checked lists $139 a month billed yearly, or $179 monthly, plus a $249 setup fee. That comes to about $1,900 in the first year. A long-time solo owner in the thread said customers rarely ask. The pack would be a runbook, a restore test and a handover checklist. Doubt: one thread, and mixed.
An AI-era project pricing worksheet. Four threads from March to October 2026 ask how to price a project when AI cuts the hours. The largest, “Is $300/HR too low these days for custom full stack?”, had 21 points and 16 comments. Replies suggested a fixed fee with normal AI use included, or pricing on the outcome. Nobody asked for a tool. Doubt: the answer is judgment, and a worksheet may not add much.
Get Eddie Larsen’s stories in your inbox
Join Medium for free to get updates from this writer.
A 30-day channel test log. A June 2026 thread, at 15 points and 25 comments, came from a founder with no audience. They had tried directories, building in public, Reddit, articles, cold outreach and micro-influencers. Reddit banned them, the articles stayed unindexed for a month, and the outreach got no replies. A spreadsheet can hold a test log. The real need is judgment about what to try next, so a template alone will not sell.
A newsletter sponsorship rate card. An August 2025 post came from the owner of an 8,000-member newsletter with no sponsor yet. It drew 3 points and 2 comments. A free brand-deal workbook already exists. Reports say Substack is adding sponsorship tools, but I could not read the source, so check that before you build.
A backup and exit kit for digital product stores. I sell on Gumroad, so I have a stake in how this one reads. A March 2026 post was titled “Gumroad suspended my account, blocked my payout, and stopped responding” and drew 6 points and 2 comments. A solo developer wrote in June 2026 that he got rejected during review with no explanation and went with Polar. The kit would list what to export before anything freezes: customer emails, files, receipts and tax settings. Doubt: moving platforms is free, the threads are small, and the idea rests on single stories.
One idea through the full check
The loudest complaint in my 50 was about a price hike. In August 2026, a Hacker News thread about a time-tracking tool whose prices jumped after an acquisition drew 112 points. One commenter posted a renewal email that put them on a $2,199.50 monthly plan. They said they had paid about $100 a month for 13 years and closed the account. Another commenter argued the old price had been very low.
The idea I tested was a small tool for owners and office managers of firms with 2 to 30 people. They forward vendor emails to one address. The tool pulls out the vendor, the renewal date and the new price, shows one list of what renews, and sends a reminder weeks ahead. I priced it at $9 to $19 a month, or $29 once.
I ran the Idea Validation skill on it. It used seven web searches and read search summaries, not full pages. It could not read Reddit or Google Trends, and I gave it no background line, so it marked founder fit as insufficient data.
Problem: medium. A poster in a QuickBooks community thread from April 2025 says a renewal went from $320 a year to $378 plus tax, with no word about the increase. A QuickBooks team member replied that a reminder email is sent and that the $320 may have been a promo price.
Demand: medium. A March 2026 survey of 781 small business owners, from a trade expo’s research desk, says 41 percent saw higher software costs. It also says 62 percent run no more than two paid subscriptions. It states no method, so I treat it as a rough signal.
Competition: low. TechStackTrack takes forwarded invoices and pulls out the vendor, price and renewal date. TrackMySubs, KeepMySubs and SpendHound cover parts of the same job. SpendHound offers renewal alerts free to companies under 1,000 employees. Free renewal-tracker spreadsheets also exist.
Monetization: low. TrackMySubs runs free, $10 and $30 a month, and TechStackTrack’s plans are 0, 39 and 79 euros a month. A price of $9 to $19 has no clear edge in that range. Go-to-market and founder fit: insufficient data.
The skill’s overall read was a weak signal. My first pass had rated this idea a maybe, so the check moved it down. It also gave three next steps. Read ten low-star reviews of the existing tools for a repeated complaint. Ask ten office managers what a missed renewal cost them. Find one community of office managers and count its members. That is the check doing its job. It cost a few searches, and it could save weeks of building.
What I cut, and the tool that already does it
These are ten ideas I dropped. Each has a free or cheap tool that covers the core, or a product that already exists.
Two browser extensions already sell search and bulk export for AI chat history, at about $5 to $10 a month. A $9 revision-round tracker for web projects already exists on Gumroad. Stripe’s customer portal has a cancellation-reason setting that is on by default. Cal.com’s free plan lists Stripe and PayPal payments, which covers a paid booking page for consultants. Plausible starts at $9 a month for privacy-friendly analytics, and Umami can be self-hosted for free.
Cloudflare Turnstile’s free plan covers a newsletter signup bot filter, with unlimited challenges. Toggl Track has a free plan for freelancers who want time capture. One renewal tracker is free for five contracts, and another costs $49 a month for 150 items, which covers a contract notice-deadline calendar. Joist’s Elite plan is $32 a month and handles change orders alongside estimates and invoices. BankConv gives the first bank statement conversion free, then charges $2.99 a document.
Each check is short. Each build would have taken weeks and launched into a market that already had a free answer.
What this says about finding ideas
The gap is rarely a missing app. In my 50, the gap was usually knowledge, setup, review or a narrow audience. The exit kit is knowledge about what to export and when. The paywall teardown is a second pair of eyes. The handover pack is a checklist and a restore test.
That pattern suggests a smaller kind of product: a guide, a checklist or a template aimed at one narrow audience. Several of my cut ideas died because a $9 template or a free workbook already existed. That should not stop you. It should tell you to price and package against what is already out there.
Run the check on your own idea
You can run the same check on an idea of your own in an afternoon.
- Find the complaint. Link it, date it and note its size. If you cannot read it, it does not count.
- Find a second, independent source. One thread is a story, and two is a pattern.
- Search for the cheap substitute. Read the vendor’s own pricing page and write down the date.
- Pick one cheap action that tests demand within a week. Do it before you build anything.
The skill I used for the full check is in the P.S. below. The one-week test is yours, and no skill can run it for you.
Special thanks to Claude for helping with the research on this list.
Thanks for reading, now go build something fun.
Eddie
P.S. If you want to test an idea before you spend money on it, I built a Claude Code skill called Idea Validation ($39). You describe the idea and run /validate. The skill searches the web for complaints, competitors, pricing and reviews, and tells you where it found nothing.
