Every few weeks, a scraper I wrote walks the open listings on Flippa, the marketplace where people sell websites, stores, apps, and content businesses.
The latest pass covered 4,433 unique listings. That dataset is usually read as a shopping list. I find it more useful as a build guide, because every listing is someone's answer to a deceptively hard question (what actually makes money online?), with a price attached.
The aggregate picture is informative on its own. The median business asks about 1.9x its claimed annual profit, roughly a two-year payback.
One in five sellers has already cut their price, by a median of 27%. About a quarter of listings claim net margins above 85%. And 12% of the listings that name their own website point at a domain that no longer resolves: a whole genre of small businesses whose listings outlived their sites.
But it's the clusters that matter if you're deciding what to create. Three patterns keep showing up among sellers who look like operators rather than escapists.
1. Vertical AI tools
Not foundation models, but small, single-purpose tools in one niche: an AI video tool for creators, a football-focused AI platform with 90K users, a podcast-to-content tool. Most run on standard stacks (Next.js plus model APIs), were built by one or two people, are less than two years old, and claim to grow mostly through organic search.
The interesting part is the asking prices: several verified listings ask the equivalent of six to twelve months of claimed profit. When tools this simple trade at those multiples, the message to a builder is not "buy this tool." It is "this buyer market exists, and the build is weeks of work, not years."
2. Skill education and content
Courses and niche content sites sell at the lowest multiples in the whole dataset, often 0.5x to 0.8x annual profit. The names are unglamorous: an Excel-focused eLearning business with 3,000+ students and 33 corporate clients; a single-word domain with strong authority doing about $4,500/month; marketing blogs and health affiliate sites with similar economics.
Two things follow. First, a thin moat is priced as a thin moat, and that is efficiency, not failure. Second, the format compounds: free tutorials feed an email list, the list buys a course, corporate clients buy licenses. Everything about it is copyable except the audience, and audiences can be earned in public.
3. Pet problem niches
The pets vertical is full of both the best and the worst outcomes in the data. On the good side: a dog nail-care store claiming roughly $2K/month that its owner says runs in 2-3 hours a week; a pet brand doing about $40K/month with 793 SKUs and a subscription option; small stores around one clever product (cooling vests, grooming) with 60-80% gross margins.
The distinction seems to be the angle: winners solve an exact-match problem (nail trimming, scratching, tear stains) with a product or guide specific enough to rank and repeat. Pet spend survives recessions; generic stores survive neither recessions nor algorithm updates.
The pattern behind the patterns
- Thin moats are the point. Standard stacks, no network effects, boring categories. That is exactly why these businesses sell for 6-12 month paybacks, and why one person can rebuild one.
- Siblings are the validation. No single listing proves much; a dozen independent sellers in the same niche, all closing at 0.5x-1x annual profit, means both the demand and the exit are real.
- Recurring beats novel. Consumables, subscriptions, and memberships show up in the healthiest listings. Trend stores show up in the graveyard.
- Assume the numbers are inflated. "Data verified" means the platform checked the connection; it is not an audit. Read the multiple as marketing, the model as signal.
If you're hunting for something to build
Shop the sell-side first. A marketplace of 4,000+ businesses is a census of models that have customers. Study the clusters, find the ones with recurring revenue and thin moats, and check how many independent sellers there are. If several are closing at a payback you could beat with a weekend of work and a few weeks of shipping, that is your idea budget: the market has already paid for the research.
The best answer to "what should I build?" is often already being sold. Someone else found the customers; you get to study exactly how.
No comments yet