// How to choose an app idea worth building
Most people pick an app idea the wrong way round: they think of something clever, build it, then go looking for people who want it. The reliable method runs the other direction. Find demand first, then build the thing that answers it.
Here is the process, and how this site does the first half of it for you.
1. Start from what people are already searching for
Every month, people type millions of searches looking for software that does not exist yet: keg tracking software, therapist waitlist management, digital carrier packets. Those searches are demand with a receipt. Someone sat down, felt a problem, and went looking for a product to buy. If you build what they were searching for, you skip the hardest part of launching, which is convincing anyone the problem matters.
You can find these yourself with keyword tools, or start from the searches already hitting this site — they are listed on the homepage, written exactly as people typed them.
2. Check three things before you commit
How many people want it. Monthly search volume for the product, not the topic. Hundreds is plenty for a niche business; you do not need thousands.
How crowded it already is. If a dozen funded companies own the space, you need a wedge: one vertical or audience the incumbents treat generically. An empty field means focus alone can win.
What one customer is worth. This is the number beginners ignore and it decides everything. Look at what advertisers pay per click for the search term. When a company pays $30 for a single click, the customer behind that click is worth thousands, which means you need dozens of customers, not thousands.
Every rated idea on this site carries these three signals, explained on the market signals page.
3. The quiet-but-valuable sweet spot
The best solo-builder markets look unattractive at a glance: low search volume, boring industry, no hype. That is exactly why they are available. Fifty searches a month for software that costs $200 a month beats fifty thousand searches for a free consumer app. Big companies ignore small markets; you should not.
4. Validate before you write code
Talk to five people in the audience and ask what they do about the problem today. If they have a painful workaround — a spreadsheet, a WhatsApp group, a folder of PDFs — you have a customer. If they shrug, walk away. Then put up a one-page site with a real price and a waitlist button. A weekend of validation saves a month of building.
5. Scope it to weeks, not months
Cut the idea until it fits a weekend or two. One audience, one workflow, one screen that matters. Ship it, charge for it, then let the people paying you decide what comes next. Every idea here is labelled with a realistic build time and comes with a paste-ready prompt for Cursor, Claude Code, Lovable, or Bolt, so the blank editor is not the thing that stops you.
→ Browse ideas rated on real demand
Common questions
How do I decide what app to build?
Work backwards from demand instead of forwards from inspiration. Find a group of people already searching for a solution, check whether anyone serves them properly, and check whether each customer is worth real money. If all three line up, build that. If you start from an idea you like and hunt for users afterwards, you are guessing.
How do I validate an app idea before building it?
Three cheap tests, in order. One: search for the product as a buyer would and see whether people are looking for it and what advertisers pay per click. Two: talk to five people in that audience and ask what they do about the problem today; if the answer is nothing, they will not pay. Three: put up a page with a price and a waitlist button. Only then write code.
How do I know if an app idea is good?
A good idea has a named audience, an existing painful workaround, evidence people spend money in the space, and a version small enough to ship in weeks. A bad idea is described with the word everyone, has no current workaround to replace, and needs six months before anyone can use it.
Is low search volume always bad for an app idea?
No, and this is where most people go wrong. Fifty searches a month for software that costs $200 a month is a better business than fifty thousand searches for something nobody pays for. Look at what advertisers pay per click: when companies pay $30 for one click, each customer is worth thousands, and you only need dozens of them.
Should I build something with lots of competition?
Only with a wedge. Competition proves the market pays, which is useful, but you cannot beat a funded incumbent at their own game as a solo builder. The reliable move is to take one specific vertical or audience the incumbent treats generically and serve it properly.