You might be wondering:
- Where do I research good ideas?
- What do I even ask about?
- Should I work on something I am passionate or interested in?
- I heard to find faults in other projects and bad reviews, should I do that? where do I even do that?
This an actionable list that will help reframe and guide you to figure out exactly this. I promise you it’s not another vague article that leaves you the same place you started. Wondering what to even do next.
Let’s start with what NOT to do.
1. Stop searching Product Hunt and other similar sites
You might be tempted on finding “What works” and go to product hunt or indie hackers but the crowd here is mostly developers, so their upvotes show what other builders love, not what a normal person pays for.
Launches by category over 100 days. The pill on each bar is how much of it is AI.
You might be tempted to think the short bars are sectors product maker crowd ignores and underserved. I should be working on those areas such as travel it only had 15 launches in 100 days, e-commerce and support.
An empty on Product Hunt does not mean empty in real life.
I fact-checked the popular “these niches are wide open” lists, and most claims fell apart. A gap is a reason to go ask real people and is telling you people aren’t launching on product hunt on those areas. It doesn’t mean it’s underserved.
2.”It’s got AI” is not an idea
Each dot is one product launched in 100 days. Orange is AI, and two out of three are.
Two out of three launches are AI. Saying “ours uses AI” now tells a customer nothing. It is the default.
It is worse than you think, actually. In the revenue data, AI first apps make about 40% more money per user, but those users also quit about 30% faster.
You should treat AI as the engine not as differentiator.
3. Stop finding your passion or what you’re good at
Nobody in the world gives a f*ck what you are passionate about or what excites you. People need services or products that solves a problem for them and they are willing to pay for that. Your passion is or should be not being broke and making money. That’s it.
4. Know whatever idea you pick will suck at first
Absolutely no one I worked with and including me knew what they where building when they started doing it. Everyone figures out or pivots as they go so whatever you pick will suck at first and is simply starting point to figure out what the customers actually want. So stop wasting time researching so much depth and asking all sort of different questions like. How much is ROI and how will it scale in revenue or how do I save cost. If you haven’t had single customer paying for whatever you’re idea is stop worrying about your revenue growth strategy.
5. Distribution is always the hardest part
In fact the biggest leverage you can do right now is start learning how to make content and build a habit of constantly posting content and learning about SEO and making hooks and getting attention to whatever you want to say. Then and only then you should go back to find out what to build. But if you already are doing this continue reading further to figure out what to build for your next side project.
Now the part you actually came for.
There is an essay every founder should read, Paul Graham’s How to Get Startup Ideas. His one big point: you do not think up a good idea by brainstorming, you notice a problem you already have. Below is that idea turned into five practical moves, each with a real story and something concrete to go do. Notice the pattern in every one: software built for a gritty, real-world business, not another app for people who already make apps.
1. Stop trying to think of ideas. Notice the problem you already have.
The worst way to find an idea is to sit down and try to invent one. That is how you get something that sounds clever but nobody wants. Graham’s word for the real move is not “think up,” it is “notice.”
Luke Hansen grew up roofing houses in his family’s business. On every job they took photos, and they kept losing track of which photo belonged to which house. He went looking for an app that could organize job-site photos by address, and there wasn’t one. So they built it in a back room of the roofing company. That is CompanyCam, an app contractors use to snap, organize and share photos from every job site. He was a roofer, not a software guy. He just could not ignore his own problem.
Do this: this week, spend two real hours inside a hands-on business you already have access to, a family shop, a friend’s trade, your day job. Watch for the clipboard, the whiteboard, the messy group chat, the pile of photos on someone’s phone. Write down the one thing they redo by hand every single day.
2. Being on the inside is your unfair advantage
This one is not about where the idea comes from, it is about why you can win it. When you already belong to a trade or a group, you get things a funded outsider never can: you speak the language, you know who to call, and those people already trust you.
Amar Ghose was not a programmer. But he owned and ran a maid-cleaning company, which meant he could walk up to other cleaning-business owners, talk to them as a peer, and get them to trust his tool in a way no outsider could. The scheduling tools he built for himself became ZenMaid, software other cleaning companies use to book jobs, route their cleaners and get paid. A slick engineer from a tech hub could copy that software and still never crack that world. He bootstrapped it to millions a year.
Do this: name a group you already belong to and are trusted by, your coworkers, your trade, your team, your corner of the internet. That trust is a door outsiders cannot open, so build for that group first and you start with an audience that already listens to you. And if no group trusts you yet, that is your real first task: pick one community, a subreddit, a Discord, a local meetup, show up every day for a month, help people and make actual friends before you ever pitch a thing. Trust is not found, it is built.
3. Run toward the boring, unsexy work everyone else avoids
Most people want to build something clean and clever. The easy ideas are already picked over. The messy, boring jobs stay wide open, precisely because nobody wants to touch them.
In construction, the ugliest problem is not building things, it is getting paid for them. Scott Wolfe was a construction lawyer who watched contractors get stiffed for months, buried in liens and paperwork. Chasing money is about the most dreaded corner of a hard industry, which is exactly why almost no one built for it. He did. Levelset is software that helps contractors and suppliers handle the paperwork and actually get paid. It has been used on millions of construction projects and sold for around half a billion dollars.
Do this: find the part of a real business that is about money or paperwork, getting paid, invoicing, permits, licenses, tax. Message three people who run that kind of business and ask one question: “what is the most painful part of getting paid?” Build for the answer they all repeat.
4. A crowded market is a green light, not a warning
Earlier I told you not to chase the crowded categories on Product Hunt. Here is the nuance, straight from Graham: “A crowded market is actually a good sign, because it means both that there’s demand and that none of the existing solutions are good enough.” The mistake is not entering a crowded market. It is copying it with no edge. If the people already there quietly hate what they are forced to use, that is your opening.
Ilir Sela grew up in a family of pizzeria owners. He watched independent pizza shops get crushed by Domino’s, which had slick apps and online ordering the little guys could never build for themselves. Online ordering was a crowded, giant-owned market. He did not avoid it, he armed the underdogs. Slice gives an independent pizzeria its own online ordering, website and app so it can go head to head with the chains. It now works with around 20,000 shops.
Do this: pick an industry being squeezed by one dominant giant, the big chain, the big platform. Talk to three small operators in it and ask what the giant does that they simply cannot match. Build the tool that lets the little guys do exactly that one thing.
5. Start stupidly narrow, for a few people who are desperate
Do not try to build something a lot of people want a little. Build something a small number of people want a lot. Ten operators who are desperate for it beat ten thousand who think it is neat.
A software engineer named Ronny Hash teamed up with one pool-cleaning guy who was drowning in paper route sheets. He rode along on the routes, watched the actual job, and built for that single narrow trade and nothing else. That is Skimmer, an app pool-service pros use to track their routes, log every pool and bill their customers. “Software for pool cleaners” sounds far too small to matter. More than 30,000 pool pros now run their whole day on it.
Do this: pick one specific trade, not “small businesses” but “pool cleaners” or “dog groomers” or “mobile mechanics.” You do not need to be one of them or hold any qualification. Ronny was not a pool cleaner, he just asked to ride along. Find one operator through a local search, a trade Facebook group or a subreddit, offer to shadow them free for a day, and watch the single worst part of their day. Then build the ugliest tool that fixes just that, for them and nobody else yet.
None of these people were geniuses or well funded. One of them was a lawyer, sure. But the others were a roofer, a house cleaner, a kid from a pizza family, and a coder who had never touched a pool in his life. No degree opened those doors. Being inside a hard, unglamorous world, or just showing up in one, did.
Stop trying to find good ideas on the internet. The answer to “but I am not in any of these worlds” is simple: get into one. This weekend, pick a trade and do one of three things. Ask one person who does it if you can shadow them for a day. Take a weekend or holiday job in it. Or join the Facebook group and the subreddit where they complain, and read for an hour. You are not trying to become a roofer or a lawyer, you are trying to watch what they hate doing. That takes zero qualification, only showing up.
Or, as Graham puts it at the end of his essay: “Live in the future and build what seems interesting. Strange as it sounds, that’s the real recipe.”
Pick one of the five, go look this week