I have built enough things backward to know how this usually goes.
I get an idea. I like it. I can already picture the website, the app, the plugin, the pricing page and probably the damn logo. Then I start building before I have a customer.
That does not automatically make the idea bad. But it does mean I need a way to learn something before customers exist.
The good news is you can validate a surprising amount of a business idea without customers. You just have to stop treating validation like a vote on whether your idea sounds cool.
You are looking for evidence, not permission
Friends saying they like an idea is not validation. An AI assistant producing a beautiful market analysis is not validation either. I know because I have had plenty of ideas that sounded great in a conversation and did absolutely nothing once they met the real world.
What I want now is evidence that the problem already exists without me creating the story around it.
Are people searching for the problem? Are they complaining about existing solutions? Are companies already charging money to solve some version of it? Are people cobbling together ugly workarounds because the current options suck?
Those signals existed before your idea did. That is what makes them useful.
Start with the problem people are already describing
Search Google, Reddit, YouTube, forums, app reviews and product reviews for the problem rather than the product you want to build.
I care a lot more about somebody writing a long angry post about a problem than somebody telling me my solution sounds awesome.
The angry person has already demonstrated pain. Now I need to understand what they tried, what failed and whether enough people have the same problem.
Competition is useful evidence
I used to look at competition as something that could ruin an idea. These days I would rather find competitors than find nothing.
Competitors tell me somebody has already done the painful work of proving people spend money in the category. Then I can look at pricing, reviews, complaints, features people hate and customers they serve badly.
If you find no competitors at all, do not celebrate yet. That can mean you found an untouched opportunity. It can also mean nobody cares enough to pay for the problem.
That is why I now treat competition as evidence instead of an automatic reason to quit.
Look at what people already pay for
This is one of the biggest changes in how I think about ideas.
I have had traffic. I have had users. I have had communities. I have had projects people thought were interesting. None of those things automatically created a business.
Money changes the quality of the signal.
Before you have your own customers, look for evidence that people already pay for competitors, freelancers, subscriptions, software, services or even manual workarounds that solve the same underlying problem.
If money is already moving, you have something worth investigating.
Use AI as a researcher, not a cheerleader
AI makes this process faster, but it can also make bad ideas sound incredibly intelligent.
I use ChatGPT and Claude constantly. I also know how easy it is to describe an idea and get back a polished explanation of why it could work.
That is useful for brainstorming. It is terrible as proof.
Ask AI to attack the assumptions instead. Ask what substitutes exist, what dependencies could kill the business, why customers would refuse to switch and what evidence would disprove your theory. Then verify the important claims outside the chat.
This is why I built Reality Check
I wanted something between having an idea and spending a bunch of money building it.
The free Marketur Reality Check is designed to pressure-test an idea instead of simply telling you it sounds promising. It looks at demand, competition, dependencies and weak assumptions so you have something concrete to investigate.
It is not a magic yes or no machine. Nothing is. But it gives you a much better starting point than excitement.
Eventually you still need humans
Desk research can tell you whether the problem exists. It can tell you who competes, what customers complain about and where money already moves.
What it cannot ultimately prove is that somebody will buy your version.
Eventually you need conversations, clicks, deposits, preorders, purchases or some other action from actual people.
But there is no reason to walk into those conversations blind. Do the research first. Learn the language. Find the assumptions. Then ask better questions.
The lesson I wish I followed more often
I love building. That is part of the problem.
Building feels productive. Research can feel like standing still. Then six months later you discover the research would have been cheaper.
So if you have no customers yet, do not use that as an excuse to skip validation. There is plenty you can learn before the first person ever pays you.
Start with evidence. Run the idea through the Reality Check. Read my full guide on how to validate a business idea. Then build when the evidence gives you a reason to build.