I have had people tell me an idea was great and then watched absolutely nothing happen when I put it in front of the market.
That is one of those lessons that sounds obvious after you learn it. Before you learn it, compliments feel a hell of a lot like validation.
I have built websites, apps, plugins, communities, crypto projects and AI projects. I have had projects work. I have had one business do more than $200,000 in sales. I have also built things people seemed excited about that went nowhere.
The difference taught me something important. People being nice to you is not market research.
Stop asking if people like your idea
If I explain an idea to a friend and ask, “Do you think this could work?” I have already screwed up the test.
They know me. They can hear that I am excited. Most people do not want to be the asshole who crushes somebody else’s idea over lunch.
So you get answers like “Yeah, I’d use that” or “That could be huge.” Those answers cost the person nothing.
What matters is what people already do.
Ask when they last had the problem. Ask what they used to solve it. Ask what annoyed them. Ask what they paid. Ask what they tried before and why they stopped using it.
Past behavior gives you something to work with. Future promises are cheap.
I have confused attention with demand before
This one gets me because attention feels good.
I have had posts get thousands of views and barely move signups. I have built communities where people interacted but that did not automatically turn the project into a sustainable business. I have seen free users show interest that disappeared the second money entered the conversation.
None of that means the attention was worthless. It means I was asking it to prove something it had not proven.
A view proves somebody looked. A compliment proves somebody liked the idea enough to say something nice. A signup proves a little more. A payment proves something stronger.
Those signals should not all get the same weight.
Ask questions that are harder to fake
When I am trying to understand an idea now, I care more about questions like these.
What happened the last time you had this problem? What are you doing about it now? What does that solution cost? What is the part you hate? Have you tried another solution? Why did you quit using it?
Notice I still have not asked whether they like my idea.
I want to understand the problem before I sell them my answer.
Strangers can be more useful than friends
Friends have a relationship to protect. Strangers usually do not.
That is why Reddit threads, support forums, competitor reviews, Facebook groups, YouTube comments and niche communities can be useful. People complain differently when they are not talking to the person who built the product.
I like repeated complaints. One angry review can be somebody having a bad day. Twenty people complaining about the same limitation is something I want to understand.
That is also why competition does not scare me the way it used to. Competitors give you customers to study and complaints to read.
Do not let AI become another friend telling you what you want to hear
AI can make this problem worse.
Give a chatbot an idea and ask whether it is good. It can give you customer personas, market sizes, pricing ideas and a beautiful launch plan before you have proven anyone gives a shit.
I use AI constantly. I also built Marketur Reality Check because I wanted something aimed in the opposite direction. Instead of helping me get more excited, I want the idea challenged.
That does not replace talking to people. It helps expose assumptions worth testing with them.
Look for commitments instead of compliments
The closer somebody gets to giving up something valuable, the stronger the signal becomes.
An email address is stronger than a like. A booked call is stronger than an email. A deposit is stronger than a booked call. A purchase from a stranger is stronger than your buddy saying he would definitely buy it later.
You do not need every idea to start with a preorder. The right commitment depends on the business. A service can be tested with actual appointments. Software might start with a paid manual version. A content business can test whether readers return and eventually buy something.
The point is to keep moving toward behavior.
The feedback I want now is the feedback that can hurt
I wrote a book called How to Fail: I Got Really Fucking Good at It. A lot of what I learned came from spending too much time protecting ideas from information that might have killed them earlier.
Now I would rather find the ugly answer before I buy another domain, burn through AI builder credits, spend weeks coding, or discover some platform dependency after I am already committed.
If everybody loves your idea, make the test harder.
If you want a second opinion before you start building, run the idea through the free Reality Check. Then take the weak spots it finds and go ask real people about what they actually do.
That combination is a lot more useful than asking ten friends if they think your idea is cool.
For the bigger process, read how I validate a business idea before spending money. The U.S. Small Business Administration also has a useful overview of market research and competitive analysis.