Two questions I ask before I pick a problem worth solving.
This site has four postmortems on it: DearMeWiser, HardChoice, CompliQuiz and CampfireX. Each one ends with a lesson. Read side by side, three of the four come back to the same thing: the problem was real enough to build for, but nobody was already paying to get rid of it.
This week I posted two questions on X that I now ask before I pick the next idea. Here they are, run against my own shut-down products, because a question only earns a place in your process if it would have caught a real mistake.
Question one: has anyone paid to make this go away?
Here is how I put it on X on November 17:
has anyone ever paid money, or hours, to make this problem go away?
if not, the pain might be real but it isn't urgent. and urgent is what people pay for.
Three of the four postmortems fail it, each in a slightly different way.
CompliQuiz told founders which compliance frameworks applied to them. In the postmortem I wrote that any founder who needs that answer "can figure it out in an afternoon." The money and the lost nights were one step further down the workflow: mapping controls, closing gaps, producing evidence, surviving the audit. The question would have passed for that step. It failed for the step I built. So the first lesson is to ask it about the exact step your product does, not about the whole space.
DearMeWiser answered your questions as an older, wiser version of yourself. The honest answer to question one was that people already had a fix that cost them nothing: one well-written prompt in ChatGPT. When the existing fix is free and already open in another tab, the answer is no, however real the feeling behind the problem is.
CampfireX got 200 clicks and zero customers in eight days. Pre-pay validation was the part that told the truth. People would read, nod, and not pull out a card. That is the same sentence as the X post, from the other side: the interest was real, the urgency was not.
Question two: do I know this problem better than most people trying to solve it?
This one is from this morning:
do i understand this problem better than most people trying to solve it?
two decades in automation and security left me knowing a lot of boring pain. maybe boring pain i know is worth more than exciting pain i'm guessing at.
The clearest yes I have had came out of a failure. While building CampfireX, almost every business owner I talked to asked some version of the same thing: how do I use AI without breaking what already works? That is a question I have spent twenty years preparing to answer, and it is where I went next, with my AI advisory practice and later The Boring AI Runbook.
NicheSlap passes question two for a different reason: I am the user. I changed my niche on X three times in seven months and did not notice until I built something that counted my posts. Question one is the part it still has to prove. The free scan tests whether people will spend their time on it, and the paid plan tests whether they will spend money. It is a few days old, so it is too early to call, and I would rather say so than pretend.
Two questions are not the whole test
HardChoice is the honest exception. It was a structured decision tool for couples, it worked, and the people who used it got value. I shut it down because of who it was for: a consumer, emotional product that needed an audience or an ad budget I did not have, sitting next to products built for founders with day jobs. Both questions can come back yes and the product can still be wrong for you. They are a filter for the problem. They do not pick your audience.
How to run them on your idea tomorrow
- Write the idea as one line about one step. Not "compliance for startups." The step: "tell a founder which frameworks apply." CompliQuiz would have failed right here.
- Name who paid, and with what. A real person or company, this year, who paid money or hours to make this go away. A tool they bought, a contractor, a spreadsheet someone maintains every week, a chore that eats Friday afternoon. If you cannot name one, you have an awareness problem, and those tend to be solved for free.
- Check the free fix. Could a smart user do it with one prompt, a template or a search? If yes, ship that as a free tool and look one step further down the workflow for the paid problem.
- Write three things you know that an outsider would not. About the people, the workaround they use today, why the obvious fix fails. If you cannot fill three lines, you are guessing at the pain.
- If both answers are yes, commit, with a date. Pick a fixed term and write down the result that would end it. Then hold to it in the dip. The method I use for that is the stick-or-quit protocol on the NicheSlap blog.
If either answer is no, do not build yet. A no before the first line of code costs an evening. A no after launch costs a postmortem.
If you are earlier, or later, than this
If you do not have a problem to test yet, start one step back. I built IkigaiNiche as a free guided way to find a person you would be glad to help and a problem of theirs worth solving.
If you already have one and you post about it on X, the next risk is drifting off it. I wrote about how to spot that in your own posts in shiny object syndrome on X.
After you pick: small bets only work if you stay long enough to find out
Related postmortems: CompliQuiz · CampfireX · DearMeWiser · HardChoice
Follow the build: @chams_builds on X