Last night I killed a CRM. My own. The one I had been building all evening.
My autumn plans are becoming more and more interesting. So much so, that I feel like my trusty Google Sheet no longer keeps me on top of all the details. So I thought a CRM was what I needed. So I prototyped one on Claude Design, ready to build. And as I did, I realized it was beautiful and utterly useless for me.
What was on the screen was something more resembling relationship management. How I am doing with that customer in general. But my problem was somewhere else. It wasn’t whether or what my relationship looked like with that customer, but rather which kinds of agreements we had made. It was meant to help me keep on top of all the projects coming and all the intricacies there. What are the next steps in each of the projects. And keeping me from overbooking myself.
So it’s more a work management tool rather than a relationship management. I am after all a product person - I need to manage information in order to see the patterns and figure out the best possible value.
And the thing is, this is what I literally do for a living. I help teams learn what to build and what not to build. I help teams understand what the core problem is that should be solved with software. And here I am thinking not problem first, but solution first.
But maybe that’s okay. As long as prototyping the solution helps me understand what the problem actually is.
Because I’ve noticed how discovery works so much more intuitively for me now. I just build the wrong thing to understand why it is wrong. And getting to the right thing is now just so much faster.
This is why I love the fact that vibe coding has become a sort of gym for me. Not for shipping. For training judgment. Understanding problems better, recognizing solutions that don’t quite hit the mark, trashing them faster. Even the ones I fall in love with. Especially those.
The thing is, the product management cycle is so long that the muscle of judgment almost doesn’t get worked out enough. You don’t get a flood of fully formed ideas, or suggestions that are past the idea level. You have to triage them knowing very little, and it’s always imperfect.
And it was impossible before, because it took so much time to prototype and to really map out the idea visually, so it didn’t always make sense.
But now you can actually put yourself into that situation. Where you have to apply judgment to more formed ideas. By building the wrong thing and then recognizing that it is wrong. And because you haven’t spent so much time building it, you’re also not so in love with it that you’d find it impossible to let go.
Ideas are a dime a dozen. Now you can visualize them, or build them, to an extent that helps you recognize good from bad.
The tool nobody will ever see might be the best training a product person gets right now.