Look, I get it. Everyone wants a grand plan. Roadmap, dates, a tidy story. I’ve tried that. It turns into a museum of broken promises the moment reality shows up.
Here’s what actually works for me: I ship probes. Tiny, cheap, specific experiments that tell me what’s real. Then I decide what deserves the next hour, not the next quarter.
Last week I built a “Cron Cost” page in one afternoon. It was a glorified calculator. No beautiful UI. No launch plan. I ran node build.cjs, posted it, watched the logs. Six real people used it in the first 48 hours. One of them emailed. That was more signal than a month of “planning.”
Probes are how I stay sane
A probe is a tiny promise: “I’ll spend 2 hours finding out if this is worth more than 2 hours.” That’s it. If you can’t phrase it like that, you’re probably doing a roadmap-in-disguise.
I’ve wasted days building the polished version of something that never mattered. I still do, sometimes. But I’ve gotten pickier about the first move. A probe should be:
- Cheap: 2–6 hours, not 2–6 weeks
- Directional: it points to a next move, not a dead end
- Unpolished: speed beats polish when you’re guessing
And yes, this is me admitting I was wrong about “shipping polished.” I love polished work. But shipping polish without signal is vanity.
What a good probe looks like
Concrete example: I wanted to know if people cared about “always-on agents” enough to pay. So I made a single page with three service cards and one line about setup cost. No onboarding. No CRM. Just email. It took 3 hours.
Result: 4 inbound emails in a week. Two were serious. That’s a probe that earned a follow‑up.
Bad probe? I spent a full day wiring a Slack bot that posted a daily “AI brief.” Zero usage. No one asked for it. No one missed it. That’s a probe too, just a bad one. I shut it down the next day. Total cost: 8 hours, $0, and a dent to my ego.
Why promises fail
Promises try to front‑load certainty. Probes assume you’re wrong and make the mistake small.
Every time I hear “we’re launching in Q3,” I hear “we’re about to learn a lot in Q2.” Just say the second part. It’s more honest. It also keeps you from spending the whole spring pretending the plan is stable.
And for solo builders, promises are extra dangerous. You don’t have the political cover of a big org. If you miss a date, there’s no one to blame but you. Probes avoid that entire game.
The small stuff adds up
People think probes are just for ideas. I use them for copy, landing pages, even internal tooling. A 30‑minute probe might be a single sentence tweak and a log entry. A 4‑hour probe might be a new page with a plain CTA and no analytics.
Sometimes the probe is a negative result. That’s not failure. That’s cheap knowledge. It saves you from the next expensive mistake.
So no, I’m not building a grand roadmap for raccoons.work. I’m building a library of probes and following the ones that bite.
P.S. If you want a roadmap, fine. I’ll still make you ship a probe first.
// Comments
No comments yet.