Three engineers in a conference room. They’ve been arguing about the right approach for two weeks. Everyone has strong opinions. Nobody has written any code.

Both sides have valid points. That’s why the debate won’t end. You can’t reason your way to an answer when the truth depends on details nobody knows yet.

So stop arguing and run a pilot. Timebox it. Two days. Build the smallest possible version that tests the core assumption. Don’t build the whole thing. Just build enough to learn whether the idea holds up.

This works especially well for technology choices, architecture patterns, and workflow changes. Anything where the honest answer is “it depends” is a candidate for a quick experiment instead of a long meeting.

One team I worked with spent three weeks debating whether to use a message queue or direct API calls between services. I told them to spend two days building a proof of concept with the queue. Took one developer a day and a half. The results made the decision obvious. Debate over.

Arguments produce opinions. Experiments produce data. When you’re stuck, stop talking and start building.