A good decision has three parts. Most of us get two.

Most bad decisions I've made weren't wrong. They were incomplete. A good decision has to score on three things at once: quality (you actually did the analysis and weighed real options), executability (the people who carry it out are bought in), and timeliness (it lands neither too early nor too late). This is the frame Management Essentials at HBS gave me, and it named something I'd felt but couldn't articulate. Miss any one of the three, and the decision tends to fail, however good the other two were.

Long ago, as a staff engineer, I sat in a meeting to pick an accounting framework for the whole company. Two options: adopt an open-source one, or build our own. After some back-and-forth that, looking back, had almost no research behind it, our lead called it: we'd build. And we missed all three legs at once. Quality was thin, the analysis just wasn't there. Buy-in was poor, it was one person's call, and the rest of us never really nodded. We ended up writing our own system by copying the open-source one feature for feature. As for timeliness: it never went live. I'm not even sure the codebase still exists anywhere in our modern repos.

That's the thing about a three-legged stool. You don't need to saw off all three to fall over. But miss all three and you don't just fall, you build something nobody asked for, that nobody owns, that arrives never.