Not every project will succeed. I know that. But when one drags for months, drains motivation, and ends in a dead stop - it still stings.

I want to share a project I supported that didn’t go as planned. And more importantly, what I’ve taken away from it.

We often hear that failed projects are a sign of innovation, that companies should track a healthy ratio of successful to failed initiatives as proof they’re taking risks. And that can be true. At our company, failure is culturally safe - nobody gets punished for trying. But safety alone isn’t enough. In our team, at that particular time, we lacked the tooling, frameworks, and support systems to fail fast and recover quickly. So even if failure was accepted, the cost of failure was still too high. That shaped how we approached the project from the start.

I can’t share specifics, but imagine this: we had a clear Set of Problems. We were at Point A and wanted to reach Point B - where those problems would be solved.

We had several possible paths. The problem was large and complex, so we had to be thoughtful. After rounds of discussion, we chose the incremental path. It felt safer. More predictable. And predictability mattered, because we didn’t have the tooling, frameworks, or capacity to absorb much risk.

That’s the first big insight: failure tolerance isn’t a cultural value alone - it’s a system constraint. You can “embrace failure” in your values doc, but if the environment can’t handle failure, it’s still unsafe.

The project dragged on. We didn’t see results. Motivation dropped. Eventually, we pulled the plug and went back to the drawing board.

Since then, I’ve been reflecting. I’m a long-term control optimizer - meaning I don’t need to control the moment, but I do need to extract something useful for the future. Here’s what I came away with.

1. Plan for impact, not predictability

Planning for predictability is often just fooling yourself.

You can’t turn people into equations. Humans aren’t linear. We are emotional, social, biased. We crave certainty and reject surprise. And so, when we plan for predictability, we tend to over-commit to the plan - because deviation feels like failure.

But when the plan becomes the point, you lose the original intent. You end up optimizing for smoothness, not outcomes.

This mindset feeds into zero-risk bias - the tendency to eliminate small risks instead of reducing the big ones. And once that sets in, creativity suffers. Speed becomes more important than insight. The project becomes a checkbox.

Planning for impact is messier. It carries risk. But it’s the only path that leads to meaningful results.

2. Stories matter more than data (sometimes)

We tried to be data-driven. We looked at metrics, measurements, and models. But we also had a story - a feeling, a shared narrative that emerged from experience and discussion. And the story said something different than the data.

We ignored it.

Jeff Bezos once said, “When the data and the anecdotes disagree, the anecdotes are usually right.”

He’s right. In hindsight, our story was the truth, but we didn’t act on it. We used data to rationalize and justify, rather than to challenge or sharpen.

Paying attention to the stories that circulate, formally and informally, is a critical skill. I underestimated it. I won’t do that again.

3. Skills are situational

Sometimes, even the most talented teams can’t deliver. That was the case here.

Over time I’ve come to believe that skills and competencies are not absolute. They’re contextual. A team that thrives in one setup can flounder in another. The same person can be brilliant, or ineffective, depending on the environment, constraints, and support systems.

So when something’s not working, the instinct is to assess the people. But often, the better move is to assess the situation.

If you want better outcomes, don’t just hire better people. Build better environments.

Raising odds in the future

There are no binary answers. There’s no guaranteed way to prevent failure. But you can raise your odds by remembering:

• Predictability is comforting, but often counterproductive.
• Stories carry truths that data can miss.
• Talent needs the right conditions to thrive.

And next time, I’ll question predictability more aggressively. I’ll listen for the story behind the numbers. And I’ll shape the environment, not just the plan.