Fast Pipelines, Slow Feedback
Feedback is oxygen for teams. Yet most of it arrives long after it could’ve helped.
By the time we finally talk about what went wrong in a sprint, the next one’s already on fire. Retros have become a sort of polite archaeology. We dig up the bones of dead decisions, marvel at how we once lived, then promptly bury them again.
We’ve built these incredibly fast technical systems—CI pipelines that deploy in seconds, feature flags flipping like light switches—and yet our human systems crawl around like dial-up in a fibre world.
You can feel the cost. The quiet drift in tone on Slack. That engineer who’s suddenly “a bit off.” The meeting that everyone leaves with a strange tension in the air but no one mentions it because, well, we’ll cover it in the retro.
We don’t need more feedback. We need it sooner.
A decent leader—and I don’t mean the title-on-the-email kind, I mean the real sort, the ones who actually notice things—learns to catch the tiny signals before they calcify. The hesitation in a pull request comment. The silence after a new process is announced. These moments are where the real feedback hides, squirming awkwardly, waiting to be named.
And it doesn’t have to be grand or ritualised. Honestly, half the time all you need is a two-sentence check-in:
“Hey, that seemed weird. You alright with how that went?”
That’s it. That’s your low-latency system.
The best teams I’ve seen treat feedback like infrastructure. They don’t wait for retros; they ship it continuously, like a background process. Small corrections, frequent merges.
(And when it’s working, you barely notice it. The air just feels lighter. People take bigger swings because they trust someone will tell them early if they’ve missed.)
It’s not about making everyone hypersensitive or turning the office into a therapy circle: it’s about shortening the emotional event loop. Because once feedback hardens into resentment, you’re debugging people, not code.
Maybe that’s the crux of it: leaders design systems, whether they mean to or not. Some systems are tuned for reflection. Others for decay.
Personally, I’d rather run a team that fails frequently and talks early—even if it’s messy. Feedback, like milk, does not age well.