Schedule the Clean-Up, or It Never Happens

Published 2022-09-13 · Updated 2026-06-23 · By Alan Wizemann

Topics: Engineering & Development, Team Building & Culture, Product Strategy & Development, Growth Leadership, Agile Methodologies

Every fast-growing company I've ever worked in carries a quiet, accumulating debt, and it isn't financial, it's technical and experiential, and it builds up in the corners while everyone's looking at the roadmap. It's the shortcuts taken to ship something on time, the report that's held together with manual steps somebody runs every Friday, the feature that works but was never actually cleaned up, the backlog of small broken things that nobody quite owns. And every one of those companies says exactly the same thing about all of it, which is that we'll fix it when things finally calm down. The trouble, of course, is that things never actually calm down, and that, right there, is the whole problem in a single sentence, and once you genuinely accept it the rest of this becomes more or less obvious.

The reason the debt never gets paid down isn't that teams don't care about it – most of them care quite a lot – it's that paying it down always loses the fight for time. When "fix the foundational mess" has to compete against "ship the thing the business is asking for this week," the business request wins every single time, because it's louder, it's more visible, and it's attached to someone important who's going to ask about it. Debt paydown is important but it is never urgent, and important-but-not-urgent work is precisely the work that never happens at all unless you force it to. So at quip we forced it, deliberately and on the calendar.

The operating cadence ran in twelve-week cycles, and at the end of each one came a week with a single, protected purpose, which was to make everything better. Fix the technical and the experience debt. Rework the reports that had gotten crusty. Build the automation everyone kept swearing they'd get to. Clear the backlogs, and address the foundational issues that had been quietly slowing everything down for months. And the critical word attached to that week, the one that did all the actual work, was non-negotiable – it wasn't "if we have the capacity," and it wasn't a stretch goal you'd hit in a good quarter, it was a fixed, recurring part of how the teams operated, as real and as immovable as any feature deadline. That one design choice, making it mandatory and putting it on the calendar in ink, is what makes the whole difference between a clean-up that actually happens and a clean-up that gets perpetually deferred into a someday that never arrives. The instant you make debt paydown optional, or contingent on having spare time lying around, you have killed it, because there is never spare time. The only version of this that survives contact with a busy roadmap is the version that's walled off and protected, where the answer to "but couldn't we use that week to ship the new feature instead" is simply no, that is not what the week is for. The protection, in other words, is the entire mechanism, the only part that actually does any work. Intentions don't pay down debt; rituals do.

The counterintuitive part, and the thing you have to be able to say to a skeptical executive, is that the clean-up week is exactly what makes the other twelve weeks fast. It feels like you're giving up productive time, a whole week, several times a year, not shipping anything new, and that feeling is hard to argue with in the moment. But debt is a tax on velocity – every crusty report, every manual workaround, every foundational shortcut makes all of your future work a little slower and a little more fragile than it needs to be. Paying it down on a regular schedule keeps the system fast and stable, which means the twelve weeks of actual building run faster than they ever would have if you'd skipped the cleanup and let the debt quietly compound. You're not losing a week of speed at all; you're buying twelve weeks of it back, at a price that only looks expensive if you forget what the alternative actually costs you.

There's a cultural payoff too, and honestly it matters more than people expect going in. Teams that are never once allowed to fix the things that bother them get demoralized, and it's not subtle – engineers know exactly where all the bodies are buried, they have to work around the same mess every single day, and being told over and over "no time, ship the next thing" grinds them down in a way that shows up in the work eventually. Giving them a sanctioned, recurring window to actually fix the stuff that's been bugging them, and to clear it rather than just talk about it in a retro, is enormously good for morale. People do better work when they're allowed to be proud of the system they're working in, and not just the features they've been forced to keep bolting onto the side of it.

I'll name the discipline this requires, because it's genuinely harder than it sounds on a slide. The pressure to cancel the clean-up week is relentless and it always sounds reasonable – there's always a launch that would benefit, a deadline that's a little tight, an executive who'd love to pull that week forward into feature work "just this once." And the moment you say yes to "just this once," you've taught the entire organization that the week is negotiable, and a negotiable ritual is no ritual at all, it'll be negotiated away within two cycles and you'll be back where you started, swearing you'll fix things when it calms down. Protecting it, repeatedly, against good-faith and well-argued pressure (the worst kind, because it's the kind you actually want to say yes to) is the actual job. The calendar is the easy part of all this; the defending of it, week after week, against people you respect, is the genuinely hard part.

The broader principle goes well past engineering, and I've come to think it's nearly universal. Any important-but-not-urgent work at all – debt paydown, foundational investment, the unglamorous maintenance that quietly keeps a system healthy – will lose to urgent work forever, unless you give it protected, recurring, non-negotiable time on the calendar. You cannot rely on it happening when things calm down, because the calm is a myth that the next quarter always disproves. You can only rely on it happening when you've built it into the rhythm of the place and committed, out loud and in advance, to defending it. So don't intend to clean up. Schedule it, protect it, and refuse to trade it away, because the mess you keep meaning to fix is only ever actually fixed on the days you decided, ahead of time, that fixing it was the only thing allowed to happen.