Sustainable Pace

Spending more hours as a deadline approaches really does raise the output of that week. The problem is what comes after. The time bought in the pushed week leaks back out over the following two or three, and the leaking side never appears as a single line the way overtime does. Defects rise, reviews slow down, people get sick, and each of those gets recorded under a different cause.

Pagedeck did this ahead of an end-of-quarter release. Two weeks of working into the night hit the date, and the release shipped. Over the next two weeks, incoming defects ran at three times the usual rate, two people took sick leave, and one asked Grace for a meeting.

Mia counted the hours. The extra time across four people over those two weeks came to 120. Fixing defects in the two weeks that followed took 90 hours, and sick leave removed another 60. More was lost than gained, and because the two numbers had never sat in the same table, nobody knew it. Sustainable pace is the rule that does this arithmetic in advance and sets a ceiling from it.

The baseline is not the maximum this week can produce but the value next week can produce as well. Five things get decided.

1. Write the ceiling as a number. Hours per week is the simplest form and wrong on its own. The same forty hours in a week chopped up by meetings and a week left whole are different weeks. So alongside the hours, the team judges directly whether it could repeat this intensity next week. Someone who filled forty hours in a week that was half meetings is commonly the more tired one afterward. Teams doing pair programming — two people writing code at one screen — find this judgment easier, because the state of the person beside you is visible.
2. Confirm with measurements, because a judgment resting on feeling collapses under a deadline. Four numbers get used: the spread in velocity, the amount of work finished per cycle; the count of newly arriving defects; sick leave and time off taken; and how long a review waits before it starts. The first rising while the last three grow means the ceiling is already behind you, and by then the correction costs a cycle.
3. Turn exceptions into a procedure. Ban overtime and you get unrecorded overtime. Instead, fix who approves it, what compensates for it, and where the cause gets written. Recording the cause is what the procedure is for, and once the same cause appears three times it has stopped being something individual effort can solve.
4. Reclaim what was spent. A pushed week means the next cycle takes on less. Without that reduction no recovery happens, and the unrecovered speed hardens into the new baseline.
5. Fix the negotiation order in advance. Agree beforehand what gets adjusted first when a date looks like slipping. Cut scope, move the date, or add people — and adding hours doesn't go on the list. Absent from the list, the conversation in the room goes somewhere else, which is the point of writing the list before the room fills up.

The place it bites most often is the meeting that decides how much goes into a cycle. Use the number produced by a pushed cycle as this cycle's baseline and that number will only ever be met by pushing again. So results from pushed cycles get excluded from the baseline or flagged. That meeting is also where this principle most often breaks: allow the exception once and next cycle treats the exception as the baseline.

Three other places.

1. Right before a release — where the principle gets tested hardest. Whether the negotiation order actually holds here is what separates a rule from a document.
2. After incident response — someone called out at 3 a.m. who then works a normal day has just worked two days. Many teams pair this with a rule that subtracts response hours from the following day's work.
3. Where people leave — check whether what recurs in exit conversations overlaps the periods when this principle broke. Burnout, the state of depleted motivation and energy from sustained overload, handled as an individual problem means the next person gets the same thing.

It isn't a principle that admits no exceptions. No team goes home on time the night the service is down. What has to be separated is a real emergency from a planning failure. Once a quarter is an emergency; every month is a planning problem or a structural one.

"Let's take on less this cycle than last. We just came off two pushed weeks."

Putting the recovery period into the plan.

"To hit the date we cut scope or move the date. Adding hours isn't on the list."

Confirming the negotiation list.

"You were on for three hours overnight, so come in this afternoon."

Applying the rule that reclaims spent time.

Pagedeck settled three things after that quarter. Results from pushed cycles get excluded from the baseline, overtime happens only with Grace's approval and compensating time off, and the cause gets written in one line. Mia collects those records each quarter.

Of the seven causes collected in the first quarter, four were the same one: external review always landed in the final week. Splitting the review into two rounds the next quarter took overtime records from seven to two. What got fixed wasn't anyone's diligence — it was the shape of the schedule.

The most common reading is that this is a principle about not working hard. It sets a ceiling on intensity, so that follows. But what the principle fixes isn't the total; it's the point at which the total stays repeatable. Between a team that produces its maximum in one week and a team that produces steadily across twelve, the larger quarterly sum usually belongs to the second, because the first adds a recovery period and a defect-fixing period after the pushed week. Read as slack, a proposal to adopt it sounds like a proposal to give up on results, and then nobody proposes it. Showing the quarterly sum alongside is the faster way to explain it.

The other common reading is that it's a forty-hour rule. Fixing one number makes judgment easy, so the temptation is to reduce it to that. But the sustainable point differs between people and differs for the same person across months. Pin it to hours and someone wearing down inside forty hours is formally fine, while someone comfortable at forty-five is in violation. The criterion is whether next week could repeat it. Hours are one input to that judgment.