China moves to rein in the 996 culture of long hours

The posts were loud because the tension is real. A home appliance giant told staff to finish by 6:20 pm to cut bureaucracy. A drone maker told teams to go home at 9 pm unless truly urgent. A prime time talk show ran for weeks with celebrities and young workers trading playbooks on how to say no. What looked like a social debate about hustle is better understood as a control system problem. Hours were the only lever managers could pull at scale, so hours became the plan.

The country has tried to stamp out the 9 to 9, six days a week norm. Labor law states eight hours a day and forty four a week, with overtime capped at one hour a day, or three if health is ensured, and thirty six a month. Yet the first five months of 2025 still averaged forty eight and a half hours a week. That gap between statute and practice is the operating environment many founders inherit. The lesson is not that people ignore rules. The lesson is that incentives, targets, and process swallow rules when they are badly designed.

A director in renewable power can put in more than twelve hours a day because half of his monthly income sits in a bonus that only clears if he hits a number. He calls those hours voluntary. They are rational. Many employees chase extra time to signal commitment or to protect income. Many managers set aggressive targets that rely on a surplus of human time as the buffer. In that environment, mandated clock off times are a speed bump. They slow the car for a moment, then the driver finds a side road.

Look under the hood and you will find three recurring failure modes. First, bonuses and stretch targets are sized for output that assumes longer days. That turns working hours into the cheapest lever to hit the forecast. Second, process debt compounds. Approvals stack. Reviews arrive late. Scopes expand mid sprint. When process is slow, time fills the gap. Third, managers reward visible sacrifice. Late messages get praise. Early departures attract side comments. Culture follows the reward, not the poster on the wall.

There is also a productivity narrative in play. If output per worker is lower than peers but the growth target is set to match peers, someone will push the hours button. It is immediate and visible. Investment and redesign are slow and political. Hours scale without paperwork. Until they cannot. That is when teams burn, quality slides, and the best people leave.

Regulatory pressure sits in the background. The ILO treats excessive working hours as a forced labor indicator. Europe is building enforcement muscle around forced labor in supply chains through a new regulation. Opinions differ on how much that touches white collar offices in major cities. Even so, export oriented firms that sell to strict markets will be asked to evidence how their labor practices meet standards. That will not be solved with a single memo about clocking off. It will be solved with a system that makes overtime unnecessary for the baseline plan.

So what should founders take from all this. A clock off rule is a constraint, not a culture. You use a constraint to force a redesign, or you perform compliance while the real system stays the same. If you cap hours without changing the design, the work moves after hours into chat threads, personal devices, and weekend hush projects. The optics improve. The load does not. Customers still expect delivery. Sales still carry targets. Finance still needs numbers to tie out.

Treat the constraint like a design brief. Start by rewriting targets in terms of repeatable throughput per person, not total hours spent. When you score a team this way, you surface bottlenecks that hours used to hide. The old review that takes two days gets a deadline before noon. The product spec that used to drift gets a lock date and a change log. The weekly number stops being lines of code, story points, or calls placed. It becomes valuable outcomes shipped, defects kept under a threshold, demos that convert, tickets resolved within a clear service level.

Align pay with the new system. High bonus ratios invite hours inflation. The more your people rely on variable pay, the more they will stretch time to protect income. This is not a moral claim. It is systems math. Shift more compensation into base for critical roles where hours have drifted into habit. Keep variable pay tied to team outcomes that reflect throughput, quality, and customer value. Remove any metric that can be gamed by simply staying online.

Rebuild the day. Set a latest send time for non urgent internal comms. Schedule handoffs that happen before the last hour of the day, not inside it. Introduce a release cadence that favors predictable sprints over long hero runs. Create an on call rotation with clear escalation rights so that urgent stays urgent, and everything else respects the boundary. These are not perks. These are operating rules that teach the system to work inside the constraint.

Retrain managers to recognize real output. Many managers learned to equate late messages with ownership. Replace that signal with a better one. Ask whether the team closed the loop with their customers on time. Ask whether the code stayed below the defect ceiling. Ask whether the campaign shipped on the date that was committed a month prior. These signals cannot be faked by staying late once a week. They require structure every day.

Fix the boring parts that hours have been covering for years. Shorten approvals. Pre decide defaults. Remove one status meeting for every non trivial dashboard you build. Replace ad hoc requests with intake forms that capture scope, priority, and deadline in a single place. Give one person the right to say no when the board, or the biggest customer, wants one more thing added mid sprint. The fastest way to de risk overtime is to remove the work that never should have entered the system.

Watch for shadow overtime. If you cap office hours, people will switch to private channels after hours unless the incentives change. Track response expectations. If the median time to first internal reply outside business hours is under thirty minutes, you have an always on culture with a new coat of paint. You can fix this by removing any career advantage linked to being reachable at night. Measure follow through the next day. Praise teams that close loops within business hours. Make it normal to defer a response until morning.

Remember customer truth. There will be genuine crunches. A launch week. A hot fix that protects customer data. A black swan supply issue. The rule is simple. Call it explicitly. Pay for it cleanly. Recover quickly. Do not let exceptional weeks write the baseline plan. Teams can rally for a sprint. They cannot live in a sprint forever.

The talent market is changing in ways founders cannot ignore. Many workers, especially younger ones, draw a harder line around time. They do not want to perform exhaustion for security. They want predictability and real progress. The firms that protect boundaries win retention. Attrition is not a line item you can ignore. The cost to replace a mid level operator, onboard them, and rebuild team trust is worth more than a few extra hours of work this quarter. Protecting time is an efficiency play dressed as empathy.

Maps differ by city and sector. Factories that rely on overtime pay to deliver livable wages will react to different levers than software teams in urban offices. Low base pay pulls workers toward extra hours by necessity. If you are building in that environment, your hours policy will fail unless you address the income design. That can mean higher base pay or tighter linkages between skill growth and pay progression. It is hard work. It is also how you break the cycle.

It is useful to note the media frame, including the primetime show that helped people talk about saying no. Dialogue moves norms. It does not fix backlogs, dysfunctional reviews, or misaligned targets. Founders do not run shows. Founders run systems. If you want the optics without the outcome, write a memo and buy dinner. If you want the outcome, rebuild the work.

Your edge sits in one question. Can your team produce predictable value in a normal week. If the honest answer is no, clock off rules will be theater. If the answer is yes, you will not need rules to keep people late. People go home when the day is designed to finish.

This is why the China overtime culture shift matters outside China. The story exposes a universal startup mistake. We try to scale results by scaling human time. That works until it breaks. Then everyone reaches for policy and posters. The better path is quieter. Redesign the target. Clean the process. Align the pay. Teach managers to praise outcomes instead of sacrifice. The clock off rule becomes a guardrail you rarely hit, not a wall you crash into every night.

Founders do not win with louder hours. They win with systems that make hours irrelevant. This posture may look soft to old playbooks. It is not. It is the only path that compounds.

Most teams do not need another hack. They need fewer excuses to avoid the rebuild. Build for normal weeks. Reward real throughput. Protect time so that focus can finally do its job. That is how you get out of the hours trap and into a company that ships on purpose. The signal is plain. A culture that glorifies exhaustion is a fragile system in disguise. The correction will not come from slogans. It will come from design.