How to Set an End Time and Keep It

Why this matters

You have set an end time before. It held for nine days. What broke it was not weakness, it was four specific mechanisms that arrive in the last hour of the day, and each one has a different fix. Owners treat a broken end time as a character problem, resolve to try harder, and get the same result, because trying harder does nothing to a phone that rings at 4:40. This is the procedure: pick the time against an anchor that has another person in it, build the closing block that makes stopping possible, instrument the four breakers, and then read your own tally instead of your own discipline.

Step 1: Set the time backward from an anchor with a person in it

Do not set the end time forward from your workload. Workload expands to whatever you give it and always will, so an end time derived from "when I finish" is not an end time, it is a hope.

Set it backward from a fixed commitment that involves someone else and starts without you: dinner at a set time, a child's pickup, a class, a standing call with a friend. The anchor matters more than the hour. A self-promise to stop at 5:30 is a negotiation you will hold with yourself every single evening, and you will lose most of them, because the person asking for one more hour is you and you have a good reason. An anchor with a person in it removes the negotiation, because the cost of the overrun is visible to someone else.

If you genuinely have no anchor, build one. It does not have to be significant. It has to be at a fixed time and it has to involve somebody who notices.

Step 2: Build the closing block, and treat it as part of the end time

Thirty minutes before the end time, you stop taking on new work and you close the day. This is not administrative fluff; it is the mechanism that makes stopping possible at all.

Without it you are mid-thought when the clock hits, so you either overrun to reach a stopping point or you leave and carry the unfinished thing home in your head, which is not stopping either. Owners who set an end time and no closing block usually report that they left on time and the evening was ruined anyway. That is the closing block missing, not the end time failing.

Three things go in it: park the unfinished work in writing, clear anything that must not wait overnight, and name tomorrow's first task. The parking note is the one that does the work, and it is three lines at most: where you got to, what the next action is, and the one thing you were about to look up. Three lines written at 5 p.m. saves twenty minutes of cold restart tomorrow, and more importantly it lets your head put the thing down.

Step 3: Instrument the four breakers

Each one arrives differently and each needs its own fix. Do not try to solve all four with resolve.

The late inbound. A call or a walk-in at 4:40 that becomes an hour. The instrument is a last-call cutoff: a stated time after which inbound work goes to the answering path and gets returned within a stated window the next morning, and a named person who handles anything that genuinely cannot wait. The cutoff needs to be earlier than your end time by at least the length of a typical call plus the closing block, or it does not protect anything.

The unfinished quote. Work that has a natural completion point somewhere past your end time. The instrument is the parking note from Step 2, plus a rule about what you start after a stated hour: after that hour you do not begin anything you cannot park in three lines. That rule is more useful than it sounds, because most end-time overruns are decided at 4:15 when you open something you should not have.

The tech who needs you at 5:15. The instrument is a same-day question deadline, published to the crew, plus a named person who holds decision authority after it. The deadline alone does not work. Techs are not calling at 5:15 to bother you, they are calling because a real thing happened and you are the only route. Give them a second route or they will keep using the only one they have.

Your own one-more-thing. The instrument is a hard count: one overrun of up to about ten minutes per week, tracked on the same tally as everything else. A budget of one is enough to absorb the genuine case and small enough that you spend it deliberately. An unlimited allowance is what you have now.

Step 4: Tally kept against broken, with the cause

One mark a day. Kept or broken, and if broken, one word for which of the four breakers did it. This takes five seconds and it is the entire diagnostic apparatus.

Run it for at least 15 working days before drawing any conclusion. Below that the sample is too small to separate a cause from a bad fortnight, and you will end up fixing the thing that happened twice rather than the thing that happens weekly.

Step 5: Read the pattern, then fix the routing before the time

The rule: when one cause accounts for 50% or more of a period's broken end times (unit: per cause, over a tally of at least 15 working days), fix the routing for that cause before you touch the end time itself. Moving the end time later does not remove the cause, it relocates the breakage to a later hour and costs you the difference.

Below 50%, no single cause dominates and the useful move is different: work the two largest causes together, or accept that your current rate is what this season looks like and re-read after it passes.

A worked tally

An owner sets 5:30 p.m. as the end time, anchored to a family dinner at 6, with a closing block starting at 5. He tallies 4 weeks, 20 working days.

First period. Kept 11, broken 9, so 55% of the 20 days held.

Causes of the 9 breaks: 5 late inbound, 2 unfinished quote, 1 tech question, 1 self-inflicted.

The read. He had this filed under discipline. The tally says otherwise: 5 of the 9 breaks, about 56% of the breaks, come from one cause. Run that against the stated rule: 56% clears the 50% threshold, on a tally of 20 working days which clears the 15-day floor. So the action is to fix the routing for late inbound, and specifically not to move the end time to 6.

There is a second pattern in the tally worth noticing. Four of the five late-inbound breaks fell on a Tuesday or a Wednesday, which are the two days he does not run calls and is therefore reachable at the shop. The inbound was not arriving more often on those days. He was simply the one who picked it up.

The fix. A last-call cutoff at 4:45. Inbound after that routes to the answering path, with a stated commitment that calls are returned within the first hour of the next business day. Anything a customer describes as an emergency goes to the on-call tech, who has authority to dispatch without asking. The cutoff sits 45 minutes before the end time, which covers a typical call plus the closing block.

Second period, another 20 working days. Kept 17, broken 3, so 85% of the 20 days held. That is a rise of 30 percentage points, from 55% to 85% of 20 working days. The remaining 3 breaks: 1 late inbound that genuinely warranted him, 1 unfinished quote, 1 self-inflicted, which was his one budgeted overrun for that week and he spent it deliberately.

Note what the residue looks like. Three unrelated single causes rather than one cause repeating. That composition is the tell that the fix landed, and it matters more than the count did. A second period at 85% where two of the three breaks were still late inbound would mean the cutoff was published but not held, which is a different problem and one you solve by talking to whoever is routing calls, not by rewriting the policy.

What he did not do. He did not move the end time later. Had he moved it to 6:30, the tally would very likely have shown roughly the same kept rate at the new time, because the 4:40 call would have become a 5:40 call and the anchor would have moved with it. Moving the time is what owners do when they have not counted the causes.

What changes the target

An on-call rotation. If you are genuinely the after-hours responder one week in three, an 85% kept rate every week is not a real target. Set it as 85% on the weeks you are not on call, and give the on-call week its own published end time, which will be later and which everyone including your household knows about in advance. A boundary that pretends the on-call week does not exist gets abandoned in the on-call week and never comes back.

A shop small enough that there is no second route. If there is nobody to route the late inbound to, the cutoff still works but its mechanism is different: the answering path becomes a message with a stated return window rather than a person, and the thing you are protecting is the last 45 minutes, not the whole evening. State the return window publicly and hit it, because a return window you miss trains customers to keep calling until they get you, which is the behaviour you were trying to change.

Peak season. Expect the kept rate to fall and keep tallying anyway. The peak-season tally is the most valuable one you will collect, because the causes that dominate under load are the ones worth building routing for. Do not suspend the end time for the season, because a boundary suspended for a season is a boundary that restarts from zero afterward.

How to tell it is actually working

• The kept rate, on the same tally, over the same length of period. Comparing a 20-day period to a 12-day one tells you nothing. Same instrument, same length.
• The composition of the breaks, which is the better signal. Scattered single causes is a working boundary. One cause repeating is an unfixed route regardless of what the count says.
• Whether the closing block survived. The first thing to get cut on a busy day is the closing block, and losing it is what turns a kept end time into a day you carried home anyway. If the tally says kept and the evenings still feel like work, check the block before you check anything else.
• Whether anyone else changed their behaviour. Within a few weeks the crew should be calling before the deadline rather than after it, and customers should be leaving messages rather than calling three times. If nothing changed on their side, the routing was written down but never actually communicated.

References

• See related: Setting After-Hours Boundaries With Customers
• See related: Work-Life Boundaries as an Owner
• See related: The Cost of an Owner Who Never Stops
• See related: The First Hour of the Day and What It Is For
• Trade-standard practice for after-hours call handling in field-service businesses