Jobs That Finished With No Checklist at All
Why this matters
This is the figure that decides whether your other checklist numbers mean anything. A completion rate is computed over checklists that exist; a pass rate is computed over items on those checklists. Jobs that finished with no checklist at all are outside both, so a shop can post excellent numbers on a shrinking fraction of its work and never see it happen.
It arrives as a bare count, which is the first problem. Twelve is an emergency at a shop that ran fifteen jobs and a rounding error at one that ran nine hundred. It is also only a proxy: it measures whether a record was created, and a created record is not a check performed. Treat it as the coverage gate on the other two numbers rather than as a quality measure in its own right.
What the number is actually made of
Completed jobs in the window with no checklist attached. One row per job, not per checklist and not per customer.
Three details decide how you read it.
The anchor is the job's completion, and the window has to be closed. A job that finished yesterday afternoon may get its checklist attached this morning by the office. Reading this count on a period still running always overstates the gap, and the overstatement lands entirely in the most recent few days, which is where anybody scanning a chart looks first.
No checklist attached is not the same as nobody started one. Three mechanisms produce the same empty result and they need different fixes. Somebody deleted an attached checklist, usually because it was the wrong template. The job was created through a route that skips template application - a recurring schedule, a bulk import, a request that came in through a portal or a booking form and was converted straight to a job. Or the job type genuinely has no template, which is not a gap at all. The count cannot tell them apart. The job's creation source can, and it is usually the fastest single cut to run.
It counts jobs, not exposure. A thirty-minute call and a full day of installation each contribute one. That is the right construction for a habit measure and the wrong one if you are trying to size risk, and it is worth saying out loud in the meeting where somebody proposes to weight it.
Give it a denominator before you give it a meaning
Two numbers, computed in this order, and the second one is the one to publish.
Raw coverage gap: uncovered completed jobs over all completed jobs in the window. Useful once, to size the problem.
Eligible coverage gap: uncovered completed jobs in job types that are supposed to carry a checklist, over completed jobs in those same job types. Both sides drop the excluded types. This is the figure that can be tracked over time, because the raw one moves every time your job mix moves, and your job mix moves every season.
Getting the denominator wrong in the usual direction - dropping the ineligible jobs out of the numerator and leaving them in the denominator - understates the gap by the share of your book that is ineligible, which in most shops is not small.
The legitimate exclusions, and why naming them protects the number
A shop that expects a checklist on every quick call will stop trusting this number inside a month, because the number will be full of jobs that never should have carried one. Write the exclusions down, as job types, before you publish anything.
The ones most shops end up excluding:
- A return visit on a job that already has a checklist. The checklist belongs to the work, not to the truck roll.
- Estimate and survey visits where no work was performed. There is nothing to check.
- Parts drops, and trips to collect equipment.
- Warranty and callback visits, which usually want a different record entirely - what was found, and why it came back.
What should not be on that list is any type excluded because the crew keeps failing to complete it. That is the gap telling you something, and moving it into the exclusions is how a coverage number becomes decoration. If a type genuinely does not need a checklist, say so and mean it; if it needs one that nobody can finish, that is a template problem and the next section is where it gets found.
Coverage gaps cluster, so cut before you conclude
The single most common mistake with this figure is reading the total and calling a meeting. Coverage gaps are almost never spread evenly, and where they cluster tells you what kind of fix is needed. Run three cuts, and run them in this order, because each answer changes how the next one reads.
Job type first. If most of the gap sits in one type, everything else you might cut by is contaminated: the technicians who run that type will look worse than everybody else purely for running it, and the hours that type is usually scheduled into will look worse purely for holding it.
Then technician, inside the affected type only. Compare a tech's uncovered share of that one type against the pooled share of everybody else who ran it. Not against the type's own overall rate, which contains his own jobs and therefore drifts toward him, and not against another tech's whole-book rate, unless the two run the same mix.
Then time of day, again inside the affected type. The last slot of the day and the peak weeks of your season are where anything optional gets dropped, and a checklist that survives a quiet Tuesday and dies at four thirty in August is a length problem, not a discipline problem.
Hold a floor of about 30 completed jobs in the window for any slice you intend to read as a rate. Below that a single job moves the figure by more than three points, which is larger than most of the differences you would be acting on. Slices under the floor get reported as counts and pooled, not ranked.
Worked example: one quarter, cut three ways
A shop closes a quarter with 612 completed jobs.
Raw. 88 completed jobs carried no checklist: 14.4 percent of completed jobs in the quarter.
Exclusions applied. The shop's written exclusion list covers four job types, which together ran 53 completed jobs in the quarter. Of those 53, 41 correctly carried no checklist and 12 carried one anyway, which is harmless. Removing those 53 jobs from both sides: 47 uncovered out of 559 eligible completed jobs, 8.4 percent. The raw 14.4 percent was computed on all 612 completed jobs and the corrected 8.4 percent on the 559 eligible ones, so they are two different bases and only the second one is trackable.
The hygiene check, before any diagnostic cut. Sort the 47 by how the job was created. Nine of them came from the shop's recurring maintenance schedules, a route that generates jobs without applying the job type's template. That is a configuration fault, it is fixed once by attaching the template to the schedule, and no conversation with any technician would have changed it. Those nine stay in the published count until the configuration is fixed, and they come off both sides of the cuts below, which run on the remaining 38 of 550 eligible completed jobs, 6.9 percent. Running the three cuts without this step spreads nine office-side jobs across whichever types, people and hours the maintenance schedule happens to favour, and every one of those cuts comes back slightly wrong.
Cut one, job type. 26 of those 38 sit in a single job type, the shop's shortest call. That type ran 118 eligible completed jobs in the quarter, and all nine of the recurring-schedule jobs sat in it as well, because a maintenance schedule generates one kind of visit, so the denominator to read it against is 109.
- That type: 26 of 109, 23.9 percent of its own eligible completed jobs.
- Everything else: 12 of 441, 2.7 percent of its own eligible completed jobs.
Those two denominators add back to the 550. Both slices clear the 30-job floor comfortably, so that is a real difference and not a small-sample artefact. It also shows what the hygiene step was protecting: left in, the nine would have read as 35 of 118, 29.7 percent, and a settings fault would have been reported as a template fault.
Cut two, technician, inside that type. Seven technicians ran it. Only one of them ran enough of it to read as a rate: 32 of the type's 109 eligible completed jobs, carrying 8 of the 26 uncovered, 25.0 percent of his own 32. The other six ran 77 between them, carrying the remaining 18, 23.4 percent of those 77. Individually all six sit under the 30-job floor, so they are pooled rather than ranked. Compare him against that pooled 23.4 percent rather than against the type's own 23.9, which contains his own jobs and drifts toward him. Even on the cleaner comparison, 25.0 against 23.4 is not a finding. Nobody here is the problem.
Cut three, time of day, inside that type. Of the 26 uncovered, 17 were scheduled into the last slot of the day. That type ran 33 of its 109 eligible completed jobs in the last slot.
- Last slot: 17 of 33, 51.5 percent of that type's last-slot eligible completed jobs.
- Rest of the day: 9 of 76, 11.8 percent of that type's earlier eligible completed jobs.
The read. The gap is concentrated in one short job type, inside that type it is concentrated in the last call of the day, and it is flat across the people who run it. That is a template-length and scheduling finding, and the fix is either a shorter template for that type or a schedule that stops putting it in the slot where the crew is trying to get home. Had the shop stopped at the headline 14.4 percent, the meeting would have been about discipline, aimed at seven people, six of whom cannot be measured at all at their volumes. Note where that leaves the risk: this figure counts jobs and not exposure, and the uncovered jobs are the shop's shortest call, so 6.9 percent of jobs is a good deal less than 6.9 percent of the work.
The ladder this produces
Work it top down and stop at the first rung that fires. Each rung's answer is what makes the rung below it readable, which is why the order is the order.
- Does it have an eligible denominator, and have the jobs created through template-skipping routes been separated out? If the exclusions are not written down as job types, stop; you cannot read a coverage gap whose denominator is an opinion. If recurring schedules, imports and converted booking requests have not been split off, fix that first, because those jobs are a settings problem masquerading as a field one.
- Is it concentrated in one job type? Then it is a template-fit problem. Either shorten the template for that type or exclude the type on purpose, in writing, and let the number go back to being honest about the rest.
- Inside the affected type, is it concentrated on one or two people who each ran at least 30 of that type in the window? Then it is habit or tooling. Go out with them once before anyone writes anything down; the usual cause is a template that asks for something that job does not have.
- Is it concentrated in a slot of the day, or in your peak weeks? Then it is capacity. The checklist is the first thing dropped when the day is running long, and the fix is in the schedule or the template length, not in a reminder.
- Is it scattered evenly across type, person and hour? Then nobody has actually been told the checklist is mandatory and what happens if it is missing. That is the cheapest fix available and the one shops skip, because a scattered gap looks like a culture problem and culture problems feel like they need a programme.
References
- See related: Checklist Completion Rate and What Completion Means, for the rate this coverage figure gates
- See related: A Checked Box Is Not a Passed Check, for the item-level number underneath both
- See related: Write a Checklist That Actually Gets Used, and Service Quality Checklist Discipline
- See related: The Job Closeout Checklist, for what a short call type's template can reasonably ask for