Jobs per Day per Technician: Why More Is Not Better
Why this matters
Every other operations number gets worse when the work gets worse. This one gets better. Split a single visit into two and the figure rises, while the shop drives an extra time, spends an extra hour on site, and bills the same customer for the same problem. A shop that watches this number climb and congratulates itself has bought a productivity story with its own margin, and the record will support the story right up until somebody puts revenue next to it.
The number treats every job as the same size
Completed jobs in the window, over technicians, over working days. The arithmetic is trivial and the assumption is not: it treats a job as a unit of roughly constant size, so that two jobs represent twice as much delivered work as one.
That holds only while your job mix holds. The moment the mix shifts toward shorter work, the number rises without a single extra hour of value leaving the truck, and nothing on the face of the figure says which happened. Job length is the hidden variable, and the shortest jobs in almost every shop are the ones you least want more of: return visits on something that was supposed to be finished.
The inversion, worked over two quarters
The same shop, the same four field technicians, 62 working days in each quarter.
| Q1 | Q2 | |
|---|---|---|
| Completed jobs | 744 | 868 |
| Jobs per technician-day | 3.00 | 3.50 |
| First visits (not a return on the same problem) | 684 | 684 |
| Return visits | 60 | 184 |
| Return visits as a share of completed jobs | 8.1% | 21.2% |
Jobs per technician-day: 744 / (4 x 62) = 744 / 248 = 3.00 in Q1, and 868 / 248 = 3.50 in Q2. That is a 16.7 percent rise against the Q1 figure of 3.00, and it is the sort of move a shop puts in a meeting.
First visits held at exactly 684 in both quarters. The shop addressed the same number of distinct customer problems and drove 124 more times to do it, because 868 minus 744 is 124 and every one of those extra completions was a second roll on a problem already visited.
Billed revenue finished Q2 within about 1 percent of Q1. So trip yield, revenue per completed job, fell: with revenue essentially flat, the ratio moves as 744 / 868 = 0.857, a 14.3 percent fall against the Q1 figure.
Those two percentages are not two independent readings. With revenue, crew and working days all held, revenue per completed job moves as the exact reciprocal of completed jobs per technician-day, and 1 divided by 1.167 is 0.857. The 16.7 percent rise and the 14.3 percent fall are one fact counted twice, in two different units: one is a count per technician-day, the other is money per job. Anyone quoting both as corroborating evidence is quoting the same evidence twice.
What the 124 extra rolls actually consumed
Take the shop's own measured figures: an average of 28 minutes of travel preceding a job, and an average of 0.9 hours on site for a return visit, which is shorter than a first visit because a return is a narrower scope by construction.
- Travel: 124 x 28 minutes = 3,472 minutes = 57.9 hours.
- On site: 124 x 0.9 hours = 111.6 hours.
- Total: 169.5 hours.
Against the quarter's field capacity of 4 technicians x 62 days x 8 hours = 1,984 hours, that is 8.5 percent of the quarter. Spread across 248 technician-days it is about 41 minutes of a technician's day, every day, for the whole quarter.
So the headline reads: a 16.7 percent rise in jobs per technician-day, bought with 8.5 percent of the quarter's field capacity, for zero additional distinct problems solved and flat revenue. Both of those percentages carry their own base and their own unit, and neither is a share of the other.
The ceiling is set by mix, which is why no cross-trade target exists
Use the same shop's figures to bound the number from above. An 8-hour field day, 28 minutes of travel ahead of each job, first visits averaging 1.6 hours on site, returns averaging 0.9.
- A day of nothing but first visits: 1.6 + 0.467 = 2.067 hours a job, so 8 / 2.067 = 3.87 jobs.
- A day of nothing but returns: 0.9 + 0.467 = 1.367 hours a job, so 8 / 1.367 = 5.85 jobs.
- The same day of first visits with each technician-day held to one sector, so 16 minutes a leg instead of 28: 1.6 + 0.267 = 1.867 hours a job, so 8 / 1.867 = 4.29 jobs.
The arithmetic ceiling for this crew ranges from 3.87 to 5.85 jobs per technician-day depending on nothing but mix. The third line is the other lever: holding a technician-day to one sector moves the first-visit ceiling by 0.42 with the mix untouched. Their actual 3.00 and 3.50 both sit inside that range, and so would almost any figure somebody quoted at them from another shop.
This is why a borrowed benchmark for this number is worthless rather than merely imprecise. A maintenance-heavy cleaning route and a heat-exchanger replacement crew are not running the same number badly, they are running two different numbers. Derive your own ceiling from your own two measurements, on-site time and travel, then read your figure as a share of that ceiling and re-derive the ceiling whenever your mix moves.
If your actual figure comes out above the ceiling you just computed, one of the inputs is wrong or the jobs are not what the record says. Three candidates, in the order they are worth checking. The field day is longer than 8 hours, which the timesheets settle in a minute. On-site time is shorter than your medians say, usually because the medians were taken from a period with a different mix. Or jobs are closing without a visit, which is what happens when a phone resolution, a no-access call or a cancellation is closed as completed. That third one is the expensive answer, because a phantom visit inflates the job count and every other per-job figure in the shop at the same time, and unlike the other two it leaves no trace in payroll.
When geography sets the ceiling, not effort
If the figure sits well below the ceiling you derived and travel per job is high, scheduling discipline will not move it, because the binding constraint is the shape of the service area rather than the shape of the day. That is the third line of the ceiling calculation above: a 12-minute cut to the average leg lifts the ceiling by 0.42 jobs per technician-day, which across the quarter's 248 technician-days is about 104 jobs of headroom that nobody has to work harder to get.
Read that as a ceiling move, not as a forecast, and read it against the travel card, which owns the question of what headroom is actually worth. Whether this shop is short of capacity or short of demand is the fork utilization answers, not this number.
Five things that raise it, and the tell that separates them
A rising figure has more than one cause and they call for opposite responses. The last column is the cut that distinguishes each from the others, and each one is available from records a shop already keeps.
| Driver | What is happening | The tell |
|---|---|---|
| Return visits | Second roll on a problem already visited | First-visit count holds flat or falls while completed jobs rise |
| Route density | Same work, less driving between stops | Travel time per job falls; on-site time per job holds |
| Job mix shift | More genuinely short work sold | On-site time per job falls and first-visit count rises together |
| Staged scope | One visit deliberately split, part on order | Completed jobs rise while revenue per first visit holds |
| Longer day | Same jobs, more hours worked | Logged hours per technician-day rise |
Two of those five, route density and a genuine mix shift, are wins. One, a longer day, is capacity borrowed from next month and shows up in premium hours. One, staged scope, is neutral and often correct. Only return visits are straightforwardly bad, and they are the one driver that leaves first visits flat, which is why that single column is worth more than the headline figure it sits beside.
The staged-scope row deserves its own note, because it is the one most often misfiled as a return. The separating fact is in the record, not in the drive: a staged visit closed the first trip as awaiting parts, with a part on order against it. A return closed the first trip as complete. If your records do not distinguish those two closures, this whole diagnosis is unavailable to you, and fixing that is a one-field change with more diagnostic value than anything else in this article.
What the number should and should not drive
It should drive a question about route density and service-area shape, and it answers that question best when read alongside travel time per job rather than alone. It should drive a review of job mix when it moves without travel moving.
It should never drive a target handed to technicians, for the reason the travel card sets out about per-person readings of a route number. Here the cheapest way to satisfy the target is to leave before the work is finished, and the two quarters above are exactly what that looks like. Nobody in that shop had to intend it: the dispatcher was clearing a backlog, the technicians were taking the next call, and the number rewarded every step.
If you want a per-person productivity figure that cannot be satisfied that way, use hours logged against a job as a share of hours available, which rises only when more real time reaches customer work, and read the job count beside it as a description of mix rather than of effort.
Where to take the 184 returns
The decision that follows Q2 is not a scheduling decision. Sort the 184 returns by job type and by who ran the first visit, and the answer is usually in the first two rows: a diagnostic step skipped under time pressure, a part the truck should carry and does not, or one technician's first visits coming back at several times everyone else's rate.
Each of those has a different fix, and none of them is visible from the number that started the conversation.
References
- See related: Trip Yield: The Number That Prices a Truck Roll
- See related: The True Cost of a Return Trip
- See related: Technician Utilization and What the Denominator Assumes
- See related: Average Travel Time and the Route Density It Implies
- See related: Zero-Revenue Jobs and Which Numbers They Touch