Total repairs completed is the number most shops put on the board. It feels like progress, it goes up over time, and it tells you almost nothing you can act on. A high repair count with a queue full of angry customers is not a good week, it is just a busy one.
The numbers worth tracking are the ones that point at a specific thing you can fix this week. Most of them are about time and where it goes, not about totals.
The vanity ones
A few numbers look important and are mostly noise on their own.
Total devices processed tells you volume, not health. Revenue per day swings with the mix of jobs and hides whether the floor is actually running well. Even a single average turnaround time, taken across every kind of repair, blurs a same-day screen swap and a two-week board-level job into one meaningless middle number.
These are fine as background. The mistake is managing to them, because none of them tell you what to do tomorrow.
The ones that point at a fix
The useful numbers break the work into stages and show you where it stalls:
- Time from intake to diagnosis, since a device sitting untouched for two days is a queue problem, not a repair one
- Time in each stage, so you can see whether jobs pile up at diagnosis, at parts, or at pickup
- Queue depth by stage, which shows the bottleneck moving around the shop day to day
- Advisor throughput, so you know whether one person is carrying the counter while another coasts
- Time waiting on parts, separated out, because it is a delay you manage differently than bench time
Each of these points at a place, not just a total. Devices sitting two days before anyone diagnoses them is something you can go fix. Two hundred and forty repairs done is not.
Reading the stages together
The value shows up when you watch the stages as a flow instead of one number at the end. A device that takes eight days is not slow everywhere. It is usually fast through most stages and stuck in one. Find the stage where things pile up and you have found the constraint that is setting your whole turnaround time.
Sonic is built around this kind of stage-level view, tracking intake time, time per stage, and where the queue bottlenecks so the constraint is visible instead of hidden inside an average. But the habit matters more than the software. Even a whiteboard that separates waiting on diagnosis from waiting on parts tells you more than a single turnaround number ever will.
Pick the one stage where devices sit longest and watch it for a week. Fixing your slowest stage does more for turnaround than any amount of pushing the whole shop to go faster, and it is the number that total repairs completed will never show you.