Track Your Comeback Rate: Your Own Quality Metric
Why this matters
"We do good work" is an opinion. Your comeback rate is a number. A comeback (also called a callback or a redo) is any time you go back to a job because it wasn't right the first time, on your dime, not for new work. It's the single most honest quality metric a shop has, because it counts failures the customer experienced, not the ones you noticed. Most shops never track it, so they can't tell a quality problem from bad luck, can't tell which tech or which job type leaks, and can't prove they're improving. Quality is cheaper than callbacks, but you can't manage what you don't measure.
Step 1: Define what counts as a comeback
Before you count anything, draw the line clearly so the number means the same thing every time:
- A comeback IS: returning to a completed job because the original work failed or was incomplete, at no charge to the customer, within a window you set (a year is a reasonable default for most repair work).
- A comeback is NOT: new work the customer ordered, a separate system failing, normal maintenance, or a problem you documented and the customer declined to fix. Those are flagged-and-declined, tracked separately.
Write the definition down. If "comeback" means whatever the tech feels like that day, the rate is noise.
Step 2: Log every one, no exceptions
The whole metric dies if comebacks get quietly absorbed and never recorded. The pressure not to log them is real (nobody likes admitting a redo), which is exactly why it has to be a no-blame habit. Capture, for each one:
- The original job and who did it.
- What failed and the likely root cause.
- The job type or system involved.
- The time and material it cost to fix.
In Manuall, tag these on the original job record so the comeback links back to the work it came from. The goal is a clean trail from failure to source, not a punishment file.
Step 3: Calculate the rate
The rate is simply comebacks divided by completed jobs over the same period, expressed as a percentage:
Comeback rate = (comebacks in the period) / (completed jobs in the period)
Pick a period long enough to have meaningful numbers (a quarter for most shops, a month if you're busy). One stray callback in a slow week swings a weekly rate wildly; don't over-read short windows.
Step 4: Slice it to find where the leak is
A single company-wide number tells you if you have a problem. Slicing it tells you where. Break the rate down by:
| Slice | What it reveals |
|---|---|
| By technician | Whether failures cluster with one person (training or habit issue) |
| By job type | Whether a specific kind of work fails more (a process or skill gap) |
| By root cause | Whether it's parts, workmanship, or diagnosis driving comebacks |
| Over time | Whether you're trending better or worse |
The slice that lights up tells you where to spend your attention. A high rate on one job type across all techs is a process problem. A high rate on one tech across all job types is a training problem. Same number, very different fix.
Step 5: Set a target and watch the trend
The absolute number matters less than the direction. Any honest shop has some comebacks; zero usually means you're not counting them. Set a target floor you consider acceptable for your trade, then watch the trend quarter over quarter:
- Trending down: your standards, training, or process changes are working. Keep going.
- Flat: you've plateaued; find the stubborn root cause that the average is hiding.
- Trending up: something changed (a new hire, a rushed schedule, a parts supplier). Find it before the reviews catch up.
Step 6: Close the loop with the crew
The number is only useful if it changes behavior. Share the rate with the crew without naming-and-shaming, frame it as the team's quality scoreboard. When a comeback's root cause is clear (a step skipped, a test not run), turn it into a standard so the same failure doesn't repeat. The comeback you analyze becomes the checklist item that prevents the next ten.
What good looks like
A shop running this well has: a clear definition everyone uses, every comeback logged without fear, a rate sliced by tech and job type, a visible downward trend, and a habit of converting root causes into standards. That shop can prove its quality to itself and to a skeptical customer, and it spends less time and money revisiting work it already got paid for once.
References
- Lean and continuous-improvement practice on defect and rework tracking.
- SBA guidance on operational metrics for small service businesses.
- See related: Setting a Quality Standard the Crew Can See, The Redo: Eat It vs Explain It, Quality Control on a Crew You Don't Watch.