Most field service dashboards suffer from the same problem: they show what happened, in great detail, without telling anyone what to do about it. Thirty tiles, all accurate, none actionable.
A useful KPI set is small and diagnostic. Each number should distinguish between causes, not just report an outcome. Here are nine that do, grouped by the question they answer.
Group 1: Is the schedule producing work?
1. Productive utilization
Hands-on job hours ÷ available field hours
The core scheduling metric. Not assigned hours — hands-on hours, measured from arrival and departure timestamps. It isolates the thing scheduling actually controls.
If you track one number from this article, track this one. Most companies discover it sits far below the "utilization" figure their scheduling software reports, for reasons covered in detail in the three utilization formulas.
2. Travel as a share of shift
Drive hours ÷ available field hours
The counterpart to productive utilization, and the first place to look when it is low. Break it down by day of week and by territory. Structural offenders show up immediately — a standing commitment across town, an account at the edge of the radius.
3. Jobs completed per technician per day
Useful as a trend, dangerous as a target. Compare it against your own calculated ceiling rather than an industry figure; the method is in how many jobs per day. A number well below your ceiling means recoverable capacity. A number at your ceiling means growth requires trucks or shorter jobs.
Group 2: Is the schedule holding together?
4. Schedule adherence
Appointments arrived within the promised window ÷ total appointments
The honesty check on your arrival windows. It is easy to advertise a two-hour window; adherence tells you whether you are actually delivering one. If you are narrowing windows, watch this weekly — a narrowing window with falling adherence means you removed the padding without removing the uncertainty.
5. Same-day reschedule rate
Appointments moved on the day of service ÷ total scheduled
This is the dispatcher-effort metric. High same-day rescheduling means the day was assembled in a way that did not survive contact with reality, and someone is spending their morning rebuilding it. It is one of the better early indicators that booking, not dispatch, is the problem.
6. Overtime hours as a share of payroll
Chronic overtime in a company that is not growing is usually a scheduling artifact, not a demand signal. Days that were infeasible when booked get resolved by running crews past the end of the shift.
Group 3: Is the work being done right?
7. First-time fix rate
Jobs resolved on the first visit ÷ total jobs
Every failed first visit is a second drive, a second appointment slot and an unhappy customer — negative capacity, not neutral. Worth segmenting by job type: a low rate concentrated in one service usually means a parts or diagnosis problem, not a technician problem.
8. Callback rate within 30 days
The quality counterweight to any productivity push. If jobs per day rises and callbacks rise with it, you have not gained capacity — you have borrowed it and will repay with interest.
Group 4: Is it worth anything?
9. Revenue per technician per day
The number that ties the rest together. Utilization can improve while revenue per technician stays flat, if the recovered hours went into low-value work. Reviewed monthly, this keeps the operational metrics honest.
All three would say "our schedule is full and we are not making money." The KPI set tells you they need three different interventions.
On benchmarks: resist the urge to compare these against published industry figures. Trade, territory density, job mix and how you define "available hours" all move the numbers more than performance does. Your own trend, measured consistently, is worth more than any external comparison.
Making them actually get used
A few practical notes from watching these get implemented badly.
Publish the definitions next to the numbers. Half the arguments about KPIs are definitional. Write down what counts as available hours, what counts as a callback, what window counts as adherent — and put it where people read the dashboard.
Segment by technician and by territory, not just company-wide. Averages hide the cases worth acting on. One territory with much higher travel share is a specific, fixable finding; the company average containing it is not.
Review the operational four weekly. Utilization, travel share, adherence and same-day reschedules respond to scheduling changes within days. Monthly review is too slow to tell whether a change worked.
Do not turn any of them into individual quotas. Job count and first-time fix in particular are easy to game in ways that cost more than they gain.
What to take away
- Track productive utilization and travel share together — one is meaningless without the other.
- Schedule adherence is the honesty check on arrival windows; watch it whenever you narrow one.
- Same-day reschedule rate is the clearest early signal that the problem is in booking rather than dispatch.
- First-time fix and callback rate keep productivity gains honest; capacity bought with rework is not capacity.
- Read the metrics as a set — the pattern between them names the cause, where a single number cannot.
- Compare against your own trend, not published benchmarks from companies doing different work.
Common questions
What are the most important field service KPIs?
Productive utilization, travel time as a share of the shift, first-time fix rate, schedule adherence, and revenue per technician per day cover most of what matters. Together they distinguish a scheduling problem from a demand problem from an execution problem, which is what a KPI set is for.
What is schedule adherence in field service?
The percentage of appointments where the crew arrived inside the promised window. It is the honest counterweight to narrowing arrival windows — a narrow window with poor adherence is worse for customers than a wider window you reliably hit.
How often should field service KPIs be reviewed?
Weekly for the operational metrics like utilization, travel share and adherence, since those respond to scheduling changes within days. Monthly or quarterly for revenue and margin metrics, which move too slowly to read weekly without mistaking noise for signal.