Ask how your dispatcher is doing and you'll probably get the company's on-time rate. That number belongs to the whole operation. A late arrival can be caused by a slot booked too tight, a job estimate that was 30 minutes short, a tech who left late, or a crash on the highway. The dispatcher only owns some of that.
If you judge the person on the board by numbers they don't control, two things happen. Good dispatchers get frustrated, and nobody fixes the real cause. This guide builds a scorecard that separates the dispatcher's part from everyone else's.
Start with who controls what
Every job passes through four hands before it is done. Each one can break it.
| Stage | Who controls it | What goes wrong |
|---|---|---|
| Booking | CSR, online booking, sales | A slot that was never reachable, a wrong job type, missing details |
| Estimate | Whoever sets durations | A 60-minute slot for a 90-minute job |
| Dispatch | The dispatcher | Wrong tech, poor sequence, slow reaction to changes |
| Field | The tech | Late start, long lunch, misdiagnosis |
The dispatcher's KPIs should come from the third row. Metrics from the other rows still matter. They just go on someone else's scorecard.
The dispatcher scorecard
Five metrics cover what a dispatcher actually decides. Each has a formula you can run from timestamps you probably already have.
1. Reaction time to schedule changes
When a job runs long, a tech calls in sick or an emergency lands, how fast does the board get fixed?
Reaction time = Time the board is updated − Time the change was known
Measure it on the events that matter: a late-running job, a cancellation, a same-day emergency. The aim is that affected customers hear before they start wondering.
2. Reassignment success
When a job has to move, does it land on someone who can actually make it?
Reassignment success = Moved jobs that arrived on time ÷ Total moved jobs
A dispatcher who moves jobs to whoever is free, without checking the drive, scores low here. One who checks where the receiving tech will be at that hour scores high.
3. Skill match rate
Skill match = Jobs sent to a tech qualified for that job type ÷ Total jobs dispatched
A job sent to the wrong tech often ends in a second visit with the right one. Count only jobs the dispatcher assigned or reassigned, not jobs booked straight to a tech by the online booking flow.
4. Drive added by dispatch decisions
This one needs care. Total drive time is mostly a booking and territory outcome. The dispatcher's part is the drive added or removed when they change the plan.
Dispatch drive delta = Drive time after dispatcher changes − Drive time as booked
A negative number means the dispatcher saved drive. A positive number on a day full of emergencies may be fine. Read it alongside the reason for each change.
5. Customer contact on changes
Contact rate = Moved or late jobs where the customer was told before the original time ÷ Moved or late jobs
Customers forgive a moved appointment far more readily than a silent one. This is entirely in the dispatcher's hands unless texts go out automatically, in which case drop it from the scorecard.
The metrics to keep off the dispatcher's card
These are good company KPIs. They are bad dispatcher KPIs, because the dispatcher is downstream of the cause.
- Overall on-time arrival rate. Split it first. The method in how to measure on-time arrival separates late from the estimate, late from the drive and late from the booking.
- Jobs per tech per day. Set mostly by what was sold and where.
- Total drive time. Set mostly at booking. Only the delta belongs to dispatch.
- First-time fix rate. Mostly booking (parts, duration) and the tech (diagnosis). Skill match is the dispatcher's slice.
The full company list lives in field service scheduling KPIs.
A worked example
Illustrative numbers for one dispatcher over one week.
On a blended scorecard, this dispatcher looks responsible for 40 late jobs. On a split one, they own 6. The other 30 point at booking and durations, which is where the fix is. If you book into slots that were never reachable, no dispatcher can save them. That is the argument for checking the drive at booking time, covered in travel-feasible scheduling.
How to collect it without a data team
You need four timestamps per job: booked start, on-my-way, arrived and done. Add a reason code whenever a job is moved. That is enough for every formula above.
Two habits make it work:
- Tag every move with a reason. Running long, sick call, emergency insert, customer request. One tap or one word.
- Freeze the booked plan. Keep the day as it was booked so you can compare it with the day as dispatched. That comparison is schedule adherence, split by who changed what.
A tech phone app that records on my way, arrived and done gives you the timestamps without anyone typing them. CrewLink's tech app records those taps and its Reports page shows on-time rate with CSV exports, so you can run the split yourself.
Read it monthly, coach it weekly
Weekly numbers for one dispatcher are noisy. One bad storm day swings everything. Review the scorecard monthly and use the weekly list of moved jobs for coaching: pick three moves, look at what the dispatcher knew at the time, and talk through the alternative.
What to take away
- The company on-time rate is not a dispatcher KPI until you split it by cause.
- Score the dispatcher on reaction time, reassignment success, skill match, drive delta and customer contact.
- Keep total drive time, jobs per day and overall FTFR on the company scorecard.
- Four timestamps per job plus a reason code on every move is all the data you need.
- If most lateness traces back to booking, fix booking. The dispatcher can't.
Common questions
What KPIs should a dispatcher be measured on?
Measure what the dispatcher decides: how fast they react to changes, whether moved jobs still arrive on time, whether jobs go to qualified techs, the drive their changes add or save, and whether customers hear about changes in time.
Is on-time arrival a dispatcher KPI?
Only the part the dispatcher caused. Split late arrivals into booking, estimate, dispatch and field causes, and put only the dispatch share on the dispatcher's card.
How many techs can one dispatcher handle?
It depends on job length, how many changes a day brings and how much the booking already gets right. Track reaction time as you add techs. When it starts climbing, the dispatcher is at capacity.
How do you measure dispatcher performance fairly?
Compare the day as booked with the day as dispatched, tag every change with a reason, and judge the dispatcher on the changes they made, not on the plan they inherited.