Dispatch KPIs for Service Businesses, The 2026 Monday Morning List

TL;DR

Most service businesses have more dispatch data than they use and fewer numbers than they need. The reports that get generated measure activity, which feels productive and changes nothing. What an owner needs on a Monday is a short list where every number maps to a decision: hire, re-price, change the window, fix the parts process, call people back.

As of September 2026, ten metrics cover almost everything a small to mid-size service company needs to run the week, and every one can be computed from records a dispatch system already creates. Below: each metric, what it tells you, what fools people, and the one lever.

One rule first. Do not chase outside benchmark numbers. Job mix, pricing, service area density and trade vary so much between shops that a published percentage tells you almost nothing about yours. The benchmark that matters is your own last quarter. Labor cost context for the capacity math is at the Bureau of Labor Statistics.

The ten-number Monday list

1. Jobs per tech per day

What it tells you: how much work your calendar converts into completed visits, per person, per day.

What fools people: the raw count on its own is close to useless. A tech who ran six filter swaps looks twice as productive as one who ran three heat pump installs, and the second may have produced four times the revenue. Read it alongside revenue per tech day and average job duration, and set the target by measuring rather than guessing, using the method in technician capacity planning.

The lever: job mix and drive time. Cluster the route and put long jobs on their own day. Telling techs to hurry produces callbacks, which shows up two metrics down.

2. Booked-to-completed rate

What it tells you: what percentage of the work you sold actually got done, which is the cleanest single measure of operational leakage.

What fools people: owners look at bookings and feel good. A booking is an intention, and one that dissolves cost you a slot you could have sold plus the marketing spend that produced the call. Every non-completion is one of three things: they cancelled, they were not there, or you could not get to it. Those have different fixes, and lumping them together is how shops solve the wrong problem.

The lever: confirmation and reminder discipline. Almost every avoidable non-completion is a communication failure a well-timed message would have caught.

3. No-show and cancellation rate

What it tells you: how much of your calendar is fiction. It carries the most direct dollar cost, because an empty slot is unrecoverable revenue and an unbilled hour of technician pay at once.

What fools people: treating cancellations and no-shows as one event. A cancellation two days out gives you time to refill the slot. A no-show at 2 p.m. with a truck in the driveway is a total loss. Separate who cancelled too: if your own reschedules drive the number, the problem is capacity planning.

The lever: deposits on the jobs that justify them, plus a reminder cadence that makes moving an appointment easier than abandoning it. Full mechanics in reduce no-shows in a service business.

4. First-time-fix rate

What it tells you: whether a visit solves the customer's problem. The quiet driver of profit, because a second trip for the same job doubles your cost and earns nothing extra.

What fools people: counting the job as fixed because it was marked complete. Compute it honestly: first-visit closures with no return trip for the same complaint, divided by all jobs that eventually closed for that complaint including the returns. If the tech came back Tuesday with a part, that job did not first-time-fix.

The lever: parts on the truck before dispatch. Most second trips are a missing part, not a diagnostic failure, so the fix is a pre-dispatch readiness check, as covered in job readiness and parts before dispatch.

5. On-time arrival within the promised window

What it tells you: whether you keep the one promise customers remember. Nobody recalls the diagnostic detail. Everybody recalls waiting three hours.

What fools people: measuring against the tech's own estimate instead of the window the customer was given. If your team quietly widened it on the way, the metric measures nothing. Low performance usually means unrealistic windows rather than slow techs, and the trade-offs between wide and tight are in arrival time windows.

The lever: windows built from your real job durations, plus live ETA so the customer hears from you before they start wondering.

6. Average response time, call to booked

What it tells you: how fast an inbound lead becomes an appointment. In most trades the fastest shop to answer wins the job.

What fools people: measuring speed to first contact instead of speed to booking. A fast callback ending in "someone will get back to you with a price" converted nothing, so the clock stops when the appointment exists. The worst version of this number is invisible: an unanswered call leaves no record, so confirm missed calls are captured at all.

The lever: answer every call and book on that call. Nothing else in dispatch returns as much for the effort.

7. Quote-to-booked conversion

What it tells you: whether your estimates turn into work. A sales metric hiding in a dispatch system, and usually the largest pile of unrealized revenue a shop owns.

What fools people: blaming price. Some losses are price. Most are silence, because a quote sent and never mentioned again converts at close to the rate of no quote at all. Segment it, since an emergency quote and a planned replacement run on different timelines.

The lever: a written follow-up cadence somebody owns, along the lines of quote follow-up.

8. Revenue per tech day

What it tells you: the real productivity number, and the one that decides whether hiring is the answer. Invoiced revenue divided by technician days worked.

What fools people: reading it as how hard people work. It mostly measures pricing and job mix, so use it to decide what work to sell, not who to praise. Strong revenue per tech day with a calendar full weeks out is a hiring signal. A full calendar with weak revenue per tech day is a pricing problem, and hiring multiplies it rather than fixing it.

The lever: pricing and the mix of work you accept. Route efficiency helps at the margin. Pricing moves it far more.

9. Callback and warranty return rate

What it tells you: the cost of work that did not hold. Every callback is a free truck roll, a reputational risk and a tech who could have been billing.

What fools people: not tracking it at all, which is the norm. Callbacks arrive as normal inbound calls and get booked as normal jobs, so unless someone flags the return at booking the number does not exist and the pattern stays invisible. That flag is the whole battle, and the process is in warranty and callback tracking.

The lever: documentation at the job, with photos and readings, plus a quality check on the job types that generate the most returns.

10. Review request-to-review conversion

What it tells you: whether your reputation engine turns happy customers into public proof. Reviews are the cheapest lead source you will ever have.

What fools people: counting reviews received instead of the conversion rate. Ten reviews from two hundred requests is a broken process. Ten from twenty is a good one worth scaling, and you cannot tell the difference unless you count the asks. Timing carries most of the result, and who you ask matters as much, as covered in get more Google reviews.

The lever: ask immediately, and route the ask to satisfied customers rather than everyone indiscriminately.

Vanity metrics to stop reporting

Some numbers feel like management and are not.

• Total calls received. Volume without conversion is a marketing statistic. Calls answered and calls booked are useful. Calls received is weather.
• Total jobs completed, unsegmented. Growing job count with flat revenue means you traded up in effort and down in margin.
• Utilization measured as hours on the clock. A tech can be busy all day driving. Billable hours against scheduled hours says something. Clock hours do not.
• Average ticket, alone. It moves when your mix moves. Read it beside job count.
• Cumulative review count. A lifetime total is a trophy. Reviews in the last ninety days are the operating number.

If a number has not changed a decision in three months, take it off the report.

What you get out of the box versus what needs a spreadsheet

Be realistic about tooling, because this is where these projects usually die.

GetTimePad includes basic reporting from the $79/mo Starter plan, along with customer records and cancellation tracking, which is enough to produce job counts, completions and cancellations without exporting anything. The $499/mo Agency plan adds an activity log with audit trail, the piece you need when the question is who moved that appointment and when, or whether a status was changed after the fact.

What GetTimePad is not is a business intelligence tool. There are no custom dashboard builders, no arbitrary pivot reports, no multi-year cohort analysis. Several metrics above, particularly revenue per tech day and quote-to-booked conversion segmented by job type, will want a spreadsheet you maintain weekly. For most shops that is a perfectly good answer, and typing the numbers in is what makes an owner look at them.

If you have outgrown that, the supported path is the REST API and outbound webhooks on the Agency plan at /api, which pull job, appointment and status data into whatever reporting stack you already run.

For choosing the dispatch system underneath, and therefore which metrics you can compute at all, the comparison work is in best dispatch software for service companies. Plan numbers are published at pricing.

How to actually run this

Pick five of the ten. Not all ten, not on the first Monday. Booked-to-completed, no-show rate, first-time-fix, on-time arrival and revenue per tech day cover most of what is wrong in most shops.

Write them down every Monday for eight weeks. Do not act on week one, because a single week is noise and you will chase a ghost. By week four you will see a direction, and a direction is what you act on. Then pull each number's one lever, one at a time, so you can tell what worked.

That is the whole point. If a number moves and nothing changes as a result, you did not need the number.

Frequently asked questions

What are the most important dispatch KPIs for a service business?

The ten that earn a weekly look are jobs per tech per day, booked-to-completed rate, no-show and cancellation rate, first-time-fix rate, on-time arrival inside the promised window, average response time from call to booked, quote-to-booked conversion, revenue per tech day, callback or warranty return rate, and review request-to-review conversion. Each one drives a specific decision rather than just describing the week.

How do you calculate first-time-fix rate?

First-time-fix rate is the share of jobs completed on the first visit with no return trip for the same problem, which you compute by counting jobs that closed as complete and dividing by all jobs that eventually closed for that issue including the return visits. Any job that generated a second appointment for the same complaint counts against it, and parts availability is usually the biggest cause of a low number.

What reporting does GetTimePad include, and what needs a spreadsheet?

GetTimePad includes basic reporting from the $79/mo Starter plan and adds an activity log with audit trail on the $499/mo Agency plan, which covers job counts, completions, cancellations and who changed what. It is a scheduling and dispatch system rather than a business intelligence tool, so trended custom dashboards still belong in a spreadsheet or your own reporting stack, fed by the REST API and outbound webhooks on Agency at /api.

How much does GetTimePad cost and which plan has the reporting?

GetTimePad is $79/mo Starter for one staff member with basic reporting and cancellation tracking, $199/mo Pro for up to five staff which adds GPS and ETA, Tech Mode, Stripe payments, review routing and two-way SMS, and $499/mo Agency for unlimited staff with the activity log, audit trail, multi-location support and REST API access. Annual billing gives you two months free and there is a 14-day free trial. See pricing.

Why is jobs per tech per day a misleading metric on its own?

Jobs per tech per day is misleading because it ignores job mix, so a tech running six small filter changes looks twice as productive as one doing three full installs that generated far more revenue. Always read it next to revenue per tech day and average job duration, and compare each tech against their own route type rather than against each other.

What is a good benchmark for these dispatch metrics?

The only benchmark that reliably means anything is your own last quarter, because job mix, service area density, trade and pricing vary so widely that an outside figure tells you almost nothing about your operation. Track direction over four to eight weeks, act on the trend, and treat any published industry percentage as context rather than a target.