How Should a Service Business Measure Technician Performance?

The leaderboard on the office wall says the top technician produced $48,000 last month. It does not say that he also produced the highest callback rate on the team, sold 31 memberships of which 19 canceled within 90 days, and turned every no-cool call into a replacement quote whether the system needed one or not. The owner is about to give him a raise. The customers who will not call back are not on the leaderboard.

A service business runs on the decisions its technicians make in the customer's home, and revenue alone is a poor way to measure those decisions. This article lays out six measures that together describe a technician's performance, shows how to combine them into a scorecard with weights, and covers the ways scorecards go wrong: gaming, seasonality, dispatch bias, and small numbers. The scorecard and the targets in it are illustrative. Your own targets should come from your own trailing data, not from a table on a website.

The six measures that matter

Billable efficiency

Billed hours divided by paid hours. A technician paid for eight hours who bills 5.2 of them runs at 65 percent. The rest is drive time, parts runs, the shop, training, the canceled call, and the second trip because the first one did not carry the part. Efficiency is the measure that ties most directly to gross margin, because every unbilled hour is paid at the full loaded rate and recovered from nobody. It is also the measure that dispatch and inventory influence most, so a low number is not always the technician's fault.

Revenue per truck per day

Total revenue produced divided by days the truck ran. This is the best single number for capacity planning because it combines efficiency, ticket size, and conversion into one figure you can compare with the fully loaded daily cost of putting that truck on the road. If a truck costs $900 a day in wages, burden, vehicle, fuel, insurance, and its share of dispatch, and it produces $1,400, you know what the next truck has to do before you buy it. Track it by technician and by truck, since technicians move between trucks and trucks move between technicians.

Average ticket

Revenue divided by completed jobs, computed separately for service, install, and maintenance. A blended average ticket is meaningless because one replacement can equal 30 repairs. Watch the service average ticket for the diagnostic-only share: a technician who closes 40 percent of calls at the diagnostic fee is either finding nothing wrong on a lot of systems or not presenting options. Both are worth a conversation, and they are different conversations.

Conversion rate

Jobs sold divided by opportunities presented. For a service technician this has two parts: repair conversion, the share of diagnosed problems the customer approved on the spot, and replacement lead turnover, the share of aging systems where the technician generated a quote or a comfort advisor visit. Track them separately. A technician with high repair conversion and no replacement turnovers is fixing 16-year-old systems that will fail again in August. A technician with the opposite pattern may be quoting replacements to customers who needed a capacitor.

Callback rate

Callbacks within 30 days divided by completed jobs, where a callback is a return trip on the same system for a related problem, at no charge to the customer. This is the quality measure and the counterweight to every revenue measure on the list. It requires a callback job type or tag in the field service system and a dispatcher who uses it consistently, which is the weak point in most companies. A callback rate that is zero every month is not a great team. It is a tagging problem.

Membership sales and retention

Memberships sold divided by eligible calls, paired with the share of those memberships still active 90 days later. The pairing is the whole point. A technician who sells 40 memberships a month that all cancel within 60 days has discovered a way to be paid twice for the same handshake. Measured with retention, membership sales become a proxy for trust, which is the thing the customer is actually buying.

Building the scorecard

A scorecard converts each measure into a score against a target, weights the scores, and sums them. The weights encode what the company values. Heavier weight on efficiency and callbacks and lighter weight on membership sales produces a different technician over 12 months than the reverse. The table below is an illustrative starting layout for a residential service technician. Install crews need a different card with job margin, hours versus estimate, and inspection pass rate in place of conversion and membership sales.

Score each measure from 0 to 100 against its target, with 100 at target and a floor at some reasonable distance below it, then multiply by the weight. A technician at target on everything scores 100. Show three columns for each technician: this month, trailing three months, and the team median. The trailing figure smooths out a slow week. The team median tells the technician where he actually stands, which the raw number never does.

Cadence matters as much as content. Efficiency and revenue per truck can be reviewed weekly in five minutes. The full scorecard is monthly, reviewed one-on-one with the service manager, with a specific coaching item agreed for the following month. Compensation reviews are quarterly and use the trailing three-month card, which removes the temptation to have one enormous month before the review and one quiet month after it.

Pitfalls

Gaming

Any single measure can be gamed, and a good technician will find the way within a month. Revenue targets alone produce over-selling and callbacks. Conversion targets alone produce technicians who present only the repairs they know will close. Efficiency targets alone produce padded hours and rushed work. The defense is pairing: revenue with callbacks, conversion with average ticket, efficiency with callback rate, membership sales with retention. A scorecard with paired measures is harder to game than the technician is motivated to try.

Seasonality

June in Phoenix is not February in Phoenix, and neither is comparable to the same month in Minneapolis. Revenue per truck per day, average ticket, and conversion all move with the weather. Compare each technician with the team for the same period, and compare the technician with his own performance in the same month a year earlier. A scorecard that shows every technician failing in the shoulder season is measuring the calendar.

Dispatch bias

The dispatcher decides who gets the 15-year-old system with the failed compressor and who gets the thermostat battery. If replacement leads always go to the same two technicians, their conversion and revenue numbers will lead the board regardless of skill. Measure opportunities per technician, not just outcomes, and review the dispatch pattern monthly. Some dispatch bias is deliberate and correct. Undisclosed bias makes the scorecard a record of the dispatcher's preferences.

Data hygiene

Miscoded job types, callbacks logged as new service calls, timesheets approved a week late, and memberships sold under the office's name instead of the technician's all corrupt the card. The scorecard is only as good as the field service system behind it, and the system is only as good as the dispatcher's habits. Assign one person to audit a sample of jobs every week. Twenty jobs, 15 minutes, and the card stays honest.

Small numbers

A technician with 12 completed jobs and one callback shows an 8 percent callback rate. The same technician with 60 jobs and one callback shows under 2 percent. Do not draw conclusions from a single month of a low-volume measure. The trailing three-month column exists for this reason, and callback rate in particular should be judged over a quarter or a rolling 100 jobs.

Every measure on the card should have a partner that catches the obvious way to cheat it. If you cannot name the partner, the measure is not ready to drive pay.

Tying it to pay, carefully

Once the scorecard has run for three months and the data has been audited, it can inform incentive pay. The principle is to pay on the composite score or on paired measures, never on revenue alone. A spiff for memberships sold should net out cancellations within 90 days. A bonus for revenue per truck should be forfeited above a callback threshold. The technician who reads the plan and finds no shortcut in it will do the one thing the plan was designed to produce, which is take good care of the customer in front of him.

Where a fractional Controller fits

The hard part of a technician scorecard is not the design. It is producing it every month from data that ties to the books. Revenue by technician has to reconcile to the income statement. Paid hours have to come from payroll, not from memory. Callbacks have to be tagged, membership cancellations have to be captured, and the burden rate behind the daily truck cost has to be current. A fractional Controller builds that pipeline: job type and tag standards in the field service system, a monthly reconciliation of system revenue to the general ledger, an unapplied labor line in cost of sales, and a scorecard the service manager receives on the third business day without asking.

A fractional CFO then uses the same numbers for the larger questions: when to add a truck, which technicians justify a comfort advisor behind them, what the membership book is worth, and how the summer will fund the winter. The scorecard is a management tool for the service manager first. It becomes a planning tool for the owner when the numbers on it are the same numbers the bank sees.