Measuring more does not always help you manage better. In a maintenance, installation or technical support company, a dashboard full of data can obscure what matters: whether we respond quickly, meet our commitments, resolve issues properly and make money on every job.
Maintenance KPIs and technical service indicators should help you make decisions, not monitor the team or decorate a presentation. Before calculating them, it is worth agreeing when each stage of the job begins and ends, using statuses consistently and recording times according to the same criteria. If each person interprets ‘resolved’ differently, the result will not be comparable.
These ten technical service metrics provide a practical foundation. There is no universal target value: an urgent industrial air-conditioning call-out, a preventive inspection and a domestic repair all have different contexts. The useful approach is to segment, monitor trends and act on the causes.
1. Response time
This measures how much time passes from the moment a request comes in until the company provides the first useful response: confirming receipt, asking for necessary information or proposing the next step. An automated acknowledgement alone does not always amount to an operational response.
Simple calculation: add up the response time for all requests during the period and divide it by the number of requests. Include the median alongside the average and review the oldest cases separately, as a few long waits can distort the result.
How to interpret it: compare it by channel, priority, customer type and time of day. If it worsens only at certain times, the problem may be coverage. If it worsens for a particular type of call-out, there may be no clear classification criteria. Define the commitment that makes sense for each service and measure performance against that commitment.
2. Resolution time
This is the time between receipt of the request and its operational resolution. It should include the actual waiting periods within the process, but it is helpful to distinguish between periods that depend on the company and those that depend on the customer, a spare part or a third party.
Simple calculation: resolution date and time minus opening date and time. Analyse the median and the distribution by job type; mixing brief inspections with complex repairs produces a figure of little use.
How to interpret it: an increase may indicate insufficient capacity, incomplete diagnoses, waits for materials or too many handovers. Do not try to reduce it by closing jobs too early. The definition of ‘resolved’ should require the work to be completed and documented in accordance with your process.
3. First-time fix rate
This indicates the proportion of jobs resolved during the first visit, without a second journey for the same cause. It is one of the most useful technical service indicators because it links diagnosis, preparation, materials availability and training.
Simple calculation: jobs resolved on the first visit divided by jobs that required a visit, multiplied by 100. Exclude or separate services that, by design, require several stages.
How to interpret it: if it falls for a family of equipment, review the information collected when the call-out was logged, the checklists and the materials prepared. It may also reveal training needs. Do not penalise a justified second visit: the aim is to avoid preventable repeat visits, not to force rushed solutions.
4. Visit punctuality
This measures the percentage of visits started within the time window promised to the customer. To calculate it properly, record the planned time, the window communicated and the actual arrival or start time.
Simple calculation: visits started within the agreed window divided by visits completed, multiplied by 100. If you work with broad time windows, also retain the number of minutes early or late to understand the spread.
How to interpret it: segment by area, route, first visit of the day and service type. Delays concentrated after certain jobs may indicate unrealistic planned durations; if they appear by area, it may be necessary to review route planning. Punctuality does not mean rushing, but rather promising reasonable windows and communicating any changes.
5. Cancellations and rescheduling
This groups together visits that are cancelled, where the customer cannot be reached or that are moved after capacity has been reserved. It is worth recording a standardised reason: customer absent, requested change, missing materials, technician unavailable, incorrect planning or another verifiable cause.
Simple calculation: cancelled or rescheduled visits divided by scheduled visits, multiplied by 100. Also calculate how many were rescheduled without losing the slot and how far in advance the change occurred.
How to interpret it: separate controllable causes from external ones. A high number of absences may justify reviewing confirmations and customer notifications; frequent internal rescheduling usually points to overload, unprepared materials or late assignments. The metric should encourage prevention, not conceal necessary changes.
6. Productive time versus travel time
This compares the time spent carrying out jobs with the time spent travelling. It is not intended to treat travel as useless—it is part of the service—but to show how much it affects available capacity.
Simple calculation: working time on visits divided by the sum of working time and travel time, multiplied by 100. Keep both figures in hours as well. Record breaks and administrative tasks separately so that they are not attributed to travel.
How to interpret it: review trends by area, day and technician, bearing in mind that rural areas or specialist services may require longer journeys. If travel time increases, try grouping visits geographically, adjusting territories or improving the sequence. Good route optimisation aims to reduce avoidable journeys without compromising priorities or schedules.
7. Actual cost per job
This shows what each job consumes, beyond the materials invoiced. It should include allocated labour, travel, materials, subcontracting and a consistent allocation of overheads where relevant.
Simple calculation: add together the direct costs of the job and the share of overheads defined by the company. For labour, multiply the hours recorded by the internal hourly cost; for the vehicle, apply your chosen time- or mileage-based method without mixing methods between periods.
How to interpret it: compare equivalent jobs and review deviations from the budget or expected cost. A high cost may arise from an insufficient estimate, a second visit, urgent purchases or incorrectly recorded time. Measuring it helps improve preparation and set prices without giving your work away.
8. Margin per job or contract
The margin shows what remains after covering the costs associated with the service. It can be analysed by job, customer, contract or type of maintenance, always using the same accounting definition.
Simple calculation: job revenue minus the actual cost of the job. To express it as a percentage, divide that margin by the revenue and multiply by 100. If there is not yet a final invoice, label the figure as estimated and update it on closure.
How to interpret it: avoid reading it in isolation. A job with a low margin may be part of a profitable contract or covered by a warranty; one with an apparently high margin may conceal unallocated hours. Look for patterns and review scope, price, purchasing and repeat visits before making a decision.
9. Backlog age
The backlog is the set of open jobs that still require action. Its age helps identify forgotten call-outs and bottlenecks that the number of outstanding jobs alone does not reveal.
Simple calculation: for each open job, subtract the date it was received from the current date. Then group jobs into age bands defined by your operations and calculate the median. Add the time in the current status to identify where the flow is stalling.
How to interpret it: review the oldest cases first and classify the cause: waiting for the customer, materials, a quotation, assignment or completion. A growing backlog does not always mean a shortage of technicians; it can also indicate duplicate jobs, poorly maintained statuses or pending closures. A well-defined flow of statuses and processes makes this metric reliable.
10. Post-service quality: satisfaction, complaints and recurrence
Quality cannot be captured in a single survey. It is worth considering three signals together: customer ratings, complaints relating to closed jobs and new incidents caused by the same issue within a period appropriate to the service.
Simple calculation: use a consistent satisfaction question and calculate the average or distribution of responses; divide complaints attributable to the service by completed jobs; and divide repeat incidents caused by the same issue by comparable closed jobs. Keep response rates visible and record the reason, not just the score.
How to interpret it: look for correlations. If repeat incidents increase but satisfaction remains stable, the survey may be arriving too early. If complaints rise without a repeat technical issue, review communication, punctuality or documentation. Read the comments and contact those involved in critical cases: the data provides direction, but the cause usually lies in the context.
What to review each day, week and month
Daily, the dashboard should help drive action: unanswered requests, visits at risk of delay, cancellations, blocked jobs and the oldest outstanding items. This is a brief operational review focused on exceptions and specific owners.
Each week, bring the team together to review response and resolution, first-time fixes, punctuality, rescheduling and the balance between productive and travel time. Compare similar jobs, identify two or three recurring causes and agree an action with an owner and a date. Changing ten things at once makes it impossible to know what has worked.
Each month, add cost, margin, backlog age and post-service quality. Review trends by service, customer and area, check the quality of data entry and adjust targets when the mix of work changes. Do not turn the target into an immovable figure: document what you want to improve, your starting point and under what conditions.
From data to a specific decision
A technical service metrics system works when each indicator has a definition, a source, an owner and an associated decision. Start with data that is reliable enough, not perfect. After a few weeks, you will be able to identify which fields are missing, which statuses are used inconsistently and which segments provide context.
Centralising jobs, visits, statuses, times and job sheets reduces the need to reconstruct activity manually. Discover how enrutar can help you organise operations and turn them into actionable information. The tool is only one part of the solution: value emerges when the team reviews the indicators regularly and turns each finding into a measurable improvement.