# How to track technician performance metrics in Salesforce First-time fix, travel, and utilisation against hours that were actually available. A count of jobs rewards the person who closed the easy ones. **Published:** 2025-12-31 **Canonical:** https://synconai.com/track-technician-performance-metrics **Author:** SynconAI Architecture Team ## In short Measure first-time fix, travel, and utilisation against the hours a technician was actually available. A count of jobs rewards short calls and ignores the hard one. Define each metric, and the hours under it, before you build the report. The hub is Agentforce Field Service and Operations, formerly Field Service. ## Sources - **Agentforce Field Service and Operations** (retrieved 2026-10-10) https://help.salesforce.com/s/articleView?id=service.fs_overview.htm&type=5 > Agentforce Field Service and Operations (formerly known as Field Service) gives you a powerful, highly customizable, mobile-friendly field service hub in Salesforce. - **Agentforce Field Service and Operations** (retrieved 2026-10-10) https://help.salesforce.com/s/articleView?id=service.fs_overview.htm&type=5 > A scheduling optimizer that allocates resources to appointments in the most efficient way possible by accounting for worker skill level, travel time, location, and other factors. ## What to take away - Define first-time fix, travel, and utilisation in writing before anyone builds a report. A column with no definition will be argued, then ignored. - Divide by available hours: the shift minus leave. A nominal five-day week flatters people who were not there, and punishes part-time crews. - Job count rewards short work. It will push behaviour toward easy appointments and away from the job that should have been finished properly. - If durations are false, travel and utilisation are false. The optimizer is using those same durations. ## The guide The national league table put the rural technician last. He had finished the hard faults and driven the long patch the territory forced on him. The city technician at the top had closed five short calls and left a repeat for the following week. The bonus followed the table. The behaviour followed the bonus. By the next quarter the difficult sites were being handed to whoever still cared, and the report said productivity was up. People still search for technician metrics in Salesforce. The hub they mean is named here: "Agentforce Field Service and Operations (formerly known as Field Service) gives you a powerful, highly customizable, mobile-friendly field service hub in Salesforce." The hub will store the appointments. It will not choose your definition of a good technician. That choice is yours, and it has to be made before the report. The scheduler beside it is specific: "A scheduling optimizer that allocates resources to appointments in the most efficient way possible by accounting for worker skill level, travel time, location, and other factors." Travel time and duration sit under both the plan and the score. If you measure people on numbers the plan is inventing, you will manage fiction. ## Three metrics, written down First-time fix: the job did not need a return visit for the same fault inside a window you name. Fourteen days, thirty days, whatever you can defend to a process owner. Write it down. Do not change it in the middle of a quarter because a region looked bad. Exclude work that was always going to be two visits, such as "inspect, then return with the part," or you will punish the person who followed the plan. The window is a business sentence, not a chart setting. A return inside the window for the same fault counts against the metric. A return the customer booked for different work does not. A second visit you scheduled on day one, because the part was never going to be on the van, does not. If you cannot say which of those three a given repeat is, the metric is not ready to be shown to anyone who will be paid on it. Travel: the time you actually record between leaving one job and starting the next, or the travel figure you have decided is the source of truth. Say which. Do not mix a guess from the schedule with a time the technician typed, and then rank people on the blend. The optimizer accounts for travel time when it builds the day. A metric that uses a different travel number will scold people for a route the product was told to build. Pick the source and keep it for the quarter. If technicians record travel on the appointment, that is the source, and the blanks are a data problem you fix, not a reason to fall back to the planned figure for the people who left it empty. Falling back rewards the blank. The person who records a long drive looks worse than the person who records nothing and inherits a tidy plan. How territory shape and the first job create that drive, and why this page will not promise a percentage reduction, is covered in [reducing technician travel time](/reduce-technician-travel-time-field-service). Measure the travel you record. Do not invent a second travel number for the scoreboard. Utilisation: time on the job divided by hours the person was available. Available means the operating hours you stored, minus leave, minus the days a contractor was not contracted. It does not mean forty hours because that is a round number. A part-time pattern divided by a full week looks lazy. A person on leave divided by a full week looks lazy. Neither was available. The denominator is the shift you could actually have scheduled. ::: note Write one sentence per metric before anyone builds a report: the numerator, the denominator, the window, and who is excluded. If the report cannot be read back as that sentence, the report is wrong. The sentence stays. ::: ## Why a job count fails A count of completed jobs makes five short calls look better than one hard failure that was fixed once. People will notice. They will split work, avoid the long serial numbers, and leave the difficult site for whoever still cares. You will then hold a meeting about culture. The metric caused it. Splitting is easy to miss if you only watch the total. One fault becomes two work orders, each closed the same day, and the count doubles. First-time fix may even look fine if you defined it badly, because neither "half" had a return. Read a sample of closed jobs with a dispatcher who knows the work. If two closures are one visit, the count is fiction, and paying on it will create more of them. If you need a volume number, keep it as context beside first-time fix, not as the ranking. Say so in the report title, or it will become the ranking anyway. People manage the column that is sorted. Put first-time fix in that column. Put volume in a column nobody is sorted by, and revisit the choice if a manager starts screenshotting it. Do not compare a rural territory with a dense city on travel or on jobs per day. The optimizer is accounting for location and travel time. The rural day contains more road. Rank people inside a territory, or adjust for it in a way you can explain in a sentence. A national league table is how you punish the person who drew the long patch. It is also how you promote the person whose patch was drawn small. Neither result is performance. Both are geography wearing a score. ## The hours under the number Available hours fail in the same places the optimizer fails, because they are the same records. A contractor stored on a five-day shift will show low utilisation on the two days they were never coming, or they will be scheduled onto those days and then marked as a no-show. Fix the hours. Then calculate. A branch stored as open on a public holiday will show a yard full of missed appointments. That is not a workforce problem. It is a calendar you refused to update. Leave has to come off the denominator in the same period the report covers. A person who took a week off, still divided by the full month, is a distorted score. So is a person who worked overtime you did not store: the numerator grew and the denominator did not, and they look like a hero or a policy breach depending on who is reading. Decide whether overtime counts as available hours. Write it in the sentence. Apply it to everyone. Duration sits under utilisation as well. Time on the job is only as honest as the duration you trust. If every job is stored as one hour and the lived work is longer, utilisation looks tidy and the day looks late. If technicians inflate the duration to survive the metric, you are paying for the inflation. The scheduling optimizer is using those same durations to build the next day. A false duration poisons the plan and the score together. Correct the work type from watched jobs before anyone is paid on the ratio. ::: warn Do not pay a bonus on utilisation while durations are still a guess. You are paying people for agreeing with a bad number, and the optimizer will keep stacking the day from that number. Fix the duration. Then open the report. ::: ## Build the report after the sentence | Metric | Numerator | Denominator or window | Exclude | | --- | --- | --- | --- | | First-time fix | Jobs with no return for the same fault | The window you wrote down and will keep for the quarter | Work that was planned as two visits | | Travel | The travel figure you named as the source of truth | Per person, inside one territory | A blend of planned travel and typed travel | | Utilisation | Time on the job, from durations you have checked | Hours available: stored shift, minus leave, minus days not contracted | A nominal full week | | Job count | Closures | Nothing you should rank on | Use it as context, or do not show it | Check a handful of people by hand against the sentence before the report is published. If the hand count and the report disagree, the report is wrong. Publishing it and promising to fix the definition later trains the crew to ignore the number, or to game the version that went out. Either habit is expensive to reverse. Review the definitions twice a year. Change them in a new quarter, with the old number left in place for the old quarter. Quiet edits to a live metric are how trust in the report ends. A region that looked bad under the old window and fine under the new one has not improved. It has been rebased. Say that in the meeting where the window changes, or the rebase will be told as a success. Decide who is in the report before you argue about the bars. A contractor on a three-day agreement is not comparable with a full-time employee on a raw job count, and they are comparable on first-time fix if the window and the exclusions are the same. A trainee who must not attend alone should not sit in a ranking built for people who close work unsupervised. Exclude them in the sentence, or give them a separate view. Dropping them into the national table and then explaining the footnote in the meeting means the table already did the damage. Supervisors who also attend need a decision in the same sentence. If their office hours count as available, a day of meetings looks like idle time. If only hours on the tools count, say that, and do not let the report scold them for a planning day you asked them to run. Mixed roles are how a sensible utilisation figure becomes a weapon. Split the view, or exclude the days that were never field capacity. The same discipline applies when a territory is split or a depot opens. The old quarter stays on the old fence. The new quarter says the fence changed, so travel moved for a reason you can name. Quietly restating history under the new fence makes a redesign look like a personal improvement, or a personal failure, depending on who changed patches. Leave the old number alone. Field Service Lightning is the search. The hub is Agentforce Field Service and Operations, formerly Field Service. It will total whatever you ask. It will not stop you ranking a job count, comparing a rural patch with a city, or dividing by hours nobody worked. Those are decisions. Make them in writing, in the sentence, before the column exists. ## Questions ### Which technician metrics are worth tracking? First-time fix, travel, and utilisation against hours the person was actually available. Define each one, including the window and the exclusions, before you build the report. A count of jobs rewards short calls and should not be the ranking. If you need volume, keep it as context beside first-time fix, and say so in the title or it will become the sort order. Rank people inside a territory. A national table punishes the long patch and flatters the short one. ### What is utilisation divided by? Hours available: the shift you stored, minus leave and minus days a contractor was not engaged. Dividing by a nominal full week punishes part-time patterns and hides absence. It also makes a person on a public holiday you forgot to close look like a no-show. Decide whether overtime and mixed office days count, and put that decision in the same sentence as the metric. Then build the report to the sentence. If the report cannot be read back as the sentence, the report is the thing that is wrong. ### How should first-time fix be defined? No return visit for the same fault inside a window you write down and keep for the quarter. Exclude jobs that were planned as two visits, such as inspect and then return with the part, or you punish the person who followed the plan. A different fault is not a miss. Changing the window because a chart looks bad makes the metric useless, and a quiet edit mid-quarter destroys trust in it. Change definitions at a quarter boundary and leave the old number on the old quarter. ### Why would the scheduling optimizer matter to a performance report? It accounts for travel time, skill, location, and other factors when it builds the day. If your metric uses a different travel figure, or a duration the optimizer was fed as a lie, you are scoring people on a plan you distorted. A short duration makes utilisation look tidy and the lived day look late. Fix the work type from watched jobs before anyone is paid on the ratio. Otherwise the bonus rewards agreement with a bad number, and the next run stacks the same impossible day. ### What is the Salesforce product called? Agentforce Field Service and Operations, formerly known as Field Service. It is the field service hub in Salesforce. Field Service Lightning is the older search name. The hub will store the appointments and total whatever you ask. It will not choose the definition of a good technician, and it will not stop you ranking a job count. Those sentences are yours; write them before the column exists. --- SynconAI is a Salesforce Select Partner and an OpenAI Select Partner, delivering architect-led Salesforce and AI programmes across the United States and Australia. Related service: Field Service (https://synconai.com/salesforce-field-service).