What a second visit really costs
The obvious cost is the travel and the hour. The real cost is larger: the customer waits days longer with equipment that does not work, a slot that could have taken new work is consumed, the technician is demoralised, and in contract work the second visit is usually unbillable.
Multiply by the share of jobs that need one and it is frequently the largest controllable loss in a service business, and almost always larger than whatever efficiency programme is currently getting attention.
It is decided before the van moves
Technicians get blamed for first-time fix. Almost all of the determinants are office decisions. Whether the job description contained enough information. Whether the equipment's history was available. Whether the likely part was on the van. Whether the slot was long enough. Whether the right person was sent.
That is good news, because office decisions are systematically improvable in a way that individual diagnostic skill is not.
The four levers, in order of effect
Most companies find the same ranking when they measure it.
- Parts availability. The commonest reason a correct diagnosis still fails to close a job. Van stock based on what is actually consumed, plus sending the likely part where history suggests it, moves this fastest.
- Information at dispatch. Equipment make, model, serial, previous faults, access arrangements and the contact who can let them in. Arriving to search for the plant room costs more than anybody records.
- Skill matching. Sending somebody who has worked on that equipment, or that manufacturer's controls, rather than whoever was free.
- Time allowance. A job booked for an hour that needs two produces either a rushed job or an abandoned one. Both come back.
Measuring it without fooling yourself
Define it before you measure it, and write the definition down. Is a planned return for an ordered part a failure? Most companies say no, as long as the part genuinely could not have been anticipated. Is a callback within thirty days on the same fault a failure? Almost everybody says yes.
Then measure it by work type. A single company-wide figure hides everything useful: planned maintenance should be close to complete every time, emergency breakdown work on unfamiliar plant never will be, and averaging them produces a number that cannot be acted on.
The parts problem, specifically
Since parts are the largest single cause, it is worth being concrete about what fixes them. Three things, in increasing order of effort.
Stock the vans from recorded consumption rather than from what somebody decided two years ago. The equipment base changes, and a van list that is never revisited is carrying spares for machines you no longer service while missing the ones you do.
Send the likely part where the history suggests it. If a unit has failed the same way twice, the third call should not be a diagnostic visit. This requires equipment-level history, which is the recurring theme of nearly every operational improvement in this trade.
Where a part genuinely cannot be carried, make the second visit immediate rather than ordinary. A planned return booked before the technician leaves site, with the part on order and the customer told, is a very different experience from a job that goes back into the queue.
The feedback loop that makes it improve
Pick the return visits from last month and read them. Not all of them, and not as a disciplinary exercise: ten jobs, with the technician in the room, asking what would have had to be true for this to finish first time.
The answers are usually concrete and cheap. A part that should be on every van. A piece of information nobody passes on. A customer site that always needs two people. Fix those and the number moves; run a campaign about trying harder and it will not.