A job that needs a second visit does not have half the margin. It usually has none, and sometimes less than none once the work it displaced is counted.
The revenue was set at the first visit. The second adds drive time, labour and a slot that could have held another job, and adds nothing to the invoice.
The causes, in the order they actually occur
The part was not on the van. By a wide margin the most common, and the most fixable.
The diagnosis was incomplete, so the wrong work was scheduled.
Not enough time was booked, so the technician ran out of day.
The work did not hold, which is a different and more serious conversation.
Fixing the first one
Find the twenty parts that account for most of your second visits. Every trade has a short list and most businesses have never written it down.
Stock those deeply and everything else barely. See stockout risk — the carrying cost of one more common part is trivial against the cost of the visit it prevents.
Check van stock on a rhythm rather than when somebody notices. A weekly restock beats a monthly audit.
Fixing the second
Ask better questions at booking. Make, model, age, symptom. Five minutes on the phone saves an hour on the road.
Photographs before the visit where the customer can send them. Increasingly normal and enormously effective.
Record what was found, so the next visit to that address starts informed rather than from nothing.
Fixing the third
If a job type consistently over-runs its scheduled time, the fix is the schedule template rather than the crew. See technician utilisation — consistently over 110% means the work is taking longer than it is booked for, on every job.
Measure it
Multi-visit rate is the number, and the important thing is to read it by job type. Installs legitimately need several visits. Service calls should not, and the two averaged together hide the problem.