Why does average ticket look wrong when jobs take more than one visit?
Because the revenue lands on one date and the work spans several. If your report counts visits, average ticket collapses; if it counts jobs, revenue appears on the invoice date rather than the day the work started. Pick one unit — the job, defined from first dispatch to final invoice — and use it everywhere, including in marketing attribution, where visit-level counting badly distorts cost per job.
The unit problem
Multi-visit work makes the field service data model visible. A replacement might be a sales visit, an install day, and an inspection. Depending on configuration, that is one job with three appointments, three jobs linked by a parent, or three unrelated jobs that a human knows are related.
Each configuration produces a different average ticket from identical work. Divide one invoice by three visits and your average ticket looks like a third of reality. Count three jobs and your job count triples while two of them show no revenue at all.
Define the job and hold the line
The workable definition for most operators is a job as a continuous customer engagement: opened at first dispatch for a given complaint at a given property, closed when the final invoice for that engagement is issued. Visits are children of the job. Revenue belongs to the job. Dates belong to both — the job has a start date and a close date, and reports should be explicit about which one they use.
Then apply the rule everywhere. Average ticket is revenue per job. Cost per job is media spend divided by jobs, not by visits. Booking rate is jobs created divided by opportunities.
Where the two dates matter most
- Marketing attribution. Credit belongs at the start date, near the lead, not at the invoice date weeks later.
- Revenue recognition. Finance wants the close date. This is a legitimate difference, not an error — label each report with the date basis it uses.
- Capacity analysis. Needs the visits, because capacity is consumed by visits, not by jobs.
- Technician credit. Ambiguous when three techs touched one job. Decide the split rule in advance and publish it.
The month-end distortion nobody catches
Jobs that start in one month and invoice in the next move revenue across the boundary. In a business with long installs, a strong month of lead generation can look like a weak month of revenue, and the marketing budget gets cut for a result that had not been billed yet.
The fix is to report both bases side by side during any period where the two diverge — revenue by job start date next to revenue by invoice date. When they disagree sharply, that is information about your cycle time, not an error to reconcile away. Joining lead data to job start dates is exactly the stitching that revenue intelligence performs, and it usually requires reading the job and invoice objects together through the platform's API rather than relying on a single canned report.
Topics: average ticket · attribution · data model · reporting
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.