Guides
A monthly KPI routine for a mobile detailing business
Eight monthly measures for a detailing van, each with its formula, its data source, the decision it changes, and the thing it cannot tell you.
What to take away
- A measure earns its place only if it can change a decision. Everything else is reporting.
- Four measures, acted on, beat twelve collected. Start with the four that touch your worst problem.
- Every measure below carries a limitation. Reading a number without knowing what it hides is how bad decisions get made confidently.
- Set your own thresholds from your own history. Borrowed thresholds describe somebody else's operation.
- Review monthly, but investigate weekly. A month is too long to notice a route falling apart.
How to define one properly
For each measure, write four lines before collecting anything: the formula, where the data comes from, who owns it, and which decision it can change. A measure with no answer to the fourth line gets dropped.
Then write a fifth line that most measurement systems omit: what this number cannot tell you. That line prevents the most common misuse, which is treating one number as a verdict.
1. Contribution per occupied hour
Formula. Price, less direct materials, fuel for the trip and payment cost, divided by hours occupied including drive, setup and teardown.
Contribution per occupied hour
- Price less materials, fuel, payment cost
- Divided by hours occupied
- Includes drive, setup, teardown
- Sourcejob records + chemical spend
Source. Job records and the materials estimate from monthly chemical spend divided by jobs.
Changes. Which services stay on the menu and which get repriced.
Cannot tell you. Whether the service brings customers who buy something else later. A low-contribution service that leads reliably to a maintenance relationship is doing a different job.
2. Drive time share
Formula. Driving minutes divided by total working minutes.
Source. Job records or the phone's own timing.
Changes. Route clustering, the minimum job value for outer zones, and eventually the service area boundary itself.
Cannot tell you. Whether the driving was avoidable. A month with one large distant commercial account looks the same as a month of scattered residential work and needs a different response.
3. Overrun rate
Formula. Jobs exceeding their planned time band, divided by all jobs.
Source. Planned time from the quote, actual from the job record.
Changes. The buffer in the schedule, and the time allowance for specific services.
Cannot tell you. Why. Diagnose by splitting it three ways: condition assessed optimistically at booking, a technique not yet trained, or an equipment limitation. The last two are addressed through the hiring and training guide and the equipment and setup guide respectively.
4. Rework visits
Formula. Return visits to correct work, divided by jobs completed.
Source. Job records. Count every return, including the ones done quietly as a favor.
Changes. Inspection sampling rate, training priorities, and sometimes the time allowed for a service.
Cannot tell you. The complaints that never became a return visit. A customer who is dissatisfied and simply does not book again is invisible here, which is why it sits alongside the rebooking rate rather than replacing it.
5. Rebooking rate
Formula. First-time customers who book again within your maintenance interval, divided by first-time customers in that period.
Source. Customer history.
Changes. Whether the follow-up process exists at all, and whether the maintenance interval you publish is the right length.
Cannot tell you. Why someone did not come back. Only asking them does that, and a short direct question a few weeks after a first visit produces more useful answers than any survey.
6. Water used per vehicle
Formula. Gallons drawn divided by vehicles completed.
Source. Tank measurement, recorded at refill.
Changes. How many jobs fit between refills, which is a hard cap on the day, and it flags a technique drifting toward a longer pre-rinse.
Cannot tell you. Whether the water was contained and disposed of properly. That is a separate check, and it carries regulatory weight rather than operational weight, which is why it appears in the audit described in the licensing and compliance guide.
7. Days to collect
Formula. Average days from invoice issued to payment received, tracked separately for residential and commercial work.
Source. Invoicing records.
Changes. Deposit policy, terms offered to commercial accounts, and how much cash you need to hold.
Cannot tell you. Whether a slow payer is a bad customer. A reliable commercial account paying slowly on agreed terms is a cash timing question, not a credit question.
8. Enquiry to booking ratio
Formula. Jobs booked divided by enquiries received, by source.
Source. Ask at the point of contact and record the answer.
Changes. Which channels continue, and whether the message is attracting people who cannot afford or do not want the service.
Cannot tell you. Whether the bookings are good ones. Pair it with drive time share, or a channel producing many cheap distant jobs will look like a success.
Choosing four to start with
| If the problem is | Track first |
|---|---|
| Busy but not profitable | Contribution per occupied hour, drive time share |
| Days ending late | Overrun rate, drive time share |
| Quality complaints | Rework visits, overrun rate |
| Cash pressure | Days to collect, contribution per occupied hour |
| Growth stalling | Rebooking rate, enquiry to booking ratio |
Pick the row that matches this quarter, act on those two for a season, then add the next pair.
Which two to track first
Problem
- Busy but not profitable
- Contribution per occupied hour
- Days ending late
- Overrun rate
- Quality complaints
- Rework visits
- Cash pressure
- Days to collect
- Growth stalling
- Rebooking rate
Track first
- Busy but not profitable
- Drive time share
- Days ending late
- Drive time share
- Quality complaints
- Overrun rate
- Cash pressure
- Contribution per occupied hour
- Growth stalling
- Enquiry to booking ratio
The monthly review
Monthly review routine
- Half an hour, same day each month
- Previous three months visible
- Did it move?
- Do we know why?
- What are we changing?
- Write the decision next to the number
Write the decision down next to the number. A review that produces no written decision produces no change, and next month you will not remember what you concluded.
Keep the underlying records in a form you can query rather than reconstruct. The financial half of that discipline is what the IRS describes as a system that clearly shows income and expenses, and the operational half has the same requirement with no regulator to enforce it.
Protecting the data these numbers live in
The records behind these measures include customer addresses, access notes and payment information, held on a phone that spends its day in a van. NIST's small business cybersecurity quick start guides and CISA's material for small and medium businesses both cover the basics that prevent most incidents: multi-factor authentication, tested backups, device locks and a plan for a lost phone.
Add one rule of your own about who may see access codes and when that access is removed, which belongs in the exit process alongside the rest of how to hire reliable staff. The tools that capture all of this are chosen using the criteria in the software and measurement guide.
Common questions
What is a good number for any of these?
There is no general answer, and any figure quoted at you describes a different operation. Set your baseline from three months of your own history and improve against it.
Should technicians see these numbers?
The ones they influence, yes, and with the definition attached. A measure used as a stick without a definition produces gaming rather than improvement.
How often should thresholds be revisited?
Once a season, and immediately after any change to route, pricing or staffing. A threshold set for a different operation stops being useful the moment the operation changes.
What if a measure never changes anything?
Stop collecting it. That is not a failure; it is the fifth line doing its job.
