Retention metrics live in your invoicing system, not in your analytics. GA4 tells you about sessions. Your books tell you about customers.
TL;DR Four numbers: lifetime value, repeat rate, average gap between visits, and retention cost. Calculate them from invoices. Use GA4 only for the acquisition half. Review quarterly, not weekly.
Where each number lives
| Metric | Source |
|---|---|
| Customer lifetime value | Invoicing or accounting system |
| Repeat booking rate | Job records |
| Average time between visits | Job records |
| Cost per retention | Marketing spend plus job records |
| Acquisition cost | Ad platforms plus GA4 |
| Which channel brought them | GA4, imperfectly |
GA4 is an acquisition tool. It struggles with the offline, multi-year, phone-heavy reality of service work. Do not try to force retention reporting into it.
Customer lifetime value
The simple version, which is good enough
Average value per job × average jobs per year × average years retained
Example: $180 average job, 1.4 jobs per year, 4.2 years retained = $1,058 lifetime value.
Use gross profit, not revenue, if you want the number that informs your ad budget. Revenue LTV overstates what you can afford to spend.
Segment it, because the average hides everything
- By service type. Emergency-only customers have very different LTV from maintenance customers.
- By acquisition channel. Referred customers frequently show markedly higher LTV than paid-search customers. If yours do, that changes where the budget should go.
- By membership status.
- By first-job value. Sometimes small first jobs produce the best long-term customers, which is the opposite of what people assume.
The channel comparison is the single most useful cut, because it can invert a decision made on cost per lead alone.
Repeat booking rate
Customers with two or more jobs, divided by total customers, over a fixed window.
Fix the window or the number is meaningless. Customers acquired in the last month have not had a chance to return.
Better: cohort it.
Of customers first served in 2024, what proportion booked again within 12 months?
Then compare cohorts. 2023 versus 2024 versus 2025 tells you whether retention is improving. A single blended figure tells you nothing.
Average time between visits
Underrated and highly actionable.
Calculate it per customer, then average.
What it tells you
- The gap is widening. Early churn signal, before anybody cancels.
- A customer well past their personal average. A specific, timely reason to contact them.
- Where the frequency opportunity is. If your average gap is 14 months on a service that should be annual, there is a reminder problem.
Use it to trigger outreach. Personal average plus 20 percent is a good trigger point, and it beats a blanket six-month rule.
Cost per retention
Retention spend divided by retained customers.
Retention spend includes loyalty rewards, gifts, newsletter tools, portal software, staff time on check-ins, referral payouts.
Compare it against acquisition cost. Retention should be dramatically cheaper. If they are close, either your retention spend is inefficient or your acquisition is unusually good.
A caution on the widely quoted claim that retaining is five times cheaper than acquiring. The figure circulates everywhere with no consistent primary source and the real ratio varies enormously by business. Calculate your own two numbers and compare those. Do not repeat the ratio as fact.
What to actually track
Keep it to a single sheet.
| Metric | Frequency | Where |
|---|---|---|
| Customers served | Monthly | Job records |
| New vs returning | Monthly | Job records |
| Revenue per customer | Quarterly | Invoicing |
| Repeat rate by cohort | Quarterly | Job records |
| Average gap between visits | Quarterly | Job records |
| Churn, voluntary and involuntary | Monthly | Billing |
| LTV by channel | Annually | Combined |
Quarterly for most of it. Retention moves slowly and weekly review produces noise and bad decisions.
Getting the data out
You do not need a data warehouse.
- Export jobs or invoices to CSV, with customer ID, date and value.
- Pivot by customer to get job count, total value, first and last date.
- Calculate the gaps and the averages from those columns.
- Repeat quarterly and keep the history.
A spreadsheet with four years of quarterly snapshots is more useful than any dashboard, because trend is the whole point.
Where GA4 fits
Use it for
- Which channels bring enquiries.
- Which pages precede a booking.
- Acquisition cost when paired with ad spend.
Do not use it for
- Lifetime value. It cannot see your invoices.
- Repeat purchase, for offline service work.
- Anything requiring a customer identity across years.
Bridge the two by recording the acquisition source on the customer record in your CRM at first contact. That single field is what lets you calculate LTV by channel later, and it costs one dropdown at booking.
What to avoid
- Vanity metrics. Total customers ever served tells you nothing about the business today.
- Blended averages that hide segment differences.
- Weekly retention reporting. The signal is not there at that resolution.
- Benchmarking against published industry figures. Definitions vary so widely that the comparison is usually meaningless. Compare against your own previous quarters.
Export twelve months of jobs to a spreadsheet this week and calculate the average gap between visits per customer. That one number will show you exactly where your reminder programme is leaking.
Need a pro to set up tracking? [BOOK A CALL]