The second cheque is where the business actually starts
The economics of renewal
A recurring service engagement that lasts twelve months is roughly twelve times more valuable than the same engagement that lasts one. The compounding here is brutal in both directions. A practice that signs many short engagements rarely becomes profitable. A practice with even modest sign-on rates and strong retention can be both. Retention, not acquisition, usually distinguishes practices that survive from those that do not.
What clients actually renew on
Clients renew when they can describe, in their own words, what they have received in the last month. Vague claims of behind-the-scenes activity rarely satisfy a renewal conversation. The discipline is to make the work visible: simple monthly summaries with concrete numbers, recovered enquiries listed, calls answered, reviews collected. The format need not be elaborate. It needs to be present and consistent.
The early warning signs of churn
Clients rarely cancel without warning. The signs typically arrive earlier: replies become slower, scheduled calls are skipped, requests for non-standard work increase. A practice that watches for those signals can usually rescue an engagement by changing the conversation. A practice that ignores them learns about cancellations during the cancellation itself.
Retention is partly relational
Beyond the visible work, the relationship itself matters. Clients who feel known, listened to, and respected by their operator tend to stay through periods of soft results that would otherwise have triggered a cancellation. This is not about being friends. It is about the small, undramatic acts of professionalism — returning calls, remembering details, not surprising the client with bad news — that compound into trust over many months.
Cohort retention curves and net revenue retention
Average tenure is a useful shorthand but it conceals two different problems. The retention curve — the percentage of clients signed in a given month who are still paying month one, three, six, twelve — is the more honest picture. A cliff at month three or four indicates a sales-fit problem (clients who weren't really sold). A slow grind from month eight onwards indicates a delivery-evolution problem (results plateau and clients drift). Different patterns, different fixes. Net revenue retention (NRR) takes this one step further: revenue from existing clients this period divided by what they paid last period, measuring whether expansion outpaces churn. NRR above one hundred per cent means the practice grows even before it signs anyone new; below one hundred, it must constantly refill. Practices serious about scale and exit watch this number every month.
Real-world parallel — Snowflake's NRR disclosure
Snowflake disclosed in its September 2020 IPO prospectus that its net revenue retention was 158 per cent — meaning every cohort of customers spent fifty-eight per cent more in the most recent twelve months than in the prior twelve, even before any new customers were added. By fiscal Q4 2022 (period ending 31 January 2022) the figure had risen to 178 per cent, before settling toward the 125 to 135 per cent range in subsequent years. The point is not to expect a small services practice to match those numbers — they reflect a high-margin software business with deep expansion-revenue mechanics. The point is that NRR is the headline metric SaaS investors examine when deciding whether a recurring-revenue business is durable enough to acquire. A services practice contemplating an exit in years three to five would be wise to start measuring this number from year one. Snowflake's S-1 and subsequent 10-Qs are public on SEC EDGAR.
If a client renews a £1,500-a-month engagement for two years, and another renews the same engagement for four months, what is the difference in their lifetime contribution? What does that imply about where attention should go?