Founders ask Collective 54 this 2 times in our records, 1 of them in 2026. The recurring revenue, account risks and client satisfaction answers on this site cover building retainers, spotting risk in accounts and gathering feedback; this page covers the number to aim for and how to measure it.
The client retention essay in the newer book says that once revenue becomes recurring, client retention becomes the constraint. Growth in a recurring revenue firm does not break when sales slow down; it breaks when clients leave. Its threshold is roughly 90 percent. Below it, churn quietly erodes progress: new revenue replaces lost revenue instead of compounding on top of it, expansion feels uphill, and forecasts feel optimistic until they are not. Above it, growth becomes calmer, expansion sticks and the business starts to look like a transferable asset rather than a personal vehicle.
The essay explains why the line sits there. A 10 percent annual churn rate implies an average client lifetime of roughly ten years, which it calls the minimum time a boutique needs to move cleanly through the grow and scale stages. So the honest answer to what churn to expect is that many firms run higher than 10 percent, and the goal is to get below it and stay there.
The client relationships chapter of the 2020 book was written for boutiques generally, including project firms. Its rule of thumb is average client tenure of three years or more, because clients who keep paying for years show the relationship is strong. It adds that no single client should be more than 10 percent of billings, and that relationships should sit with the firm rather than with one employee. Its checklist asks whether client relationships are an asset on the balance sheet and whether that asset is appreciating.
As an inference, the two figures fit together. Three years is a floor that shows clients value the work. The 90 percent line is the bar for a firm that has moved to recurring revenue and wants to scale and sell; the newer book is the one to follow where they differ.
The essay warns that many firms have built recurring billing, not recurring revenue. Recurring billing describes how clients pay. Recurring revenue describes why clients stay. A firm can move clients to monthly invoices, label an engagement a subscription and add auto-renewal, and none of that guarantees the client will stay when the renewal moment arrives. As an inference, this matters for measurement: a client still on a monthly invoice is not proof of retention if the work, the sponsor or the value has quietly faded.
The fee quality chapter of the 2020 book makes a related point about project work. Boutique work is temporary; clients are renting you, and at some point they stop paying the rent because they no longer need the work or take it in-house. That is why the chapter wants contracts longer than twelve months and services that build on one another.
Collective 54 publishes no retention formula, so what follows is an inference that organizes the published material. Keep the measures few and use the same definitions every year.
Client retention: the clients active at the start of the year who are still active at the end, divided by the clients active at the start. New clients won during the year are left out, so growth cannot hide losses. Churn is the remainder.
Revenue retention: the revenue from those same starting clients this year compared with last year. As an inference, this shows whether you are keeping the relationship but shrinking it, which client counts miss.
Tenure and concentration: the average age of your client relationships, checked against the three-year floor, and the share of billings from your largest client, checked against the 10 percent limit.
Fee mix: the fee quality chapter suggests roughly 60 percent of fees from existing clients and 40 percent from new ones. As an inference, a firm far above 60 percent may be retaining well but has forgotten how to hunt, and one far below it is probably losing clients faster than it thinks.
The essay says retention is not one thing, because each type of recurring revenue churns for a different reason. Retainers rarely fail from dissatisfaction; they fail from perceived stagnation. Subscriptions churn from value opacity, when clients cannot say what they get. Outsourcing contracts erode through silent replacement, as another vendor or an internal team takes over. Fractional executive roles suffer executive drift, as leadership teams mature or change. Long-running projects break down through momentum decay. As an inference, report the retention rate for each revenue type separately, because a blended number can hide one line that is leaking.
The essay is direct that software retention playbooks fail in services. Software companies retain customers on telemetry: usage, logins and adoption that the product reports automatically. Professional services firms have none of this. Value is experienced through conversations, judgment, outcomes and trust. It adds that health scores in the last era were based on opinion rather than evidence, reviews were episodic and backward-looking, and surveys captured snapshots rather than trajectory. As an inference, a retention rate tells you what happened. It does not tell you who is about to leave.
The essay describes the signals that matter and where they live: meeting conversations, email tone, responsiveness, executive presence and shifts in engagement. Its list of what AI can detect early includes disengagement, value confusion, relevance drift, executive withdrawal and priority loss, interpreted differently by revenue type. It describes a division of labor in which AI does about 80 percent of the work, monitoring every account continuously, and people do the 20 percent that needs judgment: re-anchoring value, handling executive change and rebuilding confidence. The account risks answer on this site covers putting that into practice.
The 2020 book tells a story from SBI, a firm that served sales leaders, whose clients stayed in their jobs for about eighteen months on average. The founder feared a buyer would see short client tenure. The data showed the sales leaders hired the firm again at their new companies. As an inference, if your work follows people, track that, because a sponsor who leaves and brings you along is a retained relationship that a simple account count would record as churn.
Collective 54 publishes no retention formula, industry churn benchmark by service type or health score. The published positions are retention as the constraint once revenue is recurring, the roughly 90 percent threshold and the ten-year lifetime behind it, recurring billing versus recurring revenue, the churn causes by revenue type, the failure of software playbooks and opinion-based health scores, the early signals and the 80 and 20 division of labor, three-year average tenure, the 10 percent concentration limit, the 60 and 40 fee mix, contracts longer than twelve months, and clients renting a boutique.
If most of your revenue is project work, as an inference, an annual churn rate means less than repeat business: track how many clients buy again within a set period, and use the three-year tenure floor.
If you have only a handful of clients, one loss swings the rate wildly; watch tenure, concentration and the signals instead.
And if a client leaves because the problem is solved, record it separately; it is still churn, but it calls for a different fix.
For recurring revenue, keep at least 90 percent of clients a year; the client retention essay says anything less makes scaling harder and more fragile, and 10 percent churn means a client lifetime of about ten years. For any boutique, the 2020 book sets a floor of three years average tenure and no client above 10 percent of billings. Measure retention among the clients you started the year with, add revenue kept, tenure and concentration, and split it by revenue type. Then watch the signals in conversations and engagement, because the rate only tells you what already happened.
The client retention essay sets roughly 90 percent a year as the line for recurring revenue firms, which implies an average client lifetime of about ten years.
Collective 54 publishes no formula. As an inference, take the clients active at the start of the year, count how many left by the end, and divide by the starting number, leaving new clients out.
The 2020 book gives a rule of thumb of three years or more of average tenure, and treats longer tenure as evidence that clients value the work.
The client retention essay says retainers rarely fail from dissatisfaction; they fail from perceived stagnation, when the client wonders whether the work is still evolving or still necessary.
Sources: Greg Alexander, The AI-Native Boutique Firm (Advantage Books, January 2027), specifically The AI Client Retention Manager for retention as the constraint once revenue is recurring, the roughly 90 percent threshold, 10 percent churn implying a ten-year client lifetime and the grow, scale and exit stages, recurring billing versus recurring revenue, churn causes by revenue type, software telemetry and the failure of software playbooks in services, opinion-based health scores and episodic reviews, where retention signals live, the early signals AI can detect, and the division of labor of about 80 percent AI and 20 percent people. Greg Alexander, The Boutique: How to Start, Scale, and Sell a Professional Services Firm (Advantage, 2020), chapter 31 for average client tenure of three years or more, no client above 10 percent of billings, institutionalized relationships, client relationships as an appreciating asset, and the SBI sales leader story; chapter 32 for the 60 and 40 fee mix, contracts longer than twelve months, services that build on one another, and clients renting a boutique. Related Collective 54 answers on this site: how do I build recurring retainer-based revenue instead of one-off projects; how do we flag account risks and upsell opportunities; how do I track and collect client satisfaction data; what is the best practice for running QBRs and account check-ins; how do I grow revenue by expanding within existing accounts. Note on scope: Collective 54 publishes no retention formula or churn benchmark by service type. The definitions of client retention and revenue retention, reading the fee mix as a retention check, reporting by revenue type, tracking sponsors who move, and the flips are inferences used here to organize the source material rather than published Collective 54 positions.
Collective 54 is the private community for founders and executives of boutique professional services firms between $5M and $50M in revenue. Members work these answers against their own numbers.