A single retainer covering 30 to 50 percent of a solo practice's monthly runway is not stable revenue. It is exposure with a friendly invoice attached. In early 2026, a primary income source ended on short notice, and every fragile assumption I had been making about that relationship surfaced inside the same week. This is the calculation I should have been running months earlier, and the three forks the math forces once you actually run it.
When a retainer is revenue and when it is exposure
A retainer feels like ARR. Same date every month, same line item, same client name. After a year or two, the brain starts treating it as a fixture next to rent and internet, and that mental category is the failure mode.
A retainer is a contract either side can cancel on thirty days' notice or less. There is no claim on next month's payment, only the goodwill of a relationship and the inertia of a recurring calendar item. Both are real; neither is revenue in the sense that "revenue" implies durability.
The way to tell the difference is the cancellation thought experiment. Pick the largest line item in the monthly book. Imagine the email arriving tomorrow: "effective the end of this billing cycle, we are not renewing." Run the next ninety days as a spreadsheet. If the spreadsheet stays calm, the line item was revenue. If it triggers the part of the brain that handles existential threats, it was exposure.
The calculation behind that thought experiment is what I should have been running quarterly. Concentration risk, weighted by churn probability, divided by buffer. When the result crosses a threshold, the line item has stopped being revenue. Relationship can still be intact. Work can still be valuable. Math has changed underneath the relationship.
The fork: how to evaluate a retainer book under concentration risk
Once concentration is real, the practice has three forks. Each one is the natural response at a different concentration level, and each one has a cost the others do not.
Option A is to ride it. Do nothing structural. Trust the relationship, deliver good work, count on the renewal. The lowest-friction path on any given month, which is why it is most operators' default.
Option B is to backfill. Sell more retainers to dilute concentration. Take per-client share from 35 percent to 15 percent by adding new retainers under the existing one. The book grows, the ratio drops, the math improves on paper.
Option C is to change the shape. Move the practice off retainers as the primary revenue mechanism and onto something whose concentration profile is structurally different. Productized offers, scoped sprints, a published off-ramp for the existing book.
Each fork solves part of the problem and creates a different problem.
Option A: ride it
Keep the book intact, deliver to scope, introduce no structural changes. Trust the relationship to absorb whatever shocks come.
What it gives the practice: continuity, no awkward conversations, no pricing transitions, no public commitments to defend. For a practice with already-low concentration (no retainer above 15 percent of monthly), this is genuinely the right answer. The math is already calm.
What it costs when concentration is high: the spreadsheet stays fragile. One bad quarter at one client, one budget freeze, one CFO replacement, one acquisition, and the line item disappears. The practice then has between zero and ninety days to either find a replacement at the same revenue level or compress its cost base. Both are emergency moves, and emergency moves break pricing discipline.
There is a hidden cost that does not show up until the cancellation happens. A practice quietly riding 35 percent concentration for two years has not been building the muscles required to replace 35 percent of revenue from cold. The pipeline atrophies, the marketing surface goes dark, the discovery process gets rusty. When the moment comes to backfill, the practice is starting from a worse position than two years earlier, because the relationship that funded everything also displaced what would have replaced it.
Option B: backfill
Sell more retainers. Replace the dependence on one anchor by spreading exposure across three or four. A 35 percent client becomes a 12 percent client when the book triples in size, and the concentration math gets quieter.
What backfill gives the practice: better-looking ratios, more total revenue, a wider relationship surface, redundancy if any one client cancels. On the surface, this is the textbook answer.
What backfill costs is more subtle and is the trap worth flagging. Retainers are time-and-attention contracts. Three retainers do not multiply revenue without multiplying attention, switching costs, status meetings, and context-loading. A solo operator who could deliver one retainer well now delivers three acceptably. The original anchor's quality degrades. The new clients also expect anchor-quality attention. Nobody is happy.
The deeper trap is that the new retainers are also retainers. Three years in, the new clients have grown into their own anchor positions. A 12 percent client becomes a 22 percent one because the practice's cost base grew to match the higher revenue. The original client churned, the practice replaced it, and the same concentration problem reappeared at a higher gross revenue line. Math did not change, just took longer to surface.
Backfill is the right answer for a team where added clients can be staffed independently. It is the wrong answer for a solo operator whose attention is the binding constraint.
Option C: change the shape
Move the practice off retainers as the primary revenue mechanism. Sunset the existing book on a published date. Productize the work into scoped offers that complete on a known timeline. Build the marketing surface around the new shape.
What it gives the practice: a revenue model where no single buyer's cancellation triggers existential math. A productized offer that completes inside thirty or sixty days has zero churn risk after delivery, by definition. Risk lives at point of sale, not for the next twelve months. That is a structurally calmer profile for a solo operator.
What it costs: the transition is real work. Existing retainer clients have to be moved off, productized offers have to actually convert, the website has to support the new shape, the pipeline has to repoint. The transition window is the most fragile period the practice will go through.
This is the option I picked. The retainer sunset on the practice's availability page came directly out of running the concentration math on my own book and concluding that no patch on the existing model fixed the underlying exposure. Mechanics are in the published off-ramp writeup, and the productized layer that replaces most retainer work is in the three-tier ladder writeup.
The actual concentration math
Here is the formula I now run quarterly.
Effective single-client risk
= (single-client monthly / total monthly revenue)
× (probability of cancellation in next N months)
÷ (months of cash buffer)
Each term matters and each gets ignored differently.
The first ratio is the easy one. Single retainer's monthly invoice over the practice's total monthly revenue. A 10K invoice in a practice doing 25K is 0.40, or 40 percent concentration. Most operators glance at this number and stop.
The second term is the one nobody runs honestly. Probability of cancellation is a number the operator's gut wants to set to zero. The honest version is between 10 and 30 percent over a six-month window regardless of how the relationship feels. Buyers churn for reasons unrelated to the operator: CFO turnover, acquisition, budget cuts, strategic pivots, champion changing roles. None are inside the operator's control. The honest baseline is 0.20 unless there is specific evidence to revise downward.
The third term is the buffer. Months of cash on hand divided by monthly burn. The math collapses fast as buffer shrinks.
Run the formula at three concentration levels.
At 10 percent concentration, 0.20 churn, six months of buffer, the effective risk is 0.003. Statistical noise. Shrug at any cancellation.
At 35 percent concentration, 0.20 churn, three months of buffer, the effective risk is 0.023. An order of magnitude higher. A cancellation costs meaningful runway and the buffer is half what it should be for that concentration level. This is the zone where most solo operators sit and convince themselves the relationship is durable enough to compensate.
At 50 percent concentration, 0.20 churn, one month of buffer, the effective risk is 0.10. Two orders of magnitude above the calm scenario. The practice is one cancellation away from compressing the cost base in real time, in a panic, with no pipeline. This is the scenario the formula is designed to catch, and this is the scenario I was sitting in without running the math.
The single-number version across an entire book is the sum of each retainer's effective risk. Anything above 0.030 in aggregate is structural exposure for a solo practice. Anything above 0.05 is the kind of book that will eventually create the week I had in early 2026.
“The honest baseline is twenty percent churn over six months for any retainer, regardless of how the relationship feels.
”
The threshold I now hold the book to is 0.020 in aggregate, with no single retainer above 0.010 individually. That maps roughly to no client over 20 percent of monthly revenue with at least three months of buffer. Both were absent when the income shock arrived.
What I would revisit
Three things I would change about how I treated this math earlier.
First, make the threshold a posted limit, not a discovered one. The 0.020 aggregate and 0.010 single-client ceilings are now practice rules. Posting them in internal documentation and reviewing quarterly would have flagged the buildup six months before the income shock. The retainer sunset on the availability page now references this math so the calendar and the score move together.
Second, keep a churn-probability log. Per-client, updated quarterly, with notes on the buyer-side champion's tenure, funding posture, recent leadership changes, and contract renewal terms. Most probability inputs are observable if the operator is watching buyer context, not just the work delivered. I was watching the work.
Third, tie the concentration score to the public availability surface. The slot count on the public page is the inverse of what can be sold next month. If the score is high and slots are open, every new retainer signed makes the score worse. Slot count and the concentration ceiling should move together so the practice cannot accidentally re-concentrate while the sunset is in motion. The four engagement shapes article frames the same idea from the engagement-design angle.
A productized practice has different concentration math. A 50K project delivering in eight weeks has no churn risk after delivery; the risk lives in the pipeline. The formula here is specific to time-and-attention contracts where revenue depends on continuous delivery against an uncancelled relationship, which is part of why the retainer-vs-productized unit economics teardown lands differently for solo operators than the surface ratios suggest.
Frequently asked questions
What counts as a retainer for this calculation?
Any monthly recurring invoice that depends on the relationship continuing rather than a specific deliverable shipping. Agency-shape monthlies, fractional executive seats, advisor contracts, content-calendar retainers, anything billed on cadence and cancelled on notice rather than completed. A milestone-based engagement that bills monthly but completes on a specific date is different. The cancellation profile is what matters, not the billing cadence.
How do I calculate replacement probability honestly?
Start with 0.20 over six months as the baseline. Revise downward only with specific evidence. Contract with mutual exit fees: 0.10. Founder-led buyer at a healthy company: 0.15. VP-level champion whose role could be eliminated in a reorg: 0.30. Recent acquisition or pending funding round: 0.30. Two or three operators quietly replaced at this client recently: 0.40. The baseline is uncomfortably high because retainer cancellations cluster around buyer-side context shifts the operator is the last to know about.
What if I am already at 50 percent concentration?
Treat the next ninety days as a recovery window. Pause new retainer signings until aggregate score drops below 0.030. Build buffer aggressively, even at the cost of growth, until the months-of-cash term is at least three. Quietly start the marketing work that would let you replace the anchor if it cancels. Open a productized-offer conversation inside the existing relationship so the anchor becomes a shape-change candidate. The goal is to discover concentration during a calm quarter, not at the cancellation email.
Is productized revenue also concentrated?
Productized revenue is concentrated at point of sale, not in the back half of the engagement. A buyer who pays for a six-week sprint cannot cancel from week four. The concentration that matters in a productized practice lives in the pipeline, not the book. If 60 percent of next month's expected revenue depends on one open lead, that is a different shape of concentration risk and calls for similar discipline at the pipeline level.
When does this concentration math not apply?
When the relationship is not actually a retainer. Equity-stake operating partnerships, employment with notice periods, advisor seats with multi-year vesting all have different cancellation profiles. The math also does not apply at the very low end; below 5 percent per client with six months of buffer, the formula returns noise. The zone where the math earns its keep is 0.10 to 0.50 single-client concentration with one to three months of buffer, which describes most solo practices most of the time.
Sources and specifics
- The income shock that anchors this article occurred in early 2026 and ended a primary income source on short notice. The math in the article would have flagged the exposure at least six months earlier.
- The 0.020 aggregate and 0.010 single-client ceilings are practice rules, not industry benchmarks; they are the thresholds that keep my own book calm at a target three-month buffer.
- The 0.20 baseline churn probability over six months is calibrated from operator anecdotes running fractional engagements between 2018 and 2026, not a public dataset.
- The retainer sunset on the availability page is dated December 31, 2026, and the productized ladder that replaces most retainer work is at /products and in the work archive.
- The fork structure (ride it, backfill, change the shape) was the actual decision tree I ran on the book in Q1 2026 before publishing the sunset.