Agency retainer pricing: how to raise rates without losing clients

TL;DR

Most agencies underprice quietly for years, then panic-raise badly. The workable playbook: know which retainers to raise (the coverage-ratio math — revenue under ~1.3× delivery cost is a decision, not a discomfort), walk in with receipts (the activity record of everything shipped this quarter is the negotiation), offer three doors instead of an ultimatum (new price for current scope, current price for reduced scope, or a restructured hybrid), time it to the renewal window after a visible win — never mid-crisis, never by surprise invoice — and sequence across the book in cohorts, worst ratios first, so one quarter's churn risk never touches more than a few accounts. Done this way, the realistic outcome across a book is most clients accepting, a few right-sizing scope, and occasionally one exit you needed anyway. The raise you don't make is the one that compounds: at typical agency margins, 10% across the book is the difference between funding growth and funding payroll anxiety.

---

There are two kinds of agencies: the ones that raised prices last year, and the ones subsidizing their clients with the founder's sleep. Inflation moves, payroll moves, platform complexity moves — a retainer that stands still for three years is a silent annual discount. Yet "we should raise prices" stays on the someday list because the fear is specific: this client, this email, this renewal call.

The fix is making the raise a system instead of a confrontation.

Step 1 — Let the math nominate, not your mood

The per-client profitability method produces one number per client: the coverage ratio — retainer revenue over loaded delivery cost. It also produces the raise list for free:

• Below 1.3: this client consumes other clients' margin. Raise, restructure, or exit — this is the "decide" band, and a raise is the kindest of the three options.
• 1.3–2.0 with scope creep: the activity share keeps growing while the retainer doesn't. Raise to re-match price to the work that's actually happening.
• 2.0+ and stable: leave it alone this cycle. Raising your healthiest accounts first is how agencies turn a pricing fix into a retention incident.

Mood-based raises target whoever annoyed you last month. Math-based raises target where the mispricing actually lives.

Step 2 — Receipts, not vibes

The raise conversation is won before it's scheduled, by the record: everything shipped this quarter, every request handled, every deliverable beyond original scope. If tasks and client communication live in one system, this is a filtered report; if they live in DMs and memory, this is the moment you discover why that matters.

The framing that works is descriptive, not apologetic: "Here's everything the engagement covered this quarter. The retainer that covers that work going forward is X." You're not asking permission to be paid more — you're re-stating what the engagement has become. Clients argue with adjectives; they rarely argue with an itemized quarter.

Step 3 — Three doors, never an ultimatum

A bare "the price is going up 20%" hands the client a binary: swallow it or leave. Offer three doors instead:

1. New price, current scope — the default door, for clients who value the engagement as-is.
2. Current price, reduced scope — name precisely what comes out. This door matters more than it looks: it converts "I can't afford the raise" from a churn event into a right-sizing event, and a smaller healthy retainer beats a lost logo every time.
3. Restructure — a different shape entirely: fewer departments but deeper, a project-plus-smaller-retainer mix, an annual commitment at a gentler number. For clients whose needs genuinely changed.

Three doors also changes the psychology: the client is choosing between versions of staying, not between staying and leaving.

Step 4 — Timing is half the outcome

• In the renewal window, 60–90 days out. Mid-term surprise raises read as a shakedown; renewal-window raises read as commerce.
• After a visible win, never during a fire. The QBR cadence from the reporting rhythm exists precisely so there's a natural, forward-looking room for this conversation — pricing premiered in a crisis call is pricing premiered as leverage.
• Live conversation first, paper second. The email confirms what the call already covered. Nobody learns their new price from an invoice.

And handle the grandfather trap deliberately: the legacy client at 2019 pricing being personally serviced by a senior lead is usually your worst ratio wearing your warmest relationship. Those get the most care — longest notice, most generous glide path — but they don't get exempted forever. Exemption is how the margin leak became structural in the first place.

Step 5 — Sequence the book in cohorts

Raising everyone at once concentrates churn risk into one quarter and floods your own team's emotional bandwidth. Sequence instead: worst coverage ratios at their next renewal, then the next band the following quarter. Each cohort teaches you the objection patterns before the next one hears them — and the capacity freed by any exits gets resold at new-client pricing, which is often the capacity board's quiet argument: when the board runs sustained-hot, raising prices on the bottom cohort is the alternative to hiring.

New business, meanwhile, starts at the new rates immediately. There is never a reason to onboard fresh clients at the prices you're trying to escape.

What actually happens

Agencies running this playbook report the same distribution: most clients accept with less drama than the founder rehearsed, a meaningful minority take door two and right-size, and occasionally one account exits — almost always one the coverage ratio had already convicted. The aggregate effect is the point: revenue up on a book that got smaller in workload, which is the only kind of growth that doesn't cost margin to service.

Pricing models that scale with value delivered rather than with friction — the argument we made for per-active-client pricing in our own product — age better precisely because they make this conversation rarer. But whatever the model: the retainer conversation you've been deferring is currently being settled by default, every month, in the client's favor.