The agency retainer conversation nobody wants to have

Why the conversation is awkward

The agency wants predictable revenue. The client wants to not pay for nothing. Both are entirely reasonable and they are being negotiated as if they were the same thing, which is why the discussion tends to circle.

The underlying problem is that a retainer sold as availability has no visible output. When a quiet month passes, the client experiences it as paying for nothing, and the agency experiences it as having reserved capacity that could have been sold. Both feel slightly cheated by an arrangement that worked exactly as intended.

Three shapes, and who each suits

Most disagreements are really about which of these is being discussed while both sides assume it is obvious.

Availability. The client is buying a response time. Suits stores where downtime is expensive and nobody internal can fix things. Fails when months pass without incident and the value becomes invisible.

Capacity. A block of hours each month. The most common and the easiest to explain. Fails when hours go unused and the client starts inventing work to justify the spend, which produces busywork neither side values.

Outcomes. Ongoing responsibility for something specific: performance, conversion, keeping the platform current. Hardest to scope and the most durable when it works, because the value is legible without anyone counting hours.

The most common failure is selling capacity while the client believes they are buying outcomes. They then measure the arrangement by results and find hours.

Naming what happens in a quiet month

The question that decides whether a retainer survives its first year, and it is almost never asked in advance.

Do unused hours roll over? For how long? If the answer is no, say so at the start and explain why, because a client who discovers it in month three feels misled even if the terms were written down.

The version that tends to work is a short rollover window, perhaps one month, which acknowledges that work is lumpy without letting a bank accumulate to the point where the client stops paying attention and the agency is carrying a large liability.

Making a quiet month visible

If a retainer is going to survive, the client needs to see something in the months where nothing broke.

The cheapest version is a short monthly note: what we checked, what we noticed, what is coming. Ten minutes to write, and it converts an invisible month into a visible one. The content does not need to be dramatic; "we checked the integration after the platform update and it is fine" is genuinely reassuring to somebody who did not know there had been an update.

Agencies resist this because it feels like manufacturing evidence of value. It is closer to the opposite: the work was real and previously invisible, and the note simply makes it legible.

Retainers rarely end because the work was poor. They end during a quiet stretch, when nobody could remember what the money was for.

When to say no to a retainer

Not every client should be on one, and proposing it universally damages the ones that should.

The client with no ongoing need. A store that is stable, low-volume and not changing does not need monthly hours. Selling one produces a client who cancels within six months and remembers the experience.

The client who wants insurance against a specific fear. Usually better served by a documented escalation path and an agreed hourly rate for emergencies than by a monthly fee that will feel wasted.

The client whose real need is training. Some stores are paying an agency to do things their team could do. A retainer entrenches that; a few sessions and some documentation resolves it, costs less, and generates more goodwill than a year of hours.

Turning down a retainer is a strong signal of good faith, and it tends to produce the project work that follows.

The alternative worth offering

For clients who do not need a retainer but do not want to lose you, the middle option is an open channel with an agreed rate.

They can ask small questions without an invoice attached, you see problems early, and anything substantial is quoted normally. No monthly fee, no unused hours, and no silence between engagements.

This costs the agency almost nothing to maintain and preserves the thing a retainer is actually protecting, which is being present at the moment the client first notices a problem. Handled well it produces more work than a retainer that gets cancelled in month seven.

Pricing it without guessing

Most retainer pricing is a number that felt right, which is why it is so often renegotiated within a year.

Work from what you are actually reserving. If a retainer commits you to responding within a day, you are holding capacity you cannot sell elsewhere, and that reservation has a cost whether or not the client uses it. Price the reservation, not the expected usage, and say that is what you are doing.

For capacity retainers, be honest about the utilisation you expect. If you know most clients use around two-thirds of their hours, pricing as though they will use all of them produces a rate that looks reasonable and delivers poor value, which the client eventually notices.

The alternative that avoids the whole problem: price outcomes rather than time. Harder to scope, and it removes the argument about unused hours entirely because hours were never the unit.

Ending one well

Retainers end, and how they end determines whether the client returns.

When a client wants to stop, resist the instinct to save it with a discount. A retainer that continues at a lower price because the client was unconvinced at the original price rarely lasts, and it devalues the arrangement for everyone else.

Instead, end it cleanly and leave the door open. Write the handover: what you were doing that now needs an owner, what is coming up, and what to watch for. Offer the open channel with an hourly rate as the alternative.

Clients who leave a retainer on good terms come back for projects. Clients who are talked into staying leave properly six months later and do not.

The review that prevents drift

Every six months, half an hour, with the client.

Two questions. What did we do in the last six months that you valued most? And what did you expect us to do that we did not?

The second question is uncomfortable and is where retainers quietly go wrong. Clients form expectations that were never agreed, do not raise them, and then cancel citing vague dissatisfaction. Asking directly surfaces the mismatch while it is still fixable, and it usually turns out to be one specific thing rather than a general problem.

The question to answer first

Before proposing anything, answer this for yourself: what are you actually reserving for this client, and what would you do with that capacity otherwise?

If the honest answer is that you have spare capacity and a retainer would fill it, you are selling availability you were not otherwise using, and the price should reflect that rather than a rate calculated as though you were turning work away.

If the answer is that you would genuinely have to decline other work, the retainer needs to be priced accordingly and the client needs to understand what they are buying. Most awkward retainer conversations come from the agency pricing the second while offering the first.

Writing it down

Whatever shape you agree, put four things in writing before it starts.

Which of the three it is: availability, capacity or outcomes. What happens to unused capacity and over what window. What is explicitly out of scope. And how either side ends it, including notice.

None of these are contentious at the start and all of them become contentious later if unstated. A page written before the first invoice prevents most of the conversations that end retainers badly, and it takes twenty minutes.

The version most agencies should offer

For most small agencies working with small stores, the honest answer is not a retainer at all. It is an open channel, an agreed hourly rate, and a short note every couple of months when something relevant changes.

No monthly fee to justify, no unused hours to explain, and no cancellation conversation in month seven. You stay present, the client stays comfortable asking small questions, and the work that follows is quoted normally.

Sell retainers to the clients who genuinely need one. Offer this to everyone else, and it will produce more work than a retainer they were talked into.

Common questions