Retainers don't fail at the pitch. They fail slowly, in the middle of year one, by a mechanism so gradual that nobody in the agency can name the month it went wrong.
The mechanism is this: scope is defined once and never again, while the client's expectations reset every month to whatever you did last month. Do a favour in March and it's the baseline in April. Nothing is agreed, nothing is argued, and the margin erodes three or four hours at a time.
What the erosion actually looks like
An $8,000 a month retainer, scoped at 50 hours, delivered by a team with a loaded cost of $110 an hour.
| Month | Hours delivered | Cost | Margin | Margin % | Effective hourly rate |
|---|---|---|---|---|---|
| 1 | 46 | $5,060 | $2,940 | 36.8% | $173.91 |
| 5 | 61 | $6,710 | $1,290 | 16.1% | $131.15 |
| 9 | 79 | $8,690 | −$690 | −8.6% | $101.27 |
| 12 | 84 | $9,240 | −$1,240 | −15.5% | $95.24 |
Across the full twelve months, with hours climbing steadily from 46 to 84, the team delivers 794 hours — an average of 66 a month against 50 scoped. Total cost $87,340 against $96,000 of revenue.
$96,000 − $87,340 = $8,660 margin, or 9.0%
A retainer that opened at 36.8% closed the year at 9.0%, and the client is delighted. That last part is what makes it hard to fix: there's no complaint, no crisis, no obvious trigger for a difficult conversation. The account is "working."
Notice the last column. The effective hourly rate falls from $173.91 to $95.24 — you are doing the same work for the same client at 45% less per hour than when you started, having never agreed to a discount.
Where the hours come from
None of these are unreasonable requests. That's precisely why they're dangerous.
Precedent. One extra round of amends in month two becomes the expected number of rounds by month five. Nobody ever said it was included; nobody said it wasn't.
Access creep. Month one has one contact. By month six there are four stakeholders, each with opinions, each in the review call, each generating follow-up.
The reporting tax. A monthly report becomes a monthly report plus a call plus a deck for the client's board. Hours you can't invoice against work you can't point to.
Relationship drag. As the relationship warms, informal requests replace formal ones. Slack replaces the brief. Turnaround expectations shorten because you're "on the team now."
Silent scope substitution. The client's priorities shift and the retainer quietly starts covering a different, larger job than the one that was priced.
Price the model, not the month
Most retainers are priced as a discounted block of hours, which is the weakest possible structure: the client sees a monthly fee and no ceiling, and you see a number that only works at a utilisation you never verify.
Three structures work better.
Capped hours with a written overage rate. "Up to 50 hours a month; hours beyond that are billed at $X." You will rarely invoice the overage. That's fine — its job is to make the ceiling real, so the month you go 30% over there is an obvious, pre-agreed conversation instead of an awkward one.
Deliverable-based. Not hours but outputs: four pieces of content, two campaign cycles, one report, defined revision rounds. Shifts the conversation from "what else can you do this month" to "what are we producing," and the answer is written down.
Outcome or value pricing. Fee attached to a result rather than an input (compared with fixed price and time and materials here). Genuinely better when it fits, and it only fits where the outcome is measurable and largely within your control — which is a narrower set of engagements than most agencies want it to be.
Whatever the structure, price at your realistic hours, not the optimistic ones. If comparable retainers historically run 20% over scope, price the 20%.
The mechanisms that keep it honest
Pricing is a one-off decision. Margin erosion is continuous, so the defence has to be continuous too.
Track hours per retainer monthly, against scope. One number per client per month, from time logged against the retainer: hours delivered versus hours scoped. Without it, everything below is guesswork.
Set a written trigger. Something like: two consecutive months above 15% over scope — 57.5 hours on a 50-hour retainer — and the account gets re-scoped or repriced. Deciding the threshold in advance removes the judgement call from the person least able to make it objectively.
Re-scope every six months, on the calendar. Not a price conversation, a scope conversation: here's what we agreed, here's what we've been doing, let's realign one to the other. Far easier to have than an annual price rise, and it usually resolves the problem without one.
Build in an annual uplift at signature. A stated yearly increase in the original agreement is administrative. The same increase raised in month fourteen is a negotiation.
Give someone the scope ledger. One person per account records out-of-scope requests as they happen — not to refuse them, to have them written down when the six-month conversation arrives. "We've absorbed 41 extra hours since March" is a fact. "It feels like we're doing a lot more" is an opinion, and it loses.
Fixing one that's already gone
Take the example above at month nine, running at 79 hours against a 50-hour scope. Two ways back to a 35% margin:
- Reprice. At the twelve-month average of 66 hours, you'd need about $11,200 a month — a 40% increase. Almost nobody wins that conversation.
- Reset the scope. Get delivery back to the 50 hours you sold, and the same $8,000 produces $2,500 of margin, or 31.3%. No price change required.
The second is easier, and it's the one agencies skip because it feels confrontational. It usually isn't. Most clients don't know what's in scope, because nobody has told them since the proposal — and a re-scope framed as "let's make sure your fee is going on the things that matter most" reads as attention, not retreat.
This week
Pick your oldest retainer. Pull the hours delivered in month one and the hours delivered last month. Multiply each by your loaded hourly cost and subtract from the fee.
Two margin numbers, one line apart. If the second is materially worse, you've found the shape of the problem across every retainer you run — and the fix is a scope conversation, not a price rise.