A new client signs. The monthly retainer looks healthy, the proposal is good, and the account team is excited. Four months later, everyone is busier and the agency has surprisingly little to show for it.
Most founders recognise some version of that story. The problem isn't always pricing. Sometimes it's staffing, scope creep, generous rounds of revision, work that gets "just sorted out" by the founder or a service mix that never made commercial sense at the agreed fee.
My earlier essay about revenue growth and agency profit looks at why top-line growth can conceal these issues. Here, let's put the economics on one account and make the problem visible enough to act on.
Know which margin you're talking about
Account contribution is not necessarily the same as company operating profit. Direct labor, freelancers and account-specific software are usually visible enough. Shared management time, facilities, subscriptions, sales expense and the founder's unrecorded work are harder to assign consistently.
Start with a fair, repeatable allocation. Use fully loaded hourly delivery costs—pay plus employer costs and relevant benefits, divided by genuinely available delivery hours, rather than salary divided by every hour on a calendar. Keep the method consistent across accounts. If your owner time is already paid and counted in delivery or overhead, do not count it twice.
Then run two views: an operating contribution before imputed owner time, and an economic margin after assigning a reasonable cost to unpriced owner work. That second number is not a formal accounting metric. It answers a founder question: would the account still be attractive if someone else had to do the job?
Use the client profitability calculator
All figures are illustrative and use a single month in US dollars. Adjust the loaded cost assumptions to your actual business. No figures leave your browser.
What a $12,000 retainer can conceal
Consider the example loaded into the calculator: $12,000 in fees; 90 planned delivery hours and another 16 hours of rework at $55 loaded cost; $900 of freelancers; $350 of dedicated tools; $1,350 of allocated shared overhead; and 10 unrecorded founder hours at a $130 assumed economic rate.
On that illustration, operating contribution before imputed founder time is $3,570—about 29.8%. After assigning $1,300 to the founder time, economic contribution is $2,270, or about 18.9%. If you require an illustrative 25% margin at that same cost, the fee would be approximately $12,973, with no allowance for future surprises. That is not a recommendation to raise your price to an arbitrary number. It is a reason to understand the actual work before choosing a price.
Find the cause, not just the embarrassing percentage
Scope creep. Look for tasks the account team performs that were never in the signed scope. Sometimes the solution is a clearer change-control conversation, not working faster.
Unpriced management. Account leadership, reporting, internal meetings and senior review have a cost. If they are essential to the service, they belong in the model.
Rework. Distinguish mistakes the agency should absorb from genuinely new requests. Your client should not fund avoidable quality failures. Neither should your team quietly deliver an unlimited new scope for free.
Utilisation assumptions. A nominal hourly salary cost can understate holidays, management, training, employer charges and unavoidable nonbillable time. Overly optimistic utilisation makes every account look wonderful on paper.
Founder heroics. If one person continually rescues the relationship, a seemingly profitable client may be consuming the one resource that has the greatest opportunity cost.
A five-question account review
Ask the account owner to bring the last three months of actual fee receipts, all staff and contractor time, change requests, missed deadlines and an honest note about how much senior support the client needed. Then ask:
- What does the client believe is included, and what did we actually deliver?
- Which work could we stop without hurting the promised outcome?
- Which expensive rework was our fault and what process would prevent it?
- Which additional requests should be priced or consciously declined?
- If this client doubled in size next month, would the economics improve or deteriorate?
Don't compare account managers by margin alone. A newly onboarded client may have intentional start-up costs; a strategically valuable client may create legitimate cross-selling opportunities. Write down the rationale and revisit it. The problem is using “strategic” as a permanent excuse for an account nobody has ever measured.
What to do with an unprofitable client
Your choices are broader than “fire them” or “suffer.” Agree a tighter deliverable, repair a broken workflow, change who approves revisions, renegotiate the fee at the appropriate point, or stop selling a service you cannot deliver well. If you do need to end an engagement, review the client decision guide and the companion concentration stress test first.
One point I try to make to agency owners: clients cannot be expected to fix economics that only you understand. Bring a clear scope, evidence and a good-faith proposal. Take responsibility for your own delivery errors before you ask the client to pay more.
Profitability is a sequence, not one mysteriously reassuring percentage
For each account, separate three layers: direct contribution after staff, freelancers and dedicated tools; allocated account contribution after a consistent share of overhead; and economic contribution after unrecorded founder effort. There isn't one universal accounting convention for allocated client profit. Be explicit about yours before comparing accounts or making pricing decisions.
Two $12,000 retainers can look identical in the revenue meeting. If account A needs $8,000 in total modelled economic cost and account B needs $13,000 because of revisions and senior intervention, one contributes $4,000 while the other destroys $1,000. More revenue from B would make the business busier and worse off without a delivery change. These are illustrative costs, not a report on Anvil.
Investigate the hours nobody reports
Ask the team to reconstruct the last ten working days of one account: contracted delivery, coordination meetings, rework, client calls, escalations, reporting and founder intervention. Compare that with the proposal. Categorise every discrepancy as (1) a mistake to fix internally, (2) an agreed inclusion, or (3) a new client request that needs a scope conversation. Don't lump all three into "scope creep".
The fastest improvement may not be a higher fee
A pricing increase sometimes makes sense, but it may be a poor response to a preventable quality issue. Repair that workflow first. Conversely, there is a limit to operational optimisation if a client is asking for fundamentally different work. Compare the revised scope and delivery plan against agency pricing models, then ask whether the relationship is disproportionately important in the client-concentration analysis. Numbers become valuable when they reveal an actual choice.
Bring one client and the real costs.
We can challenge the scope, founder effort and pricing assumptions before you conclude an account needs to be dropped or repriced.
EXPLORE FOUNDER ADVISORY ↗