[drop_cap]E[/drop_cap]very lever in this playbook has been pulled once by the time you reach this stage. Leads come in through a defined path. Proposals go out on a system, not a scramble. Retainers step up when the moment calls for it. Clients refer the next client. What none of that guarantees is that you know which lever is actually holding the business back — because a business, unlike a checklist, has exactly one binding constraint at any given time, and it is rarely the one that’s loudest.
This is Stage 10, the last stop in the Refine band, and it closes the loop rather than opening a new lever. The discipline here is finding the single weakest point in the system and fixing that one thing, instead of tightening all nine levers a little and learning nothing about which of them mattered.
The Pipeline Math
Walk your funnel the way you’d walk a P&L — one number per seam, computed the same way every time.
Enquiries this period: X
→ Qualified (real budget, real fit): X ( X / enquiries = X% )
→ Proposals sent: X ( X / qualified = X% )
→ Deals won: X ( X / proposals = X% )
→ Retained past month 3: X ( X / won = X% )
Average deal size: $X
Average retainer length: X months
Win rate (won / proposals sent): X%
Fill in your own numbers before drawing any conclusion. The value of this table isn’t the numbers themselves — it’s computing each seam SEPARATELY rather than looking at “leads” and “revenue” as if they were two facts instead of five gates in a row. A weak win rate and a weak qualification rate look identical from the outside — fewer deals close — and call for completely different fixes. One means your proposals aren’t compelling. The other means you’re proposing to the wrong people.
Utilisation and Effective Hourly Rate
An agency’s health hides in two numbers most owners never compute directly.
Utilisation is the share of your team’s paid hours actually spent on billable client work, versus internal meetings, business development, and the quiet hours nobody tracks. A team running at 55% utilisation isn’t failing at delivery — it’s carrying a hidden cost nobody’s pricing into the retainers, because the rate card assumes something closer to 75%.
Effective hourly rate is what you actually collected on a given engagement, divided by the hours it actually took — not the rate you quoted, not the rate on the proposal, the real number after every scope creep, revision round, and unbilled favor. Compute this per client, not just across the whole agency, and a pattern usually appears fast: your best-loved client and your most profitable client are often not the same client.
This second number is the one that quietly reframes what a “good” client is. A client who pays a high retainer but consumes triple the agreed hours in scope creep may have a lower effective rate than a smaller client who stays inside scope. Size and profitability are correlated, not identical — and the gap between them is exactly where margin gets eaten without anyone deciding to eat it.
The Client Mix Decision
Pull up your client list sorted by revenue and ask the concentration question directly: what share of total revenue does your single largest client represent? There’s no universal safe number, but there is a clear danger zone — a client above roughly a third of revenue means their decision to leave, downsize, or simply have a bad quarter internally becomes existential for you, not just inconvenient. That’s the feast-and-famine fear in its most concrete form: not “business is slow,” but “one phone call could end the agency.”
The honest test for firing a client isn’t whether they’re difficult — difficult clients who pay well and respect scope are worth keeping. The test is whether the effective hourly rate you computed above, combined with the emotional and reputational cost of the relationship, would let you say yes to this client again today, knowing what you know now, if they walked in cold. If the honest answer is no, the client is a candidate for a scope renegotiation, a price increase, or a graceful exit — not because they’re a bad person to work with, but because the numbers you just computed say the relationship is costing more than it’s returning.
Firing a client is never costless, and it’s rarely the first move. Try the price and scope conversation first — a client who’s genuinely valuable but underpriced or over-scoped often responds well to a direct, evidence-backed renegotiation, because you’re bringing them the same numbers you just computed for yourself. Reserve the exit for the client where that conversation has already failed, or where the relationship itself, not just the economics, has gone bad.
Finding the Weakest Lever
With the funnel, the utilisation number, and the client mix in front of you, the temptation is to fix all three. Resist it. Pick the one that’s actually binding — the one where improving it would move the whole business, not just one metric in isolation.
A funnel with a strong win rate but weak qualification is capacity-constrained on the wrong end — you’re closing well but wasting proposal effort on poor fits. A healthy funnel with low utilisation is a delivery problem no amount of new business will fix — you’re already not using the capacity you have. A strong funnel and strong utilisation with one client at 45% of revenue is a concentration risk that no operational fix addresses — only a deliberate decision about client mix does.
Name the constraint in one sentence before choosing what to work on next quarter. If you can’t name it in one sentence, you haven’t looked closely enough yet — go back to the three numbers above and find the one that’s actually moving the needle, rather than the one that’s simply easiest to think about.
AI Earns Its Place Here
Feed the tutor your funnel numbers at each seam, your rough utilisation estimate, and your client revenue breakdown, and ask it to compute the seam-by-seam percentages and flag which single number looks most out of line with a healthy range for an agency your size. Ask it explicitly to separate “this stage has a real problem” from “this stage is being starved by the stage before it” — a weak win rate fed by a weak qualification process isn’t a proposal problem, no matter how it looks from the win-rate number alone.
Push back on anything that reads as five recommendations. The value of this stage is picking one lever, and a model asked for a general health check will hand you a list — ask again, specifically, for the single constraint that’s binding the whole system right now, and the one change worth making before the next quarterly review, not the whole punch list.
The Loop Closes Here
This is the end of the playbook, and it hands you back to the beginning. Stage 0 built the hook that brought the first stranger in. Everything since has been about turning that stranger into a client who stays, grows, and refers the next one — and this stage is the one that tells you which part of that chain to strengthen next. Run the Friday numbers, name the weakest lever, fix that one thing, and when it’s no longer the weakest lever, come back here and find the next one. The playbook doesn’t end. It just tells you where to look.