Expertise
Four domains, one decision.
The four levers below are not handled separately: each is paid for out of the others. Adding channels costs creative pace. Cutting a cost can break your ability to serve. The whole job is to hold those trades at the level of the accounts, rather than improving one lever while ignoring what it takes from the other three.
Acquisition
Where does the next pound go, and why there rather than anywhere else?
What I usually find
Budget concentrated on one or two channels, results measured on last click or on the platform’s own numbers, and creative production turning out variants instead of testing ideas. The average almost always hides one very profitable kind of customer and another digging a hole.
What gets decided
- The split between channels, and the level of spend beyond which each one stops returning what it returned.
- The role each channel plays: capturing demand that already exists, creating it, or keeping customers. That is more useful than comparing their cost per customer.
- The creative pace this requires: genuinely different ideas per period, not files delivered.
- How to measure: what advertising adds on top, holdout tests, models, or simply asking the customer — according to what your volume genuinely supports.
- Opening a new channel as a dated workstream, with a starting budget and a written success threshold — not as a good intention.
What gets delivered
- Channel split plan, with saturation thresholds and a starting budget for secondary channels.
- Measurement rules written down: what we trust, what we do not, and why.
- Creative brief: ideas to test, order of testing, and when to stop each one.
- A written scope for each media supplier, and the numbers they are judged on.
What I measure
- What an extra customer truly costs, not the figure the platform reports
- Share of revenue attributable to the most concentrated channel
- Margin left once the customer is paid for, by kind of customer
- Number of genuinely different creative ideas tested per period
What it is for
A split that takes a budget increase without melting the margin, whose numbers hold up in front of a finance director, and that does not collapse when a platform changes its rules overnight.
Growing volume
What breaks first if volume doubles?
What I usually find
A decision to sell more, taken without anyone checking what the business can absorb behind it. The step change happens, then delivery times stretch, support drowns, creative quality drops, and the cost of each order rises at exactly the moment it was supposed to fall.
What gets decided
- Finding what really blocks, measured rather than guessed: creative, logistics, support, servers, cash, or simply a decision nobody takes.
- The order in which you clear them: clearing one moves the problem somewhere else, better to know where.
- The pace of the climb: the cash a step change forces you to front before it returns anything.
- What to keep in-house and what to send out, at this level — the answer will change at the next one.
What gets delivered
- Map of what blocks: what each link can absorb, and where it breaks.
- Costed order of clearing, with the cost and lead time of each step.
- Cash forecast tied to the step change, kept separate from the revenue forecast.
- Stop conditions: the signals on which the volume increase is paused.
What I measure
- Cost of an order at current volume, then at target volume
- Real capacity of the weakest link
- Cash to front in order to clear the step change
- Gap between volume sold and volume served within the promised window
What it is for
A step change cleared without each sale earning less, a cash position that was not caught out, and the next blockage named before it shows up.
Cost
Which costs fund growth, and which ones simply endure it?
What I usually find
Costs stacked layer by layer: tools bought for a need that no longer exists, suppliers renewed out of habit, servers sized for an old peak, and an acquisition cost watched closely while three heavier lines are reviewed by nobody.
What gets decided
- What a new customer costs, measured against the margin it buys rather than the revenue it triggers.
- What it costs to serve an order: fulfilment, support, servers, payments, refunds.
- Software and servers, line by line, against what each one actually produces.
- The cost of the structure itself: teams, suppliers, and this engagement included — if it is no longer the best use of the money, I say so.
What gets delivered
- Full cost of serving an order, recalculated and documented line by line.
- Inventory of tools: annual cost, usage actually observed, recommended decision.
- Reduction plan in order, with the risk of each cut written down.
- The conditions under which a reduced line can be reopened.
What I measure
- Margin left per order, per account or per active user
- Cost of software and servers as a share of revenue
- Share of costs nobody has reviewed in the last twelve months
- Cost of the structure as a share of the revenue it steers
What it is for
Margin rebuilt without cutting blind: lines reduced because they produced nothing, not because they stood out in a spreadsheet.
Team leadership
Who decides what, and on the strength of which number?
What I usually find
A capable team that ships a great deal and decides very little, suppliers judged on numbers they chose themselves, and a weekly meeting that reviews tables without ever producing a dated decision.
What gets decided
- How the in-house team is organised: which roles, who decides what, and which hires not to make.
- What stays in-house and what goes out, workstream by workstream, with the reasoning written down.
- Decision meetings: how often, with whom, on what agenda, and above all what has to come out of them.
- The numbers each person is judged on, chosen before the period rather than after it.
What gets delivered
- A map of who decides what, with names on it.
- Weekly meeting framework: format, numbers to prepare, decisions expected.
- Role definitions and assessment grid for the hires that are going ahead.
- Remit written into the contract, and judging numbers, for each supplier.
What I measure
- Time between a problem surfacing and the decision that follows
- Share of open workstreams with a named owner and a date
- Number of parallel workstreams per person
- Age of decisions that have never been revisited
What it is for
A team that decides on its own, suppliers judged on numbers they did not choose, and an organisation that holds without me.
The starting point is always the same.
A diagnostic that recalculates what each sale truly earns and names what blocks. The four domains above are not a menu: the diagnostic decides which one you start with.