Profitable growth: connect contribution, causality and cash

Amazon defines advertising cost of sales, or ACoS, as advertising spend divided by advertising revenue, expressed as a percentage. Return on advertising spend, or ROAS, reverses that ratio. They summarise the relationship between spend and credited revenue; neither formula includes product costs, fulfilment costs or fixed overhead. 1

ACCA defines contribution as revenue less variable costs and distinguishes it from profit after fixed costs. 2 For this series, contribution after advertising means net sales less the relevant product and variable selling costs, then less the corresponding advertising spend. Reconcile the treatment of discounts, refunds, taxes, returns and timing before comparing reports.

Illustrative economics. Assume £10,000 of attributed revenue is on exactly the same net basis as the finance calculation, pre-advertising contribution is 30%, and spend is £2,000. ROAS is 5 and ACoS is 20%. Estimated contribution after advertising on those sales is £3,000 minus £2,000, or £1,000, before fixed costs. If the pre-advertising contribution were only 15%, the same advertising ratios would leave negative £500.

Under those narrow assumptions, an ACoS equal to the pre-advertising contribution percentage leaves zero contribution after ads on the attributed revenue. This is not whole-business break-even, and it is not an incremental-profit test. It fails as a shortcut when the denominators, product mix, tax basis or attribution assumptions differ.

Think at the margin. Average historical return does not establish the return on the next pound spent. Expansion may change auction costs, placements, customer mix, stock requirements or promotional intensity. Compare the estimated additional contribution with the additional spend and costs; record which parts are observed and which are assumptions.

Leadership implication. Set targets from the business model rather than copying a universal “good ACoS”. Portfolio averages can conceal weak products and subsidised growth. Review product-level economics where reliable, then reconcile to the total business so allocation rules do not create fictitious precision.

Incrementality concerns the outcome caused by an activity relative to what would otherwise have happened. The latter is called the counterfactual: the comparison world that cannot be observed directly for the same customer at the same moment. Controlled experiments use comparison groups to estimate it. Google's Conversion Lift documentation provides a practical example. 3

A well-designed experiment can support a causal estimate, but its conclusion remains bounded by the population, treatment, period and measurement method. Exposure overlap, insufficient data and implementation failures can weaken the result. A statistically inconclusive result is not proof that the effect is zero. A precisely measured revenue effect is not automatically profitable.

Before a study, specify the primary outcome, the smallest effect worth acting on, the comparison design and the planned decision. Avoid repeatedly checking noisy results and stopping only when they look favourable. Where rigorous testing is not feasible, label observational findings as such and make the uncertainty part of the budget decision.

Retention introduces time. A repeat-purchase rate needs a defined cohort and observation window. A cohort is a group sharing a starting event, such as a first observed purchase in a particular month. Compare cohorts at a similar age: a twelve-month-old cohort has had more opportunity to repeat than a two-month-old cohort. Durable products and replenishable products also require different interpretations.

Customer lifetime value should not be confused with lifetime revenue. A decision model should consider expected future contribution, retention-related costs, uncertainty and the timing of cash flows. Discounting recognises that a pound expected later is not equivalent to an available pound today. Forecast assumptions should be explicit and tested against mature cohorts when possible.

Illustrative caution. If first-order acquisition loses £5 of contribution, a hoped-for £10 future contribution does not by itself justify the investment. Ask how many customers actually repeat, when they repeat, what serving them costs and whether the company can fund the gap. Subscription enrolment is not the same as a guaranteed series of completed, profitable orders.

Leadership implication. Treat retention as an observed economic relationship to develop, not an unlimited justification for current losses. Acquisition and retention should share one contribution model, with clear boundaries between measured performance and forecast value.

Profit measures economic performance under accounting rules; cash flow records movements of cash. Inventory purchases, supplier terms and the timing of receipts can create a gap between the two. ACCA's working-capital framework defines the cash operating cycle as inventory days plus receivables days minus payables days. 4

Illustrative cycle. Sixty inventory days, fourteen receivables days and thirty payables days produce a forty-four-day cash operating cycle. This is a simplified operating measure, not a prediction of a marketplace settlement schedule. Actual funding needs also depend on payment timing, deposits, reserves, fees, tax, refunds and the scale of operations.

As sales grow, more stock may need to be purchased before the associated cash is collected. A profitable unit can therefore increase the financing requirement when many more units must be funded. Supplier credit and faster stock turnover can reduce that need, while longer lead times and slower sell-through can increase it.

Build a cash forecast from actual dates and obligations rather than applying the cycle formula mechanically. Include purchase commitments, the point when inventory becomes sellable, expected receipt timing, advertising payments and unavoidable overhead. Model slower sales and delayed replenishment as well as the base case.

Inventory is also a risk position. Excess stock can lock up cash and create disposal or markdown exposure; insufficient stock can prevent the business from fulfilling demand. Neither maximum availability nor minimum inventory is universally optimal. The trade-off depends on uncertainty, product life, service requirements and funding capacity.

Leadership implication. A campaign budget and a purchasing plan should not be approved independently when they depend on the same sales forecast. Identify the peak cash requirement, the downside case and the response if assumptions fail. A commercially promising campaign may need to be paced differently because the business cannot safely fund the associated inventory.

Bring three views to the same decision: contribution, causal evidence and cash. Contribution asks whether transactions support the business. Causal evidence asks how much additional value the activity creates. Cash asks whether the plan can be financed through its least comfortable point.

For each proposed expansion, document the expected benefit, incremental costs, funding requirement, uncertainty range and alternative use of the money. Define a stop or revision rule before scaling. The appropriate rule depends on the business; it is not a universal percentage supplied by a platform.

Conclusion. Profitable growth is not a single dashboard ratio. It is a sequence of defensible decisions about demand, delivery and capital. Leadership adds value by making the connections explicit, correcting assumptions when evidence changes and refusing to confuse a persuasive narrative with a measured commercial result.