Build a marketing scorecard that connects traffic to customer value

Design the scorecard around a causal hypothesis

Acquisition is a sequence of commitments: attention, arrival, suitable interest, first value, and repeat value. A useful scorecard pairs each transition with a possible failure mechanism. Low qualified inquiry volume can result from weak distribution, poor audience fit, or an unclear destination. One overall conversion percentage cannot distinguish them.

Choose one outcome the team owns, one leading indicator, and a guardrail. An example is activated accounts, supported by setup completion, with support time per activated account as a guardrail. Record the eligible population and the time allowed for activation.

Give the team a shared measurement dictionary

The definitions in this table are starting points, not permission to blend systems. Keep a field for scope, event source, identity method, reporting delay, and owner. Retention can mean weekly use, renewed subscription, or repeat purchase; choose the behaviour that represents continuing value for your product.

Reconstruct an acquisition decision from contribution

Consider an illustrative $1,000 in campaign-attributed revenue. Subtract $550 variable delivery costs, $300 media, and $100 additional campaign production cost: contribution after campaign cost is $50 before overhead. Reporting 3.33× ROAS alone hides most of this picture. Calling that $50 incremental profit also goes too far without evidence that the sales would not otherwise have occurred.

For planning, estimate allowable acquisition cost from conservative contribution and the amount you need to retain. Include sales effort, onboarding, refunds, and support where relevant. Separate one-time setup from repeatable campaign costs, and show both views when deciding whether a channel can scale.

Align cohorts, windows, and denominators

Do not divide this week’s spend by this week’s customers when most deals take two months to close and call the result definitive CAC. Use acquisition cohorts or clearly label the period-based approximation. Similarly, compare retention only for cohorts that have reached the same age. Customers acquired yesterday have not yet had the opportunity to renew.

Averages can also mislead. If two channels have 10 leads each from 100 and 1,000 sessions, the combined rate is 20/1,100, or 1.82%. The denominator supplies the weight. Keep counts beside percentages and distinguish a percentage-point change from a relative percentage change.

Use targets as constraints to test

Assume $80 contribution before acquisition and a requirement to retain $30. Allowable CAC is $50. At a 20% qualified-lead close rate, allowable cost per qualified lead is $10. At a 4% click-to-qualified-lead rate, allowable CPC is $0.40. This scenario is arithmetic, not a forecast; each rate needs a comparable source and a plausible downside case.

If observed costs exceed the limit, investigate the constraint before changing the bid: fit, qualification, offer clarity, close process, delivery cost, or measurement. Optimising cheap clicks can make the business worse when customer quality falls.

Make the review produce an accountable action

Add a decision column to the scorecard. “Activation fell” becomes “Observe five incomplete setups and inspect the permission step by Friday.” “ROAS increased” becomes “Confirm refund maturity and whether the audience mix changed.” Keep a record of decisions made under uncertainty, then revisit them when delayed outcomes arrive.

Use LTV estimates carefully in a young business. Prefer realised cohort contribution and bounded scenarios over an optimistic infinite lifetime. A scorecard should expose what you do not know while keeping the team focused on a useful next observation.

Continue the traffic and growth learning path

Find the next useful guide in this 16-part series →