Trial growth is an intermediate result
A SaaS acquisition model can reward channels that generate many trials without generating retained customers. Keep the acquisition question separate from product activation, paid conversion and retention. Each stage can change for reasons beyond advertising, including onboarding, pricing, product reliability and sales follow-up.
- Define acquisition outcome
- Wait for cohort maturity
- Estimate retained contribution
- Constrain the change
Choose an outcome you can observe consistently and a decision horizon that finance recognises. If you model paid starts, report the uncertainty in the later value per customer. If you model revenue, explain how existing subscriptions and new acquisition enter the baseline. The break-even ROAS guide explains why revenue is not the same as contribution.
A budget recommendation needs credible response estimates, consistent value assumptions and feasible constraints. The marketing budget decisions learning path connects this idea to the assumptions, practical checks and next steps.
To work through the connected decisions, read MMM for profit: choose contribution over revenue alone and B2B MMM: model pipeline without confusing sales-cycle lag.
Compare cohorts on equal terms
Imagine channel A produces 100 incremental paid starts at £200 each and channel B produces 80 at £250 each. Acquisition spend is £20,000 in both cases. If their expected contribution over the chosen horizon is £300 and £450 per start respectively, the illustrative net contributions after advertising are £10,000 for A and £16,000 for B. Ranking by acquisition count alone reverses that comparison.
Those value estimates are assumptions, not facts created by the MMM. Use cohort evidence and sensitivity ranges. A new channel with little follow-up should not inherit the retention of a mature channel without justification. Compare cohorts at the same age and separate actual retained value from extrapolated value.
Watch for product changes that coincide with media
A new onboarding flow launched during a media increase can improve paid conversion. Record that intervention and investigate its relationship with the outcome. However, activation may also mediate the effect of attracting different customers; controlling for it mechanically can remove part of the effect you want.
Keep a causal sketch of acquisition, activation, subscription and retention. Decide whether the model targets total acquisition impact or a narrower effect conditional on product behaviour. Do not combine both interpretations in one channel ROI table.
Make a decision robust to uncertain lifetime value
Evaluate a limited budget shift under conservative, central and optimistic retained-value assumptions. Include sales or support capacity where it constrains growth. A channel that wins only under the most optimistic retention extrapolation is a candidate for a smaller test, not necessarily a large allocation.
Use the budget allocation guide to compare marginal rather than historical average returns. Maintain a separate cohort dashboard that tests the value assumptions as customers age. If the acquisition model improves while cohort economics deteriorate, revisit the objective before celebrating the forecast.
A useful monthly review records acquisition data revisions, conversion maturity, retention by cohort, the range of value assumptions and the spend movement approved. This makes the recommendation auditable when the business changes pricing or product packaging.
Sources and further practice
This article’s worked example is illustrative, not a measured client result. Use the linked methodology to check the assumptions before applying it.