Quick answer
A channel quality score summarizes whether a traffic source produces customers who get value and stay. It is a decision aid, not a universal metric; disclose its weights and always show the raw numbers behind it.
Key takeaways
- Missing retention time is not zero retention.
- Publish raw components and weights beside a custom score.
- Use a score only when it supports a clear decision.
Choose a small set of outcomes
Start with qualified signup rate, activation rate, paid conversion, and early retention. Add acquisition cost when reliable cost data exists. Avoid including many correlated metrics, such as both pricing clicks and checkout views, just to create an impressive composite.
Use enough observations to avoid unstable rankings. A channel with two visitors and one customer should not automatically beat one with hundreds of conversions.
Construct a transparent score
An illustrative model could weight activation 35%, paid conversion 35%, and thirty-day retention 30%, after normalizing metrics to comparable scales. The weights express business priorities; they are not scientific constants.
Publish the components next to the score. If retention has not matured for a new cohort, mark it incomplete instead of inserting a convenient zero or maximum.
Segment by campaign intent
Brand search, education articles, creator sponsorships, and remarketing appeal to different audiences. A single channel average may hide the fact that one campaign is working while another is wasting spend.
Compare the same observation windows and consistent attribution rules. Account for refunds or low-quality payments when reviewing commercial outcomes.
Use the score as a conversation starter
If community traffic wins on activation but loses on initial purchase, inspect buying cycle and trial length. If paid search wins on purchase but churns quickly, review promise-versus-product fit before increasing budget.
Refuse to score an immature cohort
Imagine two illustrative channels, each with twenty trial accounts. One cohort is forty days old; the other is four days old. A thirty-day retention component is observable for the first cohort but not the second. Entering zero for the second would confuse missing time with observed churn.
Use a worksheet with raw counts, eligible accounts, cohort start, evaluation date and a status such as complete or still maturing. Calculate a composite only when the components share a defensible observation window. Keep the weighting formula alongside the result, and retain the previous formula when the business changes it.
Revfaro can provide tracked goals, acquisition context and recorded revenue. A custom quality score or account-retention model belongs in your own analysis unless explicitly supported by a tool. Use goals to define observable milestones, then ask whether the score changes an actual decision. If every channel receives the same action regardless of score, simpler component reporting may be more useful.
Related reading
- Landing Page Attribution: Which Page Really Starts a Customer Journey?
- Email Campaign Analytics: From Newsletter Click to SaaS Subscription
- UTM Parameters Explained: A Clean Tracking System for SaaS
Frequently asked questions
Should every team use the same channel weights?
No. Weights should follow the company's stage and economics, and changes should be documented.
Can a quality score replace CAC or LTV?
No. It helps prioritize investigation but does not replace unit economics or retention analysis.