Tracking the wrong metric at a funnel stage is almost as unhelpful as tracking nothing — it creates false confidence or false alarm about a stage that's actually fine (or actually broken). This is a stage-by-stage metrics framework; if you're trying to diagnose a specific leak rather than choose what to track, see my conversion funnel leakage analysis method, and for the channel and messaging layer underneath these metrics, my full-funnel campaign map.
Awareness stage
- Reach and impressions — directionally useful, but easy to inflate meaninglessly with low-quality traffic; always pair with a quality signal.
- Branded search volume — a stronger signal than reach, since it indicates people are actively seeking you out by name rather than passively seeing an ad.
- Share of voice — useful for understanding relative visibility against named competitors in a category, though harder to measure precisely than the other two.
Consideration stage
- Content engagement depth — time on page and scroll depth on key comparison or educational content, not just pageviews.
- Email opt-in rate — the rate visitors convert into a nurturable contact, a stronger signal than raw traffic volume.
- Demo or trial request rate — the clearest consideration-stage signal that someone is evaluating you specifically, not just researching the category broadly.
Decision stage
- Close rate — the percentage of qualified opportunities that convert to paying customers.
- Sales cycle length — a lengthening cycle often signals friction (unclear pricing, too many stakeholders, unresolved objections) worth investigating before it shows up in close rate.
- Customer acquisition cost (CAC) — the fully-loaded cost of winning a customer, evaluated against their lifetime value, not in isolation.
Retention and expansion stage
- Net revenue retention (NRR) — whether existing customer revenue is growing or shrinking net of churn and downgrades, arguably the single most important metric for a subscription business.
- Churn rate — tracked separately for voluntary (customer choice) and involuntary (failed payment) churn, since they require entirely different fixes.
- Lifetime value (LTV) — feeds directly back into what CAC and CPA targets should be at the top of the funnel, closing the loop.
| Stage | Primary Metric | Common Vanity-Metric Trap |
|---|---|---|
| Awareness | Branded search volume | Raw impressions with no quality or intent signal attached |
| Consideration | Demo/trial request rate | Pageviews without engagement depth or opt-in context |
| Decision | Close rate and CAC vs. LTV | Deal count alone, without accounting for deal size or close rate |
| Retention | Net revenue retention | Total customer count, which can grow while revenue per customer shrinks |
The vanity vs. actionable distinction
A metric is actionable if a specific, identifiable action would move it and you'd know why. Raw impressions, follower counts, and total pageviews rarely meet that bar on their own — they're fine as context, but shouldn't anchor a funnel-stage report without a paired quality or conversion signal next to them.
FAQ
What's the single most important funnel metric for a subscription business?
Net revenue retention (NRR) is generally the most important metric for a subscription business, since it captures whether existing customer revenue is growing or shrinking net of churn and downgrades — a business can add new customers steadily while NRR quietly erodes the base, which is a more serious long-term problem than a slow new-logo month.
- NRR reflects the health of the existing customer base, not just new acquisition.
- A business can look healthy on new-customer count while NRR signals an underlying retention problem.
Why should voluntary and involuntary churn be tracked separately?
Voluntary churn (a customer choosing to leave) and involuntary churn (a failed payment or expired card) require entirely different fixes — voluntary churn points to product, pricing, or value problems, while involuntary churn is usually a billing and dunning workflow issue. Blending them into one churn number hides which problem is actually driving the loss.
- Involuntary churn is often fixable through better payment retry and dunning workflows alone.
- Voluntary churn requires product or value-proposition investigation, not a billing fix.