Skip to content
OperationsMetric 29 of 35

Conversion Funnel

From visitor to paid: where the leaks are.

Abderrahim Ben Jazia, PhDBy · Founder of FlowMRR, previously built and sold AIS (Avance Immédiate Services), French tax-credit softwareUpdated:

Put this metric to work: explore FlowMRR's MRR movement analytics to track revenue movements, retention and forecasts from your Stripe account, or see how FlowMRR compares as a Baremetrics alternative for Stripe.

The sequential steps between first touch and paying customer, with conversion rates at each step. Visitor → Signup → Activation → Trial → Paid. The funnel exposes which step is leakiest, which is the only one worth optimising at a time. Only the last step, trial to paid, lives in Stripe; that is the step FlowMRR measures, not visits, sign-ups or activation.

Worked example

10,000 visitors → 600 signups (6%) → 360 activated (60%) → 90 paid (25% of activated, 0.9% of visitors). Compare each step with its own historical and segment baseline before deciding where to optimise; these percentages alone do not identify the biggest opportunity.

Other definitions you will see

Product-led teams often start the funnel at sign-up or at a free account rather than at a visit, and sales-led teams use lead → qualified opportunity → closed deal. Step conversion can be measured within a period or by cohort, following the same sign-ups through every step; the two differ when steps take weeks.

Illustrative comparisons

Visitor → Paid

These unsourced ranges are illustrative scenarios, not measured industry benchmarks, percentiles or recommended targets. Their qualitative labels are editorial interpretations. Compare your own cohorts using the same segment, period, currency and measurement basis.

PLG strong
1–3%
Middle scenario
0.3–1%
Friction-heavy
< 0.2%

Why it matters

The funnel is the only way to know which optimisation moves the needle. A new homepage hero might double signups but do nothing to paid; a new onboarding flow might triple activation but cap at the same paid number because the next step is the leak. Measure end-to-end, fix the worst step, repeat.

Common mistakes

  • Optimising a step that already meets its baseline instead of testing the weakest step for your segment.
  • Confusing volume with conversion (more signups + worse conversion can be a net loss).
  • Not segmenting by source — paid traffic and organic convert very differently.

Calculating it from Stripe data

Stripe sees only the bottom of the funnel: the trial and the first payment. Visits, sign-ups and activation live in your web analytics and product database, so a full funnel joins those sources to Stripe by customer email or ID. On the Stripe side, a trial start is a subscription created with a trial period, and a conversion is that subscription becoming active and paying its first invoice when the trial ends.

Classify each trial once, per subscription, with exactly one outcome. Converted: the trial ended and the first charge succeeded. Payment failed: the trial ended but the first charge did not go through, which is a billing problem rather than a decision against the product. Churned: the trial was cancelled. In trial: the trial end is still ahead. A subscription that Stripe paused at trial end because no payment method was on file is a failed conversion, not one to leave out. Exclude incomplete and incomplete_expired subscriptions, which never started, and count sign-up trials only: a free period added later to a paying subscription is not a trial start.

The rate is converted ÷ (trials started − trials still running). Dividing by every trial started, including those in progress, understates conversion for each recent cohort and makes the latest month look worst. Count per subscription, not per item, so a base plan plus seats is one trial, and value converted MRR on the headline basis. FlowMRR applies these rules to its funnel, its tables by product and by trial length, and its cohort heatmap, so all of them reconcile.

Reading the result

Group trials by the month they started and compare cohorts at the same age, since a cohort whose trials have not all ended is not finished. A falling rate with a rising share of payment failures points to card collection, for example whether you ask for a card upfront; a rising share of cancellations points to activation and onboarding. Compare conversion by trial length as well: if longer trials do not convert better, they mostly delay revenue and the decision.

Read the funnel in absolute numbers as well as rates. A step whose rate improves while the number of paying customers stays flat has usually lost volume upstream. Small numbers mislead: with a dozen trials, one customer moves the rate by several points. FlowMRR's investor report grades trial conversion only from 10 completed trials and marks the figures inconclusive below that.

Frequently asked questions

How do I track trial conversions in Stripe?

List subscriptions that started with a trial and check each one after its trial end: active with a paid first invoice is a conversion, canceled is a lost trial, past_due or unpaid is a failed first payment. Webhooks such as customer.subscription.trial_will_end and customer.subscription.updated let you record changes as they happen. Leave trials that are still running out of the denominator.

Should I require a credit card for a free trial?

It is a trade-off you can measure. Asking for a card upfront usually means fewer trial starts, with conversion happening automatically unless the customer cancels; no card means more trials and a conversion that needs an active step from the customer. Compare paying customers per visitor rather than the trial conversion rate alone, because the two designs fill the funnel differently.

What happens in Stripe when a trial ends without a payment method?

It depends on the subscription's trial settings for a missing payment method: Stripe can cancel the subscription, pause it, or create an invoice and try to collect it. A paused or cancelled subscription has not converted; an invoice that stays unpaid moves the subscription into the usual failed-payment states.

Why is my latest trial cohort's conversion rate so low?

Often because the cohort is not finished. Trials still running should be left out of the denominator; otherwise every recent cohort looks worse than it will end up. Check how many trials in the cohort are still running, and compare cohorts only at the same age.

Contact us
FlowMRR

Let's talk

A question, a technical issue or an idea for a partnership? Write to the FlowMRR team.