Trial-to-paid conversion and the month-2 churn cliff
Recurly's 2024 State of Subscriptions puts the median trial-to-paid conversion rate at 50%, and Recurly's network research has B2B trials converting around 66% and B2C around 56.8%. Treat those as card-on-file numbers, not universal ones - cardless opt-in trials convert far lower. And treat conversion itself as the wrong finish line: a trial that converts and cancels in month 2 bought you a refund risk, not a customer. Month-2 churn after conversion is the honest signal.
This post covers what the public trial benchmarks actually measure, why opt-in and opt-out trials are not comparable, how WooCommerce Subscriptions handles trials mechanically, and the one metric that folds all of it into a fair verdict.
The published numbers disagree, and that is the story
Public trial-to-paid benchmarks range from under 20% to over 60% because they measure different trial designs on different populations. Nobody's benchmark is your benchmark.
The figures we could verify against primary sources, all from Recurly's network:
| Figure | What it measures | Source |
|---|---|---|
| 50% median trial-to-paid conversion | Recurly network, all industries, 2023 data | Recurly 2024 State of Subscriptions |
| 66% B2B, 56.8% B2C | Recurly network by segment | Recurly subscriber acquisition research |
| 46% in 2021 falling to 33% in 2024 | Free trial conversion trend | Recurly, "How to fix the free trial problem" |
Why are Recurly's numbers so much higher than the 15-25% figures quoted in most SaaS content? Because Recurly is a billing platform: a trial in its data typically starts at checkout with a payment method on file. That selects for high-intent customers and converts by default when the trial lapses. The much lower numbers usually quoted for opt-in trials (no card up front) come from agency client panels and self-reported surveys, not from any billing platform's published data, so we are not repeating them here as benchmarks. The direction, though, is undisputed across every source we checked: card-required trials convert a large multiple of cardless trials per trial started, and attract far fewer trial starts.
Also note Recurly's own trend line: 46% in 2021 down to 33% in 2024, inside one network with a consistent definition. Trials are getting weaker even where the measurement is stable.
Opt-in vs opt-out changes what "conversion" means
An opt-out trial (card collected up front) converts by inaction: do nothing and you become a paying customer. An opt-in trial (no card) converts by an affirmative purchase. The first inflates conversion and pushes the real decision into the first paid cycles. The second deflates conversion and forces the decision before revenue starts.
That is why the two models cannot share a benchmark. An opt-out trial "converts" everyone who forgot the end date. Some of those conversions were real intent; the rest surface in months 1 and 2 as cancellations, refund requests, and disputes. The same Recurly report puts the median signup success rate at 89% - about one initial subscription charge in nine fails - and a card tokenized 30 days before its first real charge has only aged since.
The conversion event is also regulated. Free-to-pay conversion is exactly what state auto-renewal laws target: California requires a reminder 3 to 21 days before the end of any free trial longer than 31 days (Bus. & Prof. Code 17602). If your trial converts by silence, the law wants the silence interrupted.
Month-2 churn after the trial is the honest signal
Take the cohort that converted, and count who is still active two paid cycles later. That one number folds trial quality, onboarding, and billing surprise into a figure that conversion rate is designed to hide.
Why month 2 and not month 1: the first paid cycle still contains administrative churn - people who meant to cancel the trial and are unwinding it through refunds and disputes. By the end of the second cycle you are counting customers who accepted two deliberate charges. And early tenure is where the mass is: Recurly's trial research notes that 66% of all cancellations happen within the first 12 months (45% within six months for food subscriptions), and Baremetrics finds monthly subscribers three times more likely to cancel within the first 90 days than later.
Here is the structure of the argument, with illustrative percentages (the numbers are made up to show the math; the trial-start ratio is the usual reason teams pick cardless):
| Flow | Trial starts | Converts | Still active after month 2 |
|---|---|---|---|
| Opt-out (card up front): 60% convert, 35% of those gone by month 2 | 100 | 60 | 39 |
| Opt-in (no card): 30% convert, 10% of those gone by month 2 | 200 | 60 | 54 |
Same conversions, very different businesses. Judged on conversion rate, the opt-out flow wins 60 to 30. Judged on paying customers surviving month 2 - the only column that compounds - it loses. Any trial change that harvests inaction (longer trial, card required, no reminder) moves the first column up and the last column down. The metric you choose decides which flow "wins", so choose the honest one.
One term-length caveat: this signal only exists on monthly billing. If the trial converts into an annual plan, the first honest read hides for a year behind the renewal cliff - the same measurement trap covered in annual plans and the churn math of term length.
WooCommerce Subscriptions trial mechanics
WooCommerce Subscriptions configures trials per product, and three defaults change your math: sign-up fees bill at checkout even when the trial is free, the trial extends the subscription term, and product-level limits are your only serial-trial guard.
From the official product documentation:
- Free trial + sign-up fee: the fee charges immediately. The docs' own example: a $75/month product with a $200 sign-up fee and a 30-day trial charges $200 at checkout. A "free" trial with a sign-up fee is really a paid trial, and your conversion numbers should be read like paid-trial numbers.
- Free trial, no fee: the customer "pays nothing at checkout". With most gateway setups the payment method is still tokenized at signup, which makes the standard WooCommerce trial an opt-out trial. Everything above about inaction conversions applies.
- The trial extends the term. A $5/month subscription for 6 months with a 1-month trial expires after 7 months. Cohort date math should start at first payment, not signup, or your month-2 metric drifts by the trial length.
- Synchronized products push the first payment to trial end, with no proration after the trial, per the synchronization docs. Trial length plus sync alignment decides the real first-charge date - check it before blaming month-1 churn on the product.
- "Limit subscription" is the trial-abuse control. The options are No limit, Limit to one active subscription, and Limit to one of any status. The last one is the only setting that stops serial trialing on the same account.
Designing the trial to survive month 2
The levers that move month-2 retention are activation during the trial, trial length matched to time-to-value, and an honest reminder before the first charge.
Activation first. Recurly's 2026 State of Subscriptions found 52% of consumers canceled a subscription in the past year due to lack of use. A trial that ends before first value converts disengaged users (opt-out) or nobody (opt-in), and disengaged converts are exactly who falls off the month-2 cliff.
Length matches time-to-value, not convention. For content-heavy services, Recurly found over half of subscribers (55%) engaged with trials lasting a month or longer. For a box or replenishment store, the honest trial is one full shipment cycle - the customer cannot experience the product faster than you can ship it.
Send the pre-charge reminder even where no law makes you. It filters out the inaction conversions you would have refunded in month 1 anyway, and the customers it "loses" were never month-2 survivors. It is also simply required in a growing set of states, per above.
Save the in-trial cancel differently. A cancellation during the trial is its own segment with its own best offer - typically extend-trial, the highest-save, smallest-population offer type. Numbers and routing are in pause vs discount; ChurnStop ships extend-trial as a standard offer type for exactly this segment.
What to measure
Four rows, reported together, judged together:
- Trial starts - the volume the opt-in/opt-out choice actually controls.
- Trial-to-paid conversion - fine as a diagnostic, banned as a success metric on its own.
- Month-2 retention of converts - the honest signal. Cohort it by signup month; do not blend.
- Effective yield - trial starts times conversion times month-2 retention. This is the number a trial redesign has to move.
Two supporting reads: month-1 refund and dispute rate (the leading indicator that conversion is harvesting inaction) and what happens to trial cancels afterward - they are the cheapest winback audience you have, per the winback sequence playbook.
If a change raises conversion and lowers effective yield, it made the funnel prettier and the business smaller. The one-number verdict on any trial experiment is row 4, measured no earlier than two paid cycles out.
What's next
- Annual plans and the churn math of term length - the other place a pricing structure quietly rewrites your churn metric.
- Offer types reference - extend-trial parameters and when the offer fires.