Reverse Trial vs. Freemium: Reading Two Cohort Reports
Objective: Given two real cohort performance tables (a reverse-trial cohort and a standard-freemium cohort for the same self-serve tool) and two downgrade-email variants, correctly diagnose which model is healthier on more than conversion rate alone, and which email variant follows the lesson's messaging guidance.
You're the growth PM at The Trade Desk, the programmatic advertising DSP, deciding whether to keep a self-serve campaign-optimization tool on standard freemium or move it to a reverse trial ahead of a broader rollout.
Compare two cohort tables across conversion rate, time-to-convert, and 90-day retention (not conversion rate alone), then evaluate two downgrade-notice email drafts against the lesson's loss-framing guidance.
Before you start
What you'll need
Free path (everything below is enough to finish)
Both comparisons are simple enough not to need a specialized tool
Paid upgrades (optional, faster/deeper)
Replaces the static comparison with a live, ongoing cohort dashboard
The process
2 steps
Step 01 of 02
The lesson's Measuring Success section says to track trial-to-paid conversion, time-to-convert, and post-conversion 90-day retention together, since reverse trials can convert faster but churn faster too.
Cohort A (reverse trial): 22% conversion, 68% convert before downgrade, 71% 90-day retention. Cohort B (freemium): 4% conversion, no downgrade event, 89% 90-day retention. Which cohort is actually healthier for the business?
Procedure
- Cohort A: 22% conversion x 71% 90-day retention = 15.6% of original signups are still paying at day 90
- Cohort B: 4% conversion x 89% 90-day retention = 3.6% of original signups are still paying at day 90
- Cohort A still wins on net retained customers despite the lower retention rate, because the conversion gap (22% vs 4%) outweighs the retention gap (71% vs 89%)
- Flag the retention gap itself as the actionable finding: an 18-point retention drop suggests some Cohort A converts are loss-averse rather than genuinely activated
Cohort comparison (n=1,000 each) A (reverse trial): 220 convert, 156 still paying day 90 (15.6% net) B (freemium): 40 convert, 36 still paying day 90 (3.6% net) Gap to investigate: A's 71% 90-day retention vs B's 89%, why do reverse-trial converts churn more?
Healthy
The comparison weighs conversion x retention together, and treats the retention gap as a finding to investigate, not a reason to reject the reverse trial outright.
Unhealthy
Picking Cohort B because 89% looks like a better number than 71%, without multiplying through to net retained customers.
What this means
Cohort A delivers over 4x the net retained-paying customers of Cohort B, but the retention gap says the reverse trial's post-conversion nurture and product fit still need work.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Retention comparison stops at the headline 71% vs 89% numbers | Segment Cohort A's day-90 churners by time-to-convert to see if fast converters churn more than slow ones | 30 min |
Step 02 of 02
The lesson says the downgrade notice should arrive at least 3 days before the actual downgrade and should frame what the user will lose, not what they need to pay, citing 'Your dashboard will become read-only on June 17' as the model.
Draft 1 arrives on the downgrade day and reads 'Upgrade now to keep your plan active!' Draft 2 arrives 4 days before downgrade and reads 'Your campaign-bid-optimization automations pause on August 24 unless you upgrade.' Which draft follows the lesson's guidance, and on which two dimensions does the other one fail?
Procedure
- Draft 1 fails timing (arrives on the downgrade day itself, giving zero advance warning) and fails framing (generic 'upgrade to keep your plan' instead of naming the specific lost feature and date)
- Draft 2 passes timing (4 days advance, exceeds the 3-day minimum) and passes framing (names the specific feature, 'campaign-bid-optimization automations', and the exact date)
- Flag that Draft 1 is a Mistake 3 case study from the lesson's Common Mistakes section, weak downgrade messaging
Draft 1: FAILS timing (0 days advance), FAILS framing (generic) Draft 2: PASSES timing (4 days advance), PASSES framing (names feature + date) Verdict: Draft 2 matches the lesson's model example
Healthy
The winning draft names the specific feature and exact date, sent with 3+ days of advance notice.
Unhealthy
A same-day, generic 'upgrade now' email with no reference to what is actually being lost.
What this means
Draft 1 is the exact failure pattern the lesson's Mistake 3 warns underperforms, timing and framing are both correctable without touching the trial mechanic itself.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| Downgrade emails are generic and same-day | Rewrite using Draft 2's pattern: name the specific feature, exact date, 4+ days advance | 30 min |
Final deliverable
A cohort comparison table with the net-retained-customer calculation for both models, plus a scored evaluation of the two downgrade-email drafts.
See a reference example
Canva, downgrade notice evaluation (excerpt) Draft sent day 27 of 30-day trial: 'Your background remover and brand kit tools return to standard on day 30.' Timing: PASSES (3 days advance) Framing: PASSES (names specific tools and the exact day)
Success criteria
You're done when you can:
- Correctly computes net retained-paid-customer rate for both cohorts (15.6% vs 3.6%) and recommends Cohort A despite its lower retention percentage
- Identifies the 18-point retention gap as a finding to investigate, not a reason to reject the reverse trial
- Correctly scores Draft 1 as failing both timing and framing, and Draft 2 as passing both