Time-limited discount for trial users: 1.87% vs 2.08%, inconclusive
Commentary byLiftArchiveOur analysis, not theirsShowing a time-limited discount to all logged-in trial users moved purchases from 1.87% to 2.08%, a gap too small to be statistically significant.
Patrick McKenzie ran this two-week test on Bingo Card Creator during a Halloween promotion and published it on the Kalzumeus blog in 2010. We count it as inconclusive, and after the $5 lost on each discounted sale, full price made more money.
- Purchases were 2.08% with the discount and 1.87% without it, about 11% higher in relative terms.
- Not statistically significant, so this is treated as a flat result.
- Each discounted sale earned $5 less, so the full-price version did better on revenue.
The change
The variant adds a deadline discount to the pricing page.
Simplified mockup drawn by LiftArchive from the shared screenshots.
Control
Logged-in trial users saw no discount.
Variant
Trial users saw a time-limited discount promoted on the first screen after login and on the pricing page.
Why it likely came out this way
Our reading of the result, not the practitioner's claim.
- New trial users had not yet decided the product was worth paying for, so a lower price did little to sway them.
- The deadline was set a few days after the peak buying window, which likely weakened the sense of urgency.
- The same discount worked very well in email to past accounts. Those people already knew the product and needed a reason to come back.
Caveats: Two-week test with no sample size shared, and the gap was below statistical significance. It ran during a seasonal peak in 2010.
Who ran it
Patrick McKenzie ran this test and published the results (Kalzumeus (patio11) blog). The numbers are theirs; the summary and commentary on this page are ours.
Read the original case study on Patrick McKenzie →
How to run this test on your site
- Hypothesis
- If we show trial users a time-limited discount after login and on the pricing page, then more will buy because the deadline gives them a reason to act now.
- Primary metric
- Trial-to-paid conversion rate
- Guardrails
- Revenue per trial user
- Refund rate
- Full-price purchases after the promotion
- Who should test it
- SaaS products with a seasonal buying window or a self-serve upgrade path. If your discounts already work for reactivating old accounts, test them on new trials separately rather than assuming the same lift.
Questions
Do discounts increase SaaS trial conversions?
Not always. In this test, a discount for all trial users gave a small, non-significant lift.
After the lower price, it made less money than full price.
Who should get discount offers?
Lapsed or inactive accounts can be a better target.
Here, the same discount worked well by email to past accounts but not on new trial users.
How should I set a discount deadline?
Match it to when your customers actually decide.
A deadline that falls after the real buying window gives people less reason to act now.