Adding an on-site Stripe card option to checkout: purchases flat
Commentary byLiftArchiveOur analysis, not theirsAdding an on-site Stripe credit card form next to PayPal and Google Checkout did not change total purchases for Bingo Card Creator.
Patrick McKenzie ran this checkout test and published it on the Kalzumeus blog in 2012. About 90% of buyers picked the card option once it was offered, but they mostly shifted from other methods rather than adding new sales.
- No significant difference in total purchases between the two versions.
- When the card option was offered, about 90% of buyers chose it, with roughly 5% each for PayPal and Google Checkout.
- No sample size or conversion rates were shared.
The change
The variant adds an on-site card form next to PayPal and Google Checkout.
Simplified mockup drawn by LiftArchive from the shared screenshots.
Control
Checkout offered two payment methods: PayPal and Google Checkout.
Variant
Checkout offered PayPal, Google Checkout and an on-site credit card form powered by Stripe.
Why it likely came out this way
Our reading of the result, not the practitioner's claim.
- People who wanted to pay were already paying with the existing options, so a new method mostly moved buyers around.
- The card form was quickly built, which may have limited any gain from keeping buyers on the site.
- A preferred payment method changes how people pay, not always whether they pay.
Caveats: No sample sizes or rates were shared. The data is from 2012, and the payment options tested have changed since then.
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 add an on-site card payment form next to our off-site payment options, then more people will complete a purchase because they can pay without leaving the site.
- Primary metric
- Completed purchases per checkout visitor
- Guardrails
- Payment method mix
- Payment failure rate
- Refund and chargeback rate
- Who should test it
- SaaS and software sellers whose checkout only offers off-site wallets or redirects. If you already take cards on-site, adding one more method is unlikely to move total sales much.
Questions
Does adding a credit card option increase sales?
Not always. Here, about 90% of buyers chose the card option, but total purchases did not change.
Should checkout keep people on-site?
On-site payment gives you control over the experience.
In this test, though, it did not lift sales on its own.
What should I measure when adding a payment method?
Measure total completed purchases, not just how many people use the new method.
A popular new option can simply take buyers from the old ones.