Product Sense Start a project
All notes

Experimentation

What a checkout test can tell you

A plan selection test at 5 Axis Health showed why it pays to measure what people buy as well as how many complete checkout.

By David Spiegel

Decide what the experiment is for

At 5 Axis Health I tested a plan selection page that changed the default term, showed prices per month and added a six-month option. The question we answered was whether the full page worked better than the original.

If your decision is whether to ship a page, testing a complete page can be the right choice. If you need to know whether a default can work elsewhere, vary the default on its own. The design of the test should match the decision you have to make.

Keep the purchase and the order in view

We saw purchase conversion rise 30% and average order value rise 76% in the same test. Looking at both told us more than looking at either alone. A different experiment could make a cheaper plan easier to buy and raise conversion while bringing in less revenue.

Before touching the page, write down the metrics you will use to make the call. For a plan test I’d use revenue per visitor with conversion and order value alongside it, then check which plans people chose.

Related work5 Axis Health+30% purchase conversion, +76% average order value

Show me where people stop.

Send me a link and tell me what you’ve noticed. I can start there.

Or write to david@heyspiegel.com