Next step
Twenty five minutes, about how you run experiments today. No access, no commitment.
33 tools on the map. 8 are wired today, each with a page saying exactly what it reads and what it writes.
Show delivery and the order total from the first checkout step, so the price stops changing at the payment step, without lowering completed orders.
Delivery is priced at the delivery step and the order total is only complete at the payment step, so the number a shopper is asked to pay is not the number they saw when they started. A shopper who has already given an address and their details meets a higher total on the last screen, and some of them leave there.
Extra costs added at checkout, for delivery, tax and fees, are the most common reason shoppers give for abandoning an order they had started, and not being able to see the total order cost upfront is on the same list [1]. The change under test does not remove the cost. It moves the moment the shopper learns it, from the last screen to the first, so the decision is made before the effort is spent. That is also how it can fail: the same transparency can move the abandonment earlier rather than remove it, which is why the primary metric is completed orders per checkout start and not the completion rate of the payment step. Nothing on this page reports a result.
Sessions that start checkout with a physical order on the web storefront. Digital-only orders are excluded, because they carry no delivery cost and their total does not change, and orders placed in the mobile app are excluded, because the summary there is a different component.
Showing the delivery cost and the order total from the first checkout step raises completed orders per checkout start, because a shopper who learns the full price at the start decides on the real number rather than on one that changes at the last screen, and a decision made before the effort is spent is less likely to be reversed at the payment form.
This is an example. No expected impact is stated because there is no evidence to state one from.
The order summary component reads the default delivery method for the shopper's country and shows its cost and the order total from the first step, behind the customer's existing feature flag. Prices, delivery charges, tax rules and the payment step are untouched: what changes is where the number appears, not what it is. Exposure is recorded when the first checkout step renders, so a shopper who leaves at the summary counts in their arm.
Ship when completed orders per checkout start clear zero at the planned sample with revenue per checkout start and the refund rate holding. Revert on a revenue or refund breach, at any point. Otherwise continue to the planned sample without peeking, and call it inconclusive if it does not clear zero there.
Twenty five minutes, about how you run experiments today. No access, no commitment.
