Whether to Require an Account at Checkout

Checkout abandonment was concentrated at a single step: the account creation screen that appeared before payment.

A mid-sized retailer with a growing repeat customer base. The marketing team had a target for registered accounts, and registration was placed before payment because that was where it captured the most people.

• The commercial target for registered customers was fixed for the year.
• Order history and returns both required an identifiable customer record.
• The existing checkout was a four-step flow that could not be rebuilt quickly.
• Fraud rules required a verified email address before dispatch.

Allow purchase to complete with an email address alone, and offer account creation on the confirmation page, prefilled from the order details, requiring only a password.

The account was needed for order history, returns and fraud checks. All of those are satisfied by an email address plus an order record. The password was the only additional thing registration provided, and it was being demanded at the moment of highest commitment and highest anxiety. Moving it after payment removed the barrier without removing the capability.

• Funnel analysis isolating the drop to the registration step rather than to payment.
• Session recordings showing users reaching registration, pausing, and leaving.
• The Baymard Institute's long-running checkout research, which consistently identifies forced account creation as a leading cause of abandonment.
• Support tickets from customers who had created duplicate accounts because they could not remember whether they had one.

• Fewer passwords captured at the point of purchase, so some customers remain guests.
• An additional record type to reconcile when a guest later registers with the same email.
• The marketing team lost the ability to claim every purchaser as a registered user.

Abandonment at the registration step disappeared because the step no longer existed. A meaningful share of customers created an account on the confirmation page, where the only remaining action was choosing a password. Total registrations were comparable, and purchases completed increased.