During regular discussion about improvements to the checkout flow, I questioned the need for the checkout Review page. We saw a certain percentage of our customers drop out of checkout on this step and we wondered if removing the page for those 'happy path' customers who could go straight from the Address & Payment page to the Receipt without any credit card errors would increase overall conversion.
The checkout Review page served 3 primary purposes:
Though the Review page performed 3 primary functions, the analytics had not been implemented to track each of them separately. We had tracking for PayPal users vs credit card customers, but no data to separate the credit card customers experiencing errors from those who were not. This lack of data should have been a red flag, but our assumption seemed entirely reasonable, so we pushed forward into design and development.
As an online shopper, I want to complete my order in the fewest number of steps, because my time is valuable
Checkout development is laborious: PCI compliance, front- and back-end field validation, error messaging and browser security must be considered at every step. Over a period of about 4 weeks, working with an off-shore team through daily standups, we rebuilt the checkout flow to ONLY require the Review page for those customers using PayPal OR receiving credit card error responses from CyberSource, our payment processor.
As part of the development, we also added the missing tracking around each of the credit card error types we received back from Cybersource.
The final flow was a thing of beauty. It was simpler for a majority of our customers, faster and had better analytics.
Sigh. Things become so clear in retrospect.
Yes, we successfully removed the need for the checkout Review page for the majority of our customers. Yes, we saved most customers a few seconds of their day and a button click. No, these changes had no effect on overall conversion.
The data flowing in over the few weeks after release indicated that the customers visiting the Review page were now only PayPal customers and those receiving credit card errors. PayPal customer continued to have high conversion rates from this page. Though we didn't know it before, it turns out the customers with credit card errors comprised all of the lost conversion on the Review page.
BEFORE:
AFTER:
This project was a clarifying moment in how incorrect assumptions created wasted development time which represented real money for the company. The intentions were good, the logic was sound, but the assumption was simply wrong. Instead of moving directly into development, the time should have been taken to validate the initial assumption. Development time to add additional error tracking would have saved approximately 95% of the overall project development time.
