We built a set of scripted, repeatable booking flows and ran them across a sample of hotel engines and Plekify stores, counting every click, field, page, and domain hand-off a shopper has to complete before an order is placed.
That count is the basis of what we call the friction formula: the cost of each click, field, page, and hand-off in a booking flow, added up per system, not estimated.
Across the engine flows in the sample, shoppers were typically asked for 14 to 29 fields before checkout completed. The four Plekify flows measured in the same sample all scored below 11, for a mean Plekify friction score of 10.9.
We also scored each flow on a second axis, Agent-Readiness, which measures whether a flow can be driven end to end by a shopping agent, not only by a person holding a browser and reading a screen.
Where a flow could not be completed at all, we kept it in the sample and reported it as blocked, rather than dropping it from the average. A flow that fails is still data.
Faster pages correlate with roughly a 10% conversion lift, per web.dev's own field data, and checkout design matters for the same reason: Shopify's checkout sits behind Shop Pay's saved-card network of 250M shoppers, part of what a lower friction score is worth to a hotel's direct channel.
The full methodology and raw numbers are open, including the scripts used to produce every score reported here.
Read the study — methodology, raw numbers and the scripted flows on GitHub.