Six screens became two, and the drop-off moved

6 → 2

screens to book

3

weeks

40%

of losses at one step

Disclosure Work delivered by our founder prior to incorporation. Client anonymised at their request.

The transferable point

Before optimising a booking flow, go through it as a customer on a phone, paying with a real card. Almost everything worth fixing is visible in that one exercise and almost nobody does it.

Discuss something similar

The situation

A provider selling dated, capacity-limited experiences direct to consumers. Traffic was healthy, the marketing spend was working, and the conversion rate was not.

The problem

The booking flow ran to six screens. Analytics showed a large share of customers reaching step four and leaving. Nobody knew why, and the assumption in the business was that the price was too high.

What we found

It was not the price. Step four asked for full passenger details, including dates of birth, before showing a total. Customers were being asked to commit effort before they knew what the thing cost — and a meaningful proportion of them were leaving to check the price elsewhere, which meant a competitor's site.

We also found the flow was rebuilding availability on every step, so a customer moving through it slowly could be told at the end that their selection was no longer available. That message was being shown to real customers several times a day and nobody had reported it, because the people who saw it had already left.

What we built

The flow was reduced to two screens: select and confirm.

Price is now shown as soon as there is enough information to compute one, and it updates as options change rather than appearing at the end. Passenger detail collection moved after payment authorisation, which is where it belongs for this product — the customer has committed, and the remaining information is administrative.

Availability is held for a short window at the point of selection and released cleanly if the customer does not proceed, so the late failure stopped happening. The hold expiry is visible to the customer rather than silent.

The whole flow is a single state machine with named transitions, which meant we could finally see, per step, exactly where people were leaving.

The result

Abandonment at the old step four disappeared, because the step did not exist. Drop-off did not go to zero — it moved to the first screen, which is where it should be, because that is people deciding whether they want the product at all. That is a much healthier shape.