Ostrelo's product pages converted well, but four in ten shoppers who reached checkout left before paying. Session recordings showed the same loop: shoppers scrolled back and forth between basket and delivery options, trying to work out when a sofa would actually arrive.

Delivery dates lived in three systems and were only shown after payment. Support answered 9,000 'where is my order' messages a month.

We spent the first week in Ostrelo's warehouse and support team, then put a paper prototype in front of shoppers by Thursday. The insight held up in every round: people decide on the delivery date first and the price second.

The new checkout shows a firm delivery window on the basket, lets shoppers pick a two-hour slot before paying and carries the same date through email, app and the driver's route. We built it as 38 components in a shared library so web and both apps ship from one source.

Checkout conversion rose 18% in the first 90 days and delivery questions to support fell by 42%. Ostrelo's own team has since shipped returns and assembly booking on the same library, without us.