The prevailing ecommerce playbook was written for fast, low-consideration purchases: optimise the first session, shorten the funnel, retarget hard, measure on last click. Applied to a £1,800 sofa or a whole-room specification, it quietly misprices everything.
What actually differs
- Consideration runs for weeks or months, across devices, often with a partner involved in the decision.
- Baskets are large and infrequent, so small conversion-rate movements swing contribution more than volume does.
- Samples, showrooms, trade accounts and phone orders mean a meaningful share of demand converts off-site.
- Delivery, lead times, returns and assembly are commercial objections, not shipping details.
Consequences worth acting on
Attribution windows built for a two-day cycle will systematically undercredit the channels creating demand and overcredit the ones capturing it. Site work should be judged on progression — sample requests, saved rooms, spec downloads, showroom bookings — as much as on checkout.
In considered categories, the job of the site is rarely to close the sale today. It is to stay in the shortlist.
Where AI discovery lands hardest
Long-consideration buyers are exactly the people asking assistants to compare options, and considered categories are where specification, availability and delivery facts decide eligibility for a recommendation. Structured, consistent commercial data matters more here than in almost any other retail vertical.
None of this means the fundamentals change. It means the sequence and the measurement do — and copying a fast-moving consumer playbook is the most expensive shortcut available.