The situation
A luxury resale marketplace listed a very large number of pre-owned luxury items, most of them one of a kind. Shoppers arrive, scan a page of products and decide quickly whether anything catches their eye.
The problem
When a catalogue is that big, the order of the products matters a lot. Items near the top get seen, and items further down may as well not exist. Too many shoppers were leaving product pages without finding something they liked, and that is lost sales for the marketplace and a missed chance for sellers.
What we did
The team looked at how shoppers had behaved in the past: which products attracted attention, in what setting, and which were passed over. From that, a model estimated how likely each product was to catch a shopper's interest.
The marketplace then used those estimates to decide the order of products on the page, putting the most promising items first. It's a bit like a good shopkeeper arranging the front window: the aim is to show what people are most likely to want.
The outcome
The share of visitors leaving product pages fell by 6%. More shoppers stayed and looked around, which gave more of the catalogue a chance of being seen.
What it meant for the business
Better visibility for the right items helps buyers find things sooner and gives sellers a fairer chance. It also showed the team that behaviour data they already collected could directly improve the shopping experience, with no new tools for shoppers to learn.
Could this work for you?
If your site has more products, articles or listings than fit on one screen, the order you show them in is a decision you can improve with your own data.
Related services: Predictive modelling and forecasting, Product analytics
This project was led by our founder, Achyutanand Jha, in an earlier role. The client is not named, and figures are those reported at the time.