+201% Meta revenue.
How Luigi's Meta ads began showing each user only clothes in stock in their size, and tripled revenue in six months.
Users saw clothes that were not available in their size
Luigi is one of the largest fashion e-shops in Greece, with its own photo studio and design team. By spring 2020 the usual tools had all been tried: landing pages, audiences, creative and bidding. The easy wins were gone.
Analysing the e-shop's data showed something small but huge: a garment the system counts as in stock is in practice sold out for a size M user when only XL is left.
That gap quietly cut purchases on every click from an ad. The analysis showed the conversion rate could rise a lot if every visitor saw only clothes available in their size.
No change to the page, the basket or the discounts would have solved it, because the problem sat further back, in the product feed. What 'available' means had to be redefined for each user.
Availability by size and by user
We redesigned the Meta product feed with labels that show the availability of each size, not one flag per product. The feed refreshes from live inventory every 15 minutes.
First-party cookies record each user's size from what they filter, what they look at and what they have bought. With no third-party data, after a few visits the system knows the user's size.
On Meta, audiences were built on the user's size, the product's size and live stock. Each group sees only the part of the catalogue it can really buy at that moment.
Dynamic creative adapts to each group. So the sales lever moved from small changes on the page to the architecture of the feed itself.
Three levels of personalisation
Feed by size
The feed records stock by size and refreshes continuously, as the one source for all ad logic.
Each user's size
First-party data shows which size each user wears, with no reliance on third-party tracking.
Ads that can be bought
Every user sees only products available in their size at that moment.
The Results
Meta revenue
Revenue from Meta ads tripled in spring and summer 2020, from a change at the root of the problem.
Conversion rate
Conversion rate rose 21% across the site: more visits turned into purchases.
Orders
Orders from ads almost tripled, because every ad showed only clothes the user could actually buy.
Feed refresh
The feed refreshes every 15 minutes, so a sold-out size leaves the ads before it causes a wasted click.
Want to find the sales you lose in the feed?
We review your product feed and find the personalisation a classic CRO does not touch.
Start a conversation