Three in ten eBag orders include a product found in a recommendation.
An online grocery in Bulgaria made recommendations part of the weekly shop, using only its website. No email, no push, no ads.
- Online grocery · Bulgaria · custom platform
- Client since 2019 · website only
- Store order data · click-attributed · no holdout
An online grocery with a long, repeated basket
eBag is an online grocery with fresh food, pantry, drugstore and pet products, delivered in time slots. In September 2026 it took about 180,000 orders a month, about 13 items each.
The store grows when shoppers add a few more of the right items and find them fast among tens of thousands of products. The same row of products for everyone wastes the few places where that happens.
eBag uses one part of Releva: recommendations on its website. From January to August 2026, traffic influenced by Releva was 10% of visitors and 69% of buyers.
Four places, each decided for each shopper
Cart page
Before checkout, each shopper sees items from their own past orders that are not in today's cart. This is eBag's busiest recommendation.
Themed home-page rows
The team names a theme, such as seasonal fruit or pharmacy, and sets one filter. The agent picks the products inside it for each shopper.
Category and product pages
“Popular in this category” on category pages and “People also bought” on product pages.
Few rules
Only available products, the shopper's language and one theme filter. No manual boosts on price or brand.
eBag's team has launched 372 themed rows since January 2025 and rotates them, with 16 live today.
The cart page earned most of it
1 June to 28 September 2026. Revenue attributed to recommendations, by placement.
| Placement | What runs | Orders per 100 clicks | Share of recommendation revenue |
|---|---|---|---|
| Cart page | Items from this shopper's past orders, not yet in the cart | 20.7 | 68% |
| Home page, 206 themed rows | The agent's pick for this shopper within each theme | 14.8 | 13% |
| Category pages | Popular in this category | 15.8 | 12% |
| Product pages | People also bought | 5.2 | 8% |
Since June 2025, orders with a recommended item rose from 20% to 30%
The themed rows went live through 2025. Other things changed too, so this is a trend, not a controlled result.
| Measure | June 2024 | June 2025 | August 2026 |
|---|---|---|---|
| Orders in the month | 74,582 | 113,983 | 177,681 |
| Orders with a recommended item | 28.0% | 20.3% | 30.4% |
| Recommendation clicks per 1,000 page views | 31.5 | 34.5 | 47.9 |
| Share of online revenue attributed to recommendations | 3.6% | 2.7% | 4.3% |
How we counted
Attributed, not measured lift. An item counts when the shopper clicked it in a recommendation and then bought it. eBag has no holdout group, so these figures are not lift.
Influenced traffic. Conversion is the share of a month's visitors who bought that month. Influenced visitors opened at least one page through Releva. They chose to click, so part of the gap is intent.
One order counts once. An order with three recommended items counts once in the 30%. April and May 2026 are left out: a counting error credited 5.7 items per order instead of the usual 2.2.
Next. A holdout group on the cart block would turn attributed revenue into measured lift. Releva sets one up automatically.
Get the method and setup notes
How each number on this page was counted, the definitions, and what was switched on in the store, in order. Enter your work email and the notes open; we also send you the link.
Thanks. Open the method and setup notes.
We use your email to know who reads our case studies, and to send the link. Privacy policy
See what the agent would add to your shoppers' carts
Thirty minutes, with your store open. We will show you where the agent would sell, and how we would measure it.
Book 30 minutes
You'll talk to one of us.