In-session personalization makes your search results more relevant to what each shopper is currently interested in. By analyzing the most recent products shoppers have viewed, Elevate adjusts search rankings in real time — helping customers find what they're looking for faster and with fewer refinements.
Why personalization?
The feature is shaped by well-established search research and customer insight — focused on improving precision through personal context. You can explore the full story in the dedicated article In-session personalization in Search: research and insights.
How it works
When a shopper performs a search, Elevate considers what that shopper has recently interacted with to adapt the search results.
-
Identify recent behavior
The system retrieves the most recently viewed product.
-
Boost products with behavioral affinity
- Products that are frequently clicked or purchased together with the recently viewed product receive a ranking boost.
- These relationships come from aggregated click-purchase associations.
-
Expand context through product similarity
To include a wider set of potentially relevant results, Elevate also uses the Alternatives Recommendation algorithm. This hybrid model combines:
- Behavioral data
- Attribute similarity
- Image similarity
The result is a search ranking that automatically prioritizes items similar or complementary to what the shopper just explored — even across different query intents.
Proven results
A/B tests show that in-session personalization for search consistently improves:
| Metric | Result |
|---|---|
| Precision | Shoppers click higher-ranked results more often |
| Engagement | Total search result clicks increase |
| Efficiency | Search refinements (new or modified queries) decrease |
These gains occur without reducing total sales or product visibility.
How to set it up
- Go to Experience > Pages > Settings > Autocomplete & Search Settings.
- Scroll down to the Search Result Page section and locate the Enable in-session personalization for search toggle.
- Switch the toggle On.
No additional data feeds or configuration steps are required.
Prerequisites
There are no prerequisites. In-session personalization uses data that Elevate already collects for recommendations and search ranking. No new data collection, API, or infrastructure changes are introduced.
Example use cases
| Industry | Shopper behavior | Personalized outcome |
|---|---|---|
| Fashion | A shopper views jeans and then searches for "T-shirt." | Results prioritize casual tops and items often bought with jeans. |
| Home & Living | A shopper views a sofa and then searches for "table." | Search results prioritize coffee tables and other living-room products. |
| Beauty | After viewing a hydrating serum, the shopper searches for "cream." | The list highlights moisturizers with similar ingredients or routines. |
Edge cases and dependencies
- Works best when a shopper has viewed at least one product.
- If no session data exists, Elevate automatically falls back to your normal ranking model.
- Session data is cached in real time; the system degrades gracefully if no context is available.
- Applies only to full search results (not autocomplete or type-ahead).
How to monitor and analyze the effects
Once personalization is enabled, you can monitor its impact in Elevate via the Search summary report, which shows how in-session personalization influences search behavior and shopping efficiency. Read more in the dedicated help article.
Security and privacy
No security or compliance risks have been identified with this feature.
- No new data storage, APIs, or infrastructure components are added.
- All behavioral data remains processed within the same framework already used for personalization recommendations on pages.
- In-session personalization for search extends this logic to search results, using the same behavioral signals and retention policies.
- As with page personalization, no user credentials or directly identifiable information are stored or exposed.
You may wish to confirm that your existing transparency or consent wording for behavioral personalization on pages also explicitly covers its use in search. In most cases, no update is required — but if your privacy notice separates "search" and "product listing" contexts, you may wish to align the language for consistency.