Recommendations
By default (API)
By default, ABTasty Recommendations requires to store, analyze and use the following data to create recommendation strategies:
- catalog
- analytics :
- pageviews
- transactions
This data shouldn’t be considered as “personal data” as it can’t be linked in any way to any individual.
Here are the columns that are stored and used to create recommendations strategies :
- catalog : SKU, price, url, stock… and any columns the client would like to provide at the SKU level
- pageviews : session_id (a random uuid that isn’t linkable to the visitor), url, timestamp
- transactions : transaction_id, SKU, timestamp
A recommended list of products is then retrieved by calling an ABTasty Recommendation strategy with parameters such as the product displayed on the page, the products in the basket…
Those parameters can’t link to an individual as well.
- ABTasty should be added to the list of providers using analytics data
- No need for visitor consent to show default recommendations
Personalized (API + Widget)
ABTasty Recommendations enables personalized recommendations based on a visitors navigation.
In this case, ABTasty requires to add a javascript tag to the website, Recos Tag, in charge to :
- store a visitors’ navigation in its browser (indexedDB)… And in its browser only.
- use this navigation to call Recos Strategies
For instance, Recos Tag can store the products viewed by the visitor during his current and previous sessions, then call a recommendation strategy with the list of product viewed by the visitor.
This personalization method uses personal data but is considered as extremely respectful of visitor’s privacy : visitor’s navigation is never stored in ABTasty servers (contrarely to conventional cookie-based method) and deleting his cache insures a complete deletion of his navigation.
- ABTasty should be added to the list of providers using analytics data
- Depending on how critical your recommendation banners are, you can either decide to place ABTasty Recos Tag behind a “Marketing finality consent” or consider them as “Necessary”
Monitored
In case you want to monitor the performance of your recommendations banners through the interface, ABTasty Recommendations will send custom events to your analytics tool and retrieve those custom events from your analytics tool.
By sending custom events to your analytics tools instead of sending it directly to ABTasty servers, our clients can cross those events with others, create their own reports in their analytics tools, access the most granular data…
To monitor the performance of a recommendation, ABTasty Recommendations crosses the custom event properties and the transactions based on the session_id or user_pseudo_id retrieved from the analytics tool.
Merchandizing
Merchandizing requires the same data as the recommendations by default.
- ABTasty should be added to the list of providers using analytics data
- No need for visitor consent to show default recommendations
Hosting
The data is stored in ABTasty servers, hosted in one of the three GCP zones depending on the site location :
- Europe
- APAC
- US
Client data isn’t shared in any case and is completely siloed.