(Pilot) Ratings & Reviews

Overview

The ratings_and_reviews dataset contains ratings and reviews that customers have submitted for your products sold on Zalando. The data includes both 'native' (submitted directly through Zalando sites and apps) and 'syndicated' (extracted from the brand's own website using a third-party aggregator) reviews --- No distinction is made between these types or review in the data. Reviews are provided in the data at model_sku level (i.e. the top level in the product hierarchy --- see Understanding Products). The config_sku in a review record indicates the colour configuration against which the review was submitted (i.e. the one the customer bought), however customers browsing Zalando sites and apps are shown all reviews for the model_sku regardless of which config_sku they select.

This data enables you to monitor customer feedback on your assortment, track ratings and review volume over time, analyse sentiment, and identify products with satisfaction or quality issues. The dataset only contains entries that have passed moderation. Rating-only entries (is_ratings_only = true) have no review_title or review_text. All data is governed by the Ratings and Reviews Terms of Use.

Property	Description
Data granularity	One row per published review (or rating-only entry). Reviews are provided at model_sku level and may appear under all config_sku variants belonging to that model on Zalando sites and apps.
History available	Latest snapshot only. The table is fully overwritten on each load; no historical partitions are kept.
Update frequency	Weekly on Tuesday 12:00 pm UTC.
Data retention	7 days (overwritten weekly on Tuesday).
Primary keys	review_id
Partition columns	None
SLOs	Every Tuesday at 12:00 pm UTC
Notice period for changes	See Versioning and Deprecation Policy
Support	Support available via partner-care@zalando.de. No 24/7 support available.

Table Reference

The table can be accessed using the following data reference from delta sharing client applications:

ratings_and_reviews_share.direct_data_sharing.ratings_and_reviews

Schema

Column name	Format	Description
account_id	string	The Zalando Partner zDirect Account ID.
model_sku	string	Zalando's reference code for a SKU (Stock Keeping Unit) at the article level.
config_sku	string	Zalando's reference code for a SKU (Stock Keeping Unit) at the article and colour level. A config SKU might have multiple sizes (referred to as Simple SKUs). This is the specific config SKU that was reviewed. Customers browsing Zalando sites and apps are shown all reviews for the model_sku regardless of which config_sku they select.
merchant_sku	string	Partner's reference code for a SKU (Stock Keeping Unit) at the article and colour level.
review_id	string	The review identifier.
review_published_at	string	The timestamp when the review was published (passed moderation), stored as a string in ISO 8601 format (e.g., 2018-03-11T17:09:01.000Z).
review_language	string	The language of the review (locale example: en_US, de_DE).
is_ratings_only	boolean	Indicates whether the review is a rating only (no review_title and review_text).
rating	int	The rating value, ranging from 1 to 5.
review_title	string	The review title.
review_text	string	The review text.

Example Measures

Below, we provide key KPIs that you can calculate using only the table ratings_and_reviews_share.direct_data_sharing.ratings_and_reviews. To ensure you calculate the values correctly, please use the code examples below.

Average rating and review volume per model

Average star rating and number of reviews for each model (model_sku). Since reviews are displayed at model level, use model_sku for review KPIs rather than config_sku.

SELECT
    model_sku,
    COUNT(*)              AS review_count,
    ROUND(AVG(rating), 2) AS avg_rating
FROM ratings_and_reviews
GROUP BY model_sku
ORDER BY review_count DESC;

Rating distribution

Count of reviews per star rating, to see how feedback is spread from 1 to 5.

SELECT
    rating,
    COUNT(*) AS review_count
FROM ratings_and_reviews
GROUP BY rating
ORDER BY rating;

Changelog (non-breaking changes)

No non-breaking changelog entries have been documented yet.

Contact Support