eBay SEO Cassini
What is the eBay Cassini Algorithm?
12 years ago we were introduced to eBays Best Match Search Algorithm and 6 years later to a patent called “Methods and systems for deriving a score with which item listings are ordered when presented in search results”
We are talking about the Cassini Search Algorithm for eBay. In this video we refer to an article written by Mark Steier to explain how the search engine functions based on this patent.
How does the eBay Cassini Algorithm work?
Every search you make on eBay gives you a list of products that are ranked depending on different scores. The most important ones are Relevance-Score, Business-Score, and Quality-Score. We can find out what each of them means from an excerpt of the US patent.
According to the graph we see that the Relevance score derives from the probability of a placed order, and therefore a transaction, if the results unfolded a certain way.
Excerpt from US Patent
In summary, the score shows how likely it is to see an order being placed if Product X was to rank high for a certain keyword. That’s why the score has been given this name. It’s about the relevance of a product for the given keyword.
The Business Rule Components represent a promotion or demotion factor determined based on the evaluation of one or more business rules. It also includes seller data that is used for rank correlation along with the optimisation of listings.
Listing Quality Score
The Listing quality which is about typical parameters such as the number of views, number of transactions, prices, shipping costs, and how long the listing has been online, Mark Steier offers an interesting graph that illustrates why sometimes it makes sense to delete a listing that is active for a long time and publish it again.
eBay Cassini Algorithm Graph
Here we see a formula that consists of three Parameters.
The Predicted Listing Score, the Observed Listing Score and a time-based Parameter called Weight.
This Parameter changes over time and has a range from 0 to 1.
Parameters of the Listing Quality Score
The value is 0 once an eBay Listing has been published. Then the Listing Quality score is composed of the price, condition, shipping methods, etc. The longer the listing is online, the higher the weight value gets. If it reaches its maximum of 1, then the predicted listing score gets multiplied by 0 which gives a result of 0 again.In this case the entire Listing Quality score is based only on the Observed Listing score.
If there is no transactions or clicks on a listing then its rank is dropping. That’s why it makes sense to delete a Listing in such a case and set it up from scratch.