How To prepare the feed file when using reccomnedations algorithm as Top by Analytics metrics

[tarun_jamana846913]

Hi I am working on a POC where I need to show most popular ancillaries based on a route. So currently I have a analytics metric Ancillaries Orders (Booking flow) and I created a dashboard in AA to visualize the data for a particular route(AUH-LHR) and then broke it down by products and I filtered the ancillaries.

Now how do I create my field and what should be the entity.id values and how should the structure of the entity be since I need to consider the route as well and it can be any route 

[kandersen]

@tarun_jamana846913 My understanding is that you would like to recommend the most popular ancillaries, based on the current route?

First of all I don't see it as an option using the Analytics data in the source for the Criteria in Target. The integration uses s.products as the key, which will require that both route and the ancillaries in s.products, which based on the screendump doesn't seem to be the case, but rather 2 seperate dimensions (disclaimer, without access to EA account i might be wrong).

This also makes me think that the routes today are not tracked as seperate SKUs in Recommendation? If not, it complicates the scenario a lot.

If routes are tracked as SKUs, then the ancillaries also need to be, and I would add meta data to the ancillaries (e.g. entity.productType = ancillaries). This would allow to create a product collection in Recommendation allowing me to configure the Criteria to only recommend products/ancillaries from this collection.

A simpler approach, which I personally would take - assuming the top ancillaries doesn't change very often (which based on the numbers you shared, I doubt is the case). I would extract the numbers from AA and create a custom criteria where i dictate the ancillaries to show in the recommendation. Again, it still requires that you're tracking routes as unique product SKUs in Recommendation.

I hope this can help you in the right direction for a solution.