No email without recommendations

[former_member711131]

Hello Experts,

We send product recommendation in one of campaign emails. We need functionality where if No Recommendation generated Exist then the email should not be send .

In email we could see Offer Recommendation provides this functionality( No email without valid offer) but couldn't find similar one for Product Recommendations.

Could you please advice what could be best way to achieve this ?

Thanks

[theodore_schachter2 · Accepted solution]

Hi Ajinkya,

There is no way to do that. The correct approach is to create a fallback.

There are many ways that you could set up your Recommendation Scenario, but here is one approach.

First:

A Recommend Step containing an optimized Query Type algorithm such as “Last Purchased Items” or “Recently Purchased Items” which uses the given recipient of the email (the contact) as their leading item. These algorithms then query for the contact's last or most recent purchases (sales orders)

Then (the fallback):

To ensure that a recommendation is provided in every email, the scenario can be further enriched in the first step with a “Top Sellers” algorithm set to low or very low importance, so that if the contact has no interactions, this algorithm can be used as a fallback, providing the top sellers in general from the your shop.

Last:

A second Recommend Step containing an Association Type algorithm such as “Often Bought Together (Interactions)”, which will take the recommendations from the preceding step as leading items and provide a personalized recommendation to the contact.

There are many different ways you could set up your scenario, but the fallback is key to avoid sending your customers an empty email.

Regards, and good luck

[theodore_schachter2]

Hi Ajinkya,

With reference to the model design I suggested above, if a given customer has no interactions from which an association can be made by the algorithm (i.e. no sales orders for products with which other customers had bought together with another product) then they would receive the fallback.

If the customer has interactions supporting an association, then they'd receive personalized recommendations.

You should be able to see the content the customer received using the Contacts app, by choosing the Interactions tab, searching their interaction timeline for the campaign email, choosing that interaction, and then checking that interaction's Content header, but you'd have to be very familiar with your data to determine if the recommendation was personalized or fallback. If you had a good idea of what your fall back algorithm usually returned, you might be able to, but the fallback recommendation might change over time depending on your overall customer interaction behavior, especially at heavy sales periods like Christmas or Black Friday, etc.

Regards,

[former_member711131]

Thanks theodore.schachter2 . Would it be possible to know which Customer Received the default recommendations and which one received the Recommendations by Recommendation Algorithm ?