Ecommerce Personalization: The Black Box Problem and Its Solution

Anyone who has ever shopped online has been in this situation at least once (in fact, more than once). Something that you never intended to buy suddenly appeared in the “recommended products”. How did it end up in your recommendations? It’s just that “smart” Big Data has once again decided everything for you, for the marketer of this store, and for everyone in general.

A typical black box

 

With an unpredictable result. It’s impossible to calculate how much this harms your sales, since recommendations are calculated individually for each visitor. And no one knows what your customers actually lebanon phone number data see in the “recommended”. However, there is a problem.

mobile phone number library

Everyone already knows about personalization, and everyone knows that it is good for sales and customer loyalty. But there is a caveat here – this only applies to the personalization that works correctly.

What do I mean by “right/wrong”? Any online store is about industry specifics. This means that the behavior of visitors is determined by very individual patterns that are valid only for this market. Or even for this site.

For example, look at any children’s store

 

There is such an obvious characteristic of the product as the age of the child. This is, one might say, a determining factor, it sets the behavior of the parent-buyer. If your child is three months old, you do not look at metal construction sets, this is logical. Accordingly, personalization in the store should work in such a way as to take this factor into account.

But on the other hand, there are manufacturers of product recommendation engines. You connect one to the store, and you get personalization. They, in the vast majority, have never heard of how to make a balance sheet in detail industry specifics. The key task of such a business is to build the most universal product so that it works more or less normally with children’s goods, and with auto parts, and with cosmetics.

As a result, something like this can happen to the store:

Why toys for children aged 4-5 ended up in related products – I’m sure data on no one knows. To be fair, it should be noted that if you scroll through the product description, you will find a block called “similar products” at the very bottom of the page. And there is something that looks like the truth:

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top