Personalize for new and anonymous shoppers

A common challenge in online commerce is around the cold start problem. That is when shoppers come to your website for the first time, and there is no data to leverage. Retailers across verticals ask us about this. So, here’s an example from one of our hypermarket and grocery vertical clients.

I have come to the site in incognito mode. So there’s no past information around behaviour or purchases. There is, however, contextual data to leverage, such as geolocation, my device, browser, my nearest store, or warehouse to personalize what I see on the homepage, I get inspiration around the trending products, top sellers in my locality and so on. Here, the top seller strategy has been picked based on the 150 Plus algorithms that are competing, and AI is selecting the best algorithm to use. Very contextual here is the section on Work from Home Essentials, that is likely a very popular category right now. What are the top products and popular categories today? What has been the most sold today? The popular products across all categories? What people are frequently Buying today, and so on and so forth.

So I have around 10 different widgets for product recommendations. All of them are based on the context, the top sellers, the movers and shakers, what is new, what is trending, provide inspiration to the shopper, even if we don’t know anything about her right now.

Let’s see how we can adopt that. And let’s go to a subcategory like fish as an example. So I click on the fish category, let’s have a look at this fresh organic sea bass. So I go through and interact with this product detail page. And then I interact with another category. For example, electronic tablets, we’ll see how AI can adapt in real time for my signals and those interactions. So I go back to the homepage and see if anything has changed. I still see the Work from Home Essentials. I also see Carrefour private label products. And that’s something they have put as merchandising rules to boost and promote products that give them higher margins. And here you see that those strategies with tablet, fish have been created in real time from the interactions I had just now. So you can see that the homepage adapted in real time, based on what I’ve just done.

So, to conclude, remember that we don’t need to set up any manual or AB tests to personalize. RichRelevance(now Algonomy) leverages a unique optimization AI engine to select the best performing algorithm for each individual from a broad set of machine learning techniques. As a result, you can expect higher engagement as well as lift in both average order value and the number of items per order.