Recommendations matching customer attributes
The recommendations are intended to be the best match of product listings to the client's expectations built upon their previous interactions with the products. In their basic form, they look for the best matches based on the entire product feed.
Unlimited choice is not always the best solution. Not only for a client whose, for example, size is not analyzed when building a recommendation but also for marketers who want to implement more or less complex marketing scenarios based on customer segmentation.
Restricting the listing of recommended products based on customer and product attributes is a solution to this problem. Here are some ready-made use cases to get you started with this feature.
Product recommendations per customer clothes size
The purpose of the recommendation is to present the customer with the products best suited to their behavioral profile built during each visit to your online shop. One way to do it is to recommend items in the customer’s size. In this Use Case, you will learn how to create such recommendations.
The client's favorite colour as an element influencing the recommendations
Colour is one of the features that significantly affect the purchase decisions in the fashion industry. For this reason, it can be very valuable to recommend products based on your customers' favorite colour. For more details on how to create such recommendations, check out this Use Case.
Recommendations depending on the customer segment
Combine the power of the prediction and recommendation engine to improve the effectiveness of recommendations. Learn how to create recommendations for customers with high lifetime value.
Recommendations for VIP clients
Learn how to create effective product recommendations for VIP customers with appropriate product price ranges whose margins allow the customer to "safely" consume the discount.
Inspirations
Information from surveys and forms
A valuable customer behavioral profile, apart from a large amount of information from interactions, i.e., events, also includes additional, but equally, important information gathered in the client's attributes. Attributes store information about your favorite brand, and size, but also many other parameters that can be used for an even more precise selection of product recommendations. Let's consider how to influence the presented recommendations by using information from surveys and forms.
Age-limited product recommendations.
Age or date of birth is one of the attributes that we use to verify access to certain products and services. This information can very well be used to enhance the presentation of selected product groups. After all, console games are bought mainly by older players...
Attributes gathered in a survey.
Short surveys can be a source of a lot of valuable information written as customer attributes. A favorite sport, leisure activity, or even a movie or book can be used to filter recommendations for a specific client. Every marketer knows what beer a fan of films about an English agent will choose ??
Based on the most convenient delivery as an attribute.
The preferred method of delivery is also an interesting attribute (information about the customer). The customer will not see in the recommendations those products, which due to their dimensions or other specific delivery conditions, will significantly increase the cost of the order and discourage them from making these and subsequent purchases.
Attributes based on events
Events store very detailed information about every interaction of the customer on the Internet, but also in the brick-and-mortar world. Aggregating this information as attributes allows this detailed knowledge to be used in advanced personalization scenarios. These are some suggestions for using attributes based on events.
Exclusion of returned or exchanged brands in product recommendations
Customers often equate the advantages and disadvantages of products with specific brands. Information from the events describing the complaints can be saved in the form of an attribute that defines the producer. Then, based on this attribute, we limit personalized recommendations to the products of those manufacturers who do not have negative associations for the customer.
The customer was not able to consume a product
Returning customers are of great value to any business. For each visit to be equally effective in terms of sales, during the customer's next visit to the store, we can remove from the recommendation those products that, according to our calculations, should not be the customer's center of interest at a given moment. So those that he has not used up yet.
Churn customers should like brands that they previously bought
A churn customer is one who has not made a purchase for a specific period. One of the nurture strategies may be communication-based on products (manufacturers) that they previously purchased. In this case, our recommendations will be based on linking the manufacturer's attribute on the product side and the favorite or most frequently purchased brand on the client attributes side.
Dominik Królikowski, Business Value Services Director
For even more interesting Use Cases describing recommendations visit our Use Case Library.
For our Partners, we have prepared a series of workshops and inspirational meetings that expand knowledge about this functionality. If you are not a Synerise Partner yet, consider registering.