When it comes to e-shops, the focus is on products and increasing sales. Consequently, personalization in e-shops is directed towards products and creating personalized offers and displays of products for each user.
Recommendation engines are tools that recommend products to visitors based on their preferences and interests.
How do they traditionally work? They first retrieve data in real-time about the browsing history of visitors and their behavior throughout the session. Then, by using rule-based algorithms, they associate these preferences with certain products. It results in a set of products that matche exactly the preferences of users.
What’s the ultimate objective of a recommendation engine? – Making accurate predictions on what the visitor might like. They have to understand perfectly how taste, trends and online behavior work together.
If we can make a loose comparison, they are like the intuitive version of forms. When they encounter a form, visitors search for something by choosing variables that describe exactly what they want. So there is some input to begin with.
Here are some examples that will help you visualize how a recommendation works:
Asos.com suggests products you can buy to complete the look:
Another online shopping platform, topshop.com, displays a “Why not try..” section, that recommends similar products with the current product you are looking at:
Currys.co.uk, a British electrical retailer, shows you what other people that are looking and the same product as you are, have search on their website:
Recommendation engines work intuitively: they read the visitor’s preferences and predict what would suit him best.
What’s the first condition for a recommendation engine to generate the best results? – Having a very precise and powerful data gathering and interpretation system. The more data they can gather and interpret about the visitor, the better predictions they can make.