Recently we’ve told you about our item-to-item recommendation system for websites and email. We’ve been hard at work improving this product ever since and now proudly announce the addition of some cool new features.8p>
- Related products
- Bought together with
Now you can combine these two basic algorithms with different settings according to your needs and even set up A/B tests.
- Create product recommendations according to your own business rule
- Fine-tune your settings at any moment to test or use different configurations of each algorithm
- Have your recommendations updated automatically.
The where and the how
While developing this feature, we aimed to create a simple yet flexible tool. For simplicity, all basic settings are pre-configured by default – just choose the recommendation type and the number of products:
And add the desired recommendations on the website:
If you want the recommended products to be within a particular price range, simply move the price slider to the needed position:
And add the recommendations to the body of your email:
Product types and categories
You can configure the recommender to work for a specific category or type of products:
You can further limit the product selection by using custom fields such as brand, material, color, etc. The big advantage here is that you can set the custom fields to work with just about any product characteristic and then use them to configure the recommender via the same slider:
Need to base your recommendations on an exact match? We’ve created the Fixed field option precisely for this occasion! Use this field to make sure all recommended products have the exact same value of one of their characteristics as the initial product.
For example, let’s say you have an online shop selling tableware and you wish to have all recommendations to belong to the same product series as the initial item. Select «Series» in the Fixed field:
And voila! All recommended products have the same series as the initial one:
This feature can be used to easily set up the «Related products» and «Bought together with» types of recommendations.
- Visual reports of configured recommendation algorithms;
- Personalized recommendations based on the user’s previous actions such as orders, page views, and other product-related data.