/ Digital Transformation

Tailored product recommendations for your online store customers

In our last article, we wrote about how data science can be used to boost sales for online stores. This article is about a tool that integrates with your online store and helps you send tailored emails via an email marketing tool like Mailchimp.


Data science insights are nice to have, but useless unless you act on them. One of the ways to act on these insights is to send tailored recommendations to your customers via email marketing.

Other strategies, which we plan to cover in future articles include:

  • Having different marketing strategies for different groups of customers (e.g. your VIP, core, high-rollers etc.)
  • Rule-based email marketing such as emails sent when a product is likely to need replenishment or a win-back promotion when a customer is at-risk of loss
  • Self-learning A/B/...Z tests that optimise your emails for timing, subject and content to maximise open and click rates

What are tailored recommendations?

The idea behind tailored recommendations is to send content that are more relevant to your customers based on what you know about them. Tailored recommendations will lead to statistically higher click-through rates on email marketing.

Here's an example of recommendations in action in an email from booking.com:

How does our recommendation algorithm work?

Recommendation algorithms are a combination of past customer behaviour and what customers are currently browsing. Our recommendation engine looks at historical purchasing patterns and recommends certain products to a customer. People love variety so we recommend different products to your customers and learn what works over time.

In our last article, we discussed customer segmentation insights that are based on customers' purchasing patterns (diagram below). We analyse your store's historical sales and groups customers according to the mix of products they have bought.

Our recommendation engine then uses these insights to suggest new or related products that customers might be interested to buy based on what others in their segment have bought.

For instance, we know that customers in Cluster 1 tend to buy Product 300105, 300106 and 300211. Then, a natural recommendation for a customer in that segment who buys only Product 300105 and 300106 is Product 300211.

So how do I implement it?

Good news is you can implement this with your favourite email marketing tools like Mailchimp.

We create email lists in Mailchimp with recommended products as merge fields. You can use these merge fields in one of your email templates to send a campaign with tailored content for each user with your branding styles.

That's it!