How to perform an e-commerce basket analysis and what to do with the data?

What is e-commerce market basket analysis?
E-commerce market basket analysis is a technique used to identify relationships between products people buy from your online store. To elaborate further, when performing an e-commerce market basket analysis you are trying to anticipate (based on data collected that after a customer selected, and added a product to cart or purchased a product) which product do they likely need next or should recommended in addition to first product added?

A real-world example will be when you buy an e-cigarette online then the website recommends you buy a selected range of flavoured vape oils for that e-cigarette. Also like in the example below, after clicking on a Charcoal Kettle Grill the site recommends Hardwood Charcoal because by monitoring customers' buying patterns, it has discovered that whenever one buys the grill, they are most likely to buy it with the Hardwood Charcoals.

Example 1

How is this done?

Unlike humans, computer systems don’t know that you bought a Grill so you need Charcoal. It will not guess that you moved into a new place, so by purchasing a Kettle then you might also need an Iron. To do this, we use e-commerce market basket analysis and makes use of recommender systems or algorithms.

There are three types of recommendation systems that you can utilise when doing e-commerce basket analysis. You can implement one or use all of them together to fine-tune your recommendations engine:

Knowledge-based Recommendation system – This type of system investigates customer product purchase behaviour and analyses the data, looking for products that are purchased together most of the time. These products will be grouped and used in recommendations on one (or more) of the products selected.

Collaborative filtering model – This system creates recommendations based on data collected on customer’s preferences and product recommendations to a similar type of customers.

Content-based Recommendations – This system uses historical sales data of customers to create predictions for recommendations.

You can go deeper by looking at the relationship between the products themselves. There are many ways to analyse the relationship between product these are called association rules. You look at the connections if they are useful, trivial and inexplicable.

Why is e-commerce basket analysis important for your business?

E-commerce basket analysis can help increase sales to your business by giving recommendations of appropriate products to customers shopping on your website. E-commerce basket analysis can be that edge that your business needs against your competitors. It will boost the credibility of your online store and reduces the time customers spend looking for products on your site. E-commerce basket analysis will give you an insight into which products compliment each other and makes you focus on those, as some of the products rarely get bought without the pairing product. After identifying your product groups you can create new product bundles that can improve your average order values.

If you require assistance with e-commerce data analysis, please contact Algorithm Agency.

About the author

Nigel is a Data Analyst with a web and app development background. He is quickly developing in SQL, Python, Javascript and front-end coding. Nigel loves transforming raw data into useful information that informs business decision-making.