DATA ANALYTICS

Market Basket Analysis: Improve Your Upsells and Cross-Sells

Julian Alvarado

Sr. Content Marketing Manager at Sigma

In the last five years, the cost of acquiring new customers has increased by over 50%. Strategies like cross-selling and upselling can help retailers generate more sales from existing customers, growing revenue without acquisition costs. But how do you know what products to promote together?

Market basket analysis gives you the product affinity insights you need to recommend the items that customers are most likely to buy alongside any given product. But it’s difficult to get basket analysis right. In this first of a three-part series on retail merchandising, we’re taking a look at how you can use basket analysis effectively to boost profitability.

Read How Data Drives a Successful Product Mix Analysis, the second post in this series, to learn how to develop a strong product mix.

What Is Market Basket Analysis?

Market basket analysis is a subset of product affinity analysis that retailers use to discover buying patterns and the relationships between items. The classic fast-food example is a burger and fries. But not all relationships are so obvious. By analyzing the contents of transactions and looking for attributes that appear together, retailers learn what combinations of items occur frequently. With a better understanding of those correlations, they can improve their upsell and cross-sell strategies.

Retailers use three association rules to analyze transaction data:

  • SupportSupport measures how frequently a group of items occurs together as a percentage of all transactions.
  • ConfidenceConfidence is the ratio of the number of transactions that include the combination of items to the number of transactions that contain the single item only.
  • LiftLift says how much better a rule is at predicting the result or how much confidence has increased that Product B will be purchased when Product A is purchased.

The power of basket data analysis is illustrated by the Whole Foods online shopping algorithm. It populates product pages with additional items under the heading “Customers who bought this item also bought.” Amazon knows that if you’re a Prime customer shopping Whole Foods for Beyond Sausage, you’re likely to be interested in Just Egg, a plant-based egg alternative.

Advantages of Product Affinity and Market Basket Analysis

But product recommendations aren’t the only way to use product affinity and market basket analysis. Here are three reasons to add them to your toolkit.

Create more targeted and personalized marketing. When you understand product relationships and purchase sequences, you can identify and track customers who have bought a given product and deliver tailored messages to them. With personalization, you’re also able to create more effective marketing campaigns.

Upsell and cross-sell more effectively. As we mentioned, identifying the connections and correlations between products and product groups helps you improve your recommendation engine rules and algorithms to increase conversions. It also tells you which products to pair together in promotions and when to suggest correlated products at the point of purchase.

Improve merchandising. When you’re able to identify consumer purchase patterns and product affinities, you can create more effective product promotions, shelf layouts, aisle displays, and more.

Why Basket Analysis Is Hard To Do

While product affinity analysis in general, and market basket analysis in particular, are extremely valuable to retailers, this type of data mining isn’t easy to do. There are three reasons why.

Volumes of siloed data are overwhelming

The volume and variety of data that it’s possible to collect can be overwhelming. Purchase history, transaction-level data, and product information are usually spread across multiple, siloed systems, and integration is a challenge.

Traditional spreadsheets can only handle so much

The billions of data points involved in shopping basket analysis require mind-boggling computation power to analyze. Spreadsheets are inadequate due to row limitations, meaning you must use multiple spreadsheets and data extracts. And even with multiple spreadsheets, you can only analyze a subset of the data at any given time. It’s impossible to see a full picture.

BI tools have the power but require code

Even if systems are integrated and compute power is available, complicated BI tools that require coding and technical expertise create bottlenecks. BI and data teams are inundated with requests because they are typically small teams responsible for supporting the entire organization. It can take days or even weeks for BI teams to give retailer merchandisers the insights they need. As a result, retailers are unable to put timely insights into action.

How To Do Market Basket Analysis — Tools and Best Practices

So, how can retailers do shopping basket analysis effectively? Here are the tools and best practices that enable efficient basket analysis:

  • A single source of truth — Using a cloud data platform, like Snowflake’s Data Cloud, to collect PoS data, online transactions, product information, clickstream, loyalty program, and omnichannel customer data gives you a single source of truth. And with modern cloud data warehouses, multiple data formats can be stored and analyzed together, eliminating data silos.
  • Cloud data tools rather than spreadsheets — Traditional spreadsheets are simply insufficient to handle the volume and velocity of data points you’ll want to analyze. However, modern cloud analytics solutions (like Sigma) allow you to analyze billions of live data points down to the individual SKU or transaction level in seconds.
  • Direct, democratic real-time access — Your analytics platform should allow business teams to perform independent analyses without having to extract data or knowing how to code or having to wait on BI teams. Bonus points if it is in an environment they are already familiar with, like spreadsheets.



Examples of Companies Doing  Basket Analysis Well

Retailers are increasingly realizing the power of basket data analysis. Let’s look at two examples of companies that are using it to power their conversions.

Quantzig’s Basket Analysis for Food Retailers — A case study by Quantzig, a global data analytics and advisory firm, helped a German food retailer increase quarterly revenue by 50% and reduce marketing cost by 15% using basket analysis.

Amazon’s Basket Analysis Data — We mentioned Amazon’s algorithm in the Whole Foods example above, but Amazon uses market basket analysis across all its offerings. Additionally, the company makes its market basket analysis data available to its marketplace partners.

Market Basket Analysis Is a Competitive Advantage for Retailers

Retailers that use market basket analysis can dramatically improve their revenues as a result of better upsells and cross-sells. And when people in various roles in the organization can mine for insights, not just BI and data professionals, you can implement changes that will improve marketing results and increase conversions.

For more on market basket analysis and other strategies retailers can use to improve profitability, read

How Data Analytics Empowers Retail Merchandisers to Thrive Through Disruption.

Ready to visualize your data for actionable insights?