Let’s say you’re an FMCG brand and your customer is searching for ingredients on your marketplace listing to make Poha. While they’re adding flattened rice to the cart, they realize they may also need some peanuts, oil, and spices, maybe even some spicy sev. What if you could automatically predict this behavior and suggest additional products as your customer is viewing your online marketplace listing? It would certainly mean more value for your customer and more sales for you.

Say hello to market basket analysis (MBA)

With analytics now playing a bigger role in digital marketing, upselling and cross-selling have become ubiquitous in online marketplace management, and market basket analysis is a must-have tool to derive valuable customer insights.

  • What exactly is it?

To put it technically, market basket analysis is a data-mining technique that can help you understand and predict purchasing patterns in a retail space by showing you how customers connect different items on a marketplace listing.

  • How does it work?

At the heart of market basket analysis is association rule learning, a machine learning technique that creates relationships among variables within a dataset. If your customers are looking for notebooks, they’re also likely to purchase pens. If they’re looking for a smartphone, they’re probably also looking to buy screen guards, earphones, or power banks. If they’re buying a sandwich, they may want a beverage to go with it. Market basket analysis discovers these connections and establishes relationships between products using statistical affinity calculations.

  • How does it help?

Analyzing your customers’ behavior and purchase patterns allows you to make improved product recommendations and promotional campaigns, as well as fine tune product bundles and placements on your online marketplace listing. Ultimately, when done right, market basket analysis can help you improve customer engagement, customer experience, sales, and ROI.

The A to Z of MBA: Basic terminologies to know

Before you jump into finding the right marketplace management services for you, it is important to equip yourself with the right knowledge, or at least the right terminology. Here are the most important ones:

  • Itemset: A collection of products your customer purchases
  • Antecedent: The first product that your customer purchases or places in the checkout cart
  • Consequent: The product purchased as a result of the antecedent item
  • Support: The probability of the antecedent event occurring, or the general popularity of the product
  • Confidence: The measure of the probability that your customer will purchase a ‘consequent’ item after the ‘antecedent’ item
  • Lift: The efficiency of the rule in finding consequences, as compared to a random selection of transactions 

The formula for market basket analysis The most popular algorithm used in market basket analysis is Apriori. It conducts frequent itemset mining as well as association rule learning, which we mentioned above, used with relational databases. Apriori leverages a bottom-up approach by extending frequent items one at a time and testing them against available datasets (explained in the formula below).

Revisiting the example of the delicious Poha, the Apriori algorithm uses this rule of association formula and tells you to promote spices and other items when customers search for ingredients – thereby increasing your average order value. 

Three types of market basket analysis techniques

There are various other algorithms used for market basket analysis, the primary difference being how differently datasets are utilized.

  • Descriptive MBA
    The most commonly used approach, this technique derives customer insights based only on past data. Descriptive market basket analysis uses statistical modeling (unsupervised learning) to rate associations. It does not make any predictions.  
  • Predictive MBA

This technique uses sequential data to identify cross-selling opportunities. Predictive market basket analysis uses supervised learning models such as regression and classification, mimicking the market to analyze cause and effect.  

  • Differential MBA

Ideal for competitor analysis, differential market basket analysis determines interesting customer behavior patterns by comparing purchase histories across stores, seasons, time periods, days of the week, etc. It can help you identify intangible customer preferences – such as why customers choose a specific marketplace, even when the same product is available at the same price on both platforms. The answer could be delivery time, customer experience, warehouse volumes, or another factor that was previously unknown.

Integrate market basket analysis easily with TenSight

Not sure how to begin integrating market basket analysis into your business?

TenSight – Tenovia’s central ecommerce analytics dashboard – allows you to make data-driven business decisions that are better, quicker, and more accurate. The dashboard comprises centralized data from all major ecommerce platforms and offers actionable insights.

With our expertise, we’ve helped a global denim brand achieve a 3x boost in conversion, 3x clickthrough rate, and 15% growth in brand SOV. We’ve helped a national footwear brand achieve a 40% boost in profitability and over 1 Cr in savings. TenSight has also helped increase regional utilization in ecommerce warehouses, and reduce delivery time and cost per shipment. To know more about TenSight and our marketplace management services, talk to us today.

Recommended Posts

No comment yet, add your voice below!


Add a Comment

Your email address will not be published.