Smart Pricing Part 4: Application of Unsupervised ML in B2B Companies

August 23, 2018 Amir Meimand

Price management is no small task for B2B pricing owners. While machine learning can set smarter pricing strategies, understanding how it works in the context of B2B and knowing how to deploy highly-relevant price guidance, and ultimately, create a more strategic pricing approach isn’t as straightforward. In this four-part series, Zilliant Director of R&D, Pricing Science, Amir Meimand will share a summary of how predictive analytics, prescriptive analytics and machine learning contribute to a smart pricing strategy for B2B companies.

Amir will be speaking at the 6th Annual Global Big Data Conference on August 30; he will present on the topic “An Efficient ML-Based Algorithm for Mining Substitute Products for eCommerce.”

In my last post, we discussed the differences between supervised and unsupervised learning algorithms and how to determine their effectiveness in a business setting. In this post, I’ll share practical applications of these algorithms in B2B.

As unsupervised algorithms are seeking to discover the hidden pattern from data, they can provide some actionable insights that enable a business to gain a competitive advantage. 

Customer segmentation: For B2B companies, customers differ across multiple dimensions, such as demographics, competitiveness of location, annual spend, dominant industry, etc. Traditional marketing methods focus mostly on one or two dimensions to group similar customers together.

Map of southwestern United States showing clusters of customers based on one dimension in Texas and California

This approach fails to capture the effect of all metrics, as a result, very different customers are grouped together. For example, it is not accurate to say that all customers in California behave similarly just because they live in the same state. Other factors, such as company size and dominant industry can strongly influence the extent of similarity. This is important because the price optimization model sets the same pricing strategy for similar customers. For example,  a revenue-aggressive strategy may be set in California because it is a competitive region, whereas a profit-aggressive strategy may be set in Florida because it is less competitive. However, there might be some large customers in Florida who will not appreciate the profit-aggressive strategy, as their margins are very sensitive to price.

The clustering algorithm captures the effects of all important factors and creates more granular and effective customer groupings.

Map of southwestern United States showing clusters of customers based on multiple dimensions in Texas and California.

Complementary and substitute products: In the price optimization process, it is very important to understand the relationships between products based on customer behavior, whether those customers are retailers or suppliers. In general, if two products are related they can be either substitutes or complements. Extracting both complementary and substitute products provides valuable knowledge for market prediction.

Mining complementary products reveals which products are frequently purchased together. Such information can be used to generate price recommendations, not only for individual products, but also for baskets of products that often are purchased together. Complementary product identification is a core design element of recommendation systems used in eCommerce.  Zilliant’s customers benefit from Cart IQ to increase average order value by delivering real-time complementary product recommendations at the time of order, based on the items in a customer’s basket.

Identifying substitute products enables a constraint to be placed on the optimization model to ensure products with similar functionality are offered at similar prices. The process of mining substitute products is based on extracting negative association rules; this process can be very expensive from a computational perspective and typically generates a lot of redundant rules and needs. Zilliant’s AI platform takes an innovative approach to discover substitute products by deriving product similarities based on corresponding association rules. This method is computationally efficient and effective.

Read the related posts in Amir’s Smart Pricing blog series:

Smart Pricing: Machine Learning Applications in B2B Pricing

Smart Pricing Part 2: Problems Addressed by ML

Smart Pricing Part 3: Unsupervised Learning Algorithms

About the Author

Amir Meimand

Amir Meimand is Zilliant Director of R&D, pricing science, where he designs and develops pricing solutions for customers and performs research in which he applies new methods to improve the current solutions as well as develop new tools. Prior to joining Zilliant, Amir helped design and develop a promotion planning and pricing platform for B2C retailers. Amir holds a dual Ph.D degree in industrial engineering and operations research from Pennsylvania State University. In his doctoral work, he applied operations research concepts to dynamic pricing and revenue management.

Follow on Linkedin
Previous Article
Food Service Distribution: What if Your Products Could Find Your Customers?
Food Service Distribution: What if Your Products Could Find Your Customers?

It’s a given that food service customers want the consumer-like ease of buying online; this blog post explo...

Next Article
Smart Pricing Part 3: Unsupervised Learning Algorithms
Smart Pricing Part 3: Unsupervised Learning Algorithms

Zilliant Director of R&D, Pricing Science, Amir Meimand gives a primer on supervised and unsupervised learn...