Wired Innovation Insights: Realizing Big Data Benefits: The Intersection of Science and Customer Segmentation
June 7, 2013-
By Neil Biehn
Everyone knows Big Data is here to stay but, per my last post, the real key to make it work for your business is to determine what value it can create that translates into a competitive advantage – that is, is it viable? One great example of Big Data working to achieve business goals can be found in one of the most common corporate practices – differentiating between customers in some meaningful way. Almost all companies segment their customers, realizing the maximum value doing so can have for both parties involved. For nearly every industry, customer segmentation provides a significant opportunity to increase revenue through a better understanding of customers – and how to best approach sales, offer and pricing strategies for them. How can Big Data help with that task? The answer lies in science-based segmentation.
Science-Based Segmentation
Science-based segmentation utilizes statistical data analysis and methods to group customers into similar buckets. The object is to create specific optimized sales and pricing strategies, given particular buying behaviors of each segment. A segmentation study begins with data and deep business understanding. Executives, sales managers, marketers and pricing experts can provide insights into what differentiates the company from a benefit and value perspective. Experienced professionals bring their hypotheses of the potentially important customer, product and offer attributes that drive customer behavior. Examples include geography, product mix, volume purchased, industry, lifecycle, purchase frequency, size and many others. IT departments struggle to gather all the potential external and unstructured data sources – and then there is the Big Data repository of all transactions.
Beyond the Data
Big Data alone cannot overcome the challenges in customer segmentation. Methods utilizing only data mining for segmentation typically miss key factors, or combinations of factors, that may not explicitly exist in the initial data. Business expertise is a critical component to defining a great segmentation. Consider the following example – while there is no field in the data that says “wallet share”, it isn’t very difficult, using data, to determine the percentage of revenue a customer spends on a specific product. If thinking of terminating a service or product offering, an organization probably wants to know how many customers have it as a large part of their total spend.
However, business expertise alone will miss the mark – it doesn’t matter how many meetings you have. The business team can only make hypotheses. Data science must be used to then test and conclude which pieces of the data actually provide real predictive value into how different customers behave differently.
Customer segmentation guided by Big Data analytics is critical when it comes to driving sales. Not only does the process identify where each customer fits, it also makes the sales team smarter and more strategic when approaching them. With science-based segmentation, organizations can more easily identify anomalous buying behavior and make intelligent product and offer recommendations that are statistically more likely to be purchased. If two customers are alike but not buying the same products, the data analysis can advise which opportunities the sales team might be missing. This is the type of Big Data viability that moves the needle in the real world.
Real World Example
To illustrate the idea of science-based segmentation as it relates to Big Data and sales, consider one major express shipping company who was faced with the constant challenge of creating contract offers for varying customers. By implementing science-based segmentation, they could leverage the massive amounts of data and the varieties of data they held about customers and determine potential attributes based on account geography, competitiveness, total volume, product centricity, and other groupings. Once they did this, they sorted customers into groups that respond similarly to contract offers. Moving away from the one-size-fits all contract policy allowed the sales team to create tailored offers and pricing strategies that would have the best chance of driving sales growth. As a result, the company received gains on a magnitude of tens of millions of dollars after six months.
Sales organizations have several different technologies at their fingertips today – and customer segmentation tools are not often discussed. Many companies, therefore, are missing a huge opportunity. The next time you hear of someone talking about why you “need to stop allowing deep discounts for low margin customers” or asks, “why aren’t we offering customer-centric solutions?,” consider bringing Big Data into the picture to make customer segmentation work for you.
Neil Biehn is vice president and leader of the science and research group at PROS.