Leveraging Share of Wallet Analytics to Deliver Tangible Impact for Your Brand

Harnessing big data and machine learning to become more relevant in your customer touchpoints by identifying Share of Wallet opportunities in your customer base

Media coverage of “big data” and “machine learning” within the marketing space has exploded thanks to their promises to revolutionize the effectiveness of brands through the power of analytics and customer insight. Business analytics are becoming “the air companies breathe and the oceans in which they swim,” according to Deloitte Analytics. And interest and attention from the CMO and C-Suite is high. There’s an emerging belief that analytics can provide a growth engine, but often analytics insight either isn’t being generated at all or, as The Harvard Business Review recently stated: it seems “a core issue with generating insights is that they don’t get to the people who can use the.”

In fact, this challenge may be worse than many perceive it to be. A recent McKinsey analysis of an ANA survey revealed, for example, that “only 10% of respondents believed they were very effective at feeding insights into customer behaviors back to the organization to improve performance.”

One space where customer data and predictive analytics have long delivered value is in direct marketing. Response, spend and uplift models have provided a solid understanding of which customer to contact, and with which message. More recently, marketers in highly promotional industries like retail have developed discount sensitivity models to manage the level of incentive required to produce a desired customer response. Furthermore, the communication channels available to marketers today have increased the complexity of this process by requiring an accounting of social, mobile, web, local marketing, etc. to understand all available touchpoints a given brand has with its customers.

Analytics teams are faced with increased complexity, but also have a greatly expanded opportunity to impact customer value. New database and data mining tools have improved the speed at which new customer data coming into organizations can be mined for insight and utilized for developing improved relevance and targeting in marketing. This is accomplished by applying machine learning-based modeling techniques to large numbers of target attributes, supported by very large customer level data sets to provide a 360-degree profile of each individual and household.

One highly actionable analytics product emerging from this work at Olson 1to1 is Customer Share of Wallet. The core idea is to create predictive wallet size models for each product level of interest for each individual customer. For retailers, this is typically a higher-level product category, but the concept can be pushed down the hierarchy all the way to the individual sku.

Analytics Chart
Convert deep customer profiles into estimates of product category wallet size through the use of machine learning algorithms

By targeting customers with low actual spend, but large predicted wallets in a given category, it’s possible to improve sales lift by being more relevant to the customer. If a customer has purchased regularly within a specific category, let’s say snacks, the result of most algorithms will be to continue to push snacks or products associated with snacks. By looking at share of wallet, a marketer can determine if there’s any remaining sales potential in snacks and if not, look for low current share of wallet categories to focus on. This process serves multiple purposes:

  1. Increases the number of product categories in which a customer is active—a key predictor of future value
  2. Withholds marketing dollars from product categories the company already dominates
  3. Increases the relevance of the message to the customer
  4. Reduces the repetitive nature of many product targeting processes
  5. Opens up a rich test and learn opportunity for continuous improvement

Share of wallet estimates can also be used as a segmentation scheme, tracked over time, to understand trends in the performance of product categories in creating and retaining high-value customers. Example deep-dives which could be performed include: looking for patterns in high share of wallet customers who were not retained; understanding the appropriate discounting strategy an individual customer should receive; improving the brand’s performance based on current and potential share of wallet; and understanding the lifecycle of the share of wallet metric and where there are natural intervention points to exploit. The resulting insight will be extremely valuable for the merchant capability, both in managing pockets of opportunity, but also in attracting vendor funding to drive specific initiatives.

A final benefit of producing a share of wallet model is it can be directly linked to the brand’s overall market share. This creates an immediate connection to metrics that senior leaders in all organizations are interested in tracking and improving; driving improved utilization and business impact.