Button: Member Login
 

Improving ROI through Customer Transaction Analysis

Jim Manzi, Anthony Bruce, Scott Setrakian
Applied Predictive Technologies
July 2009
www.predictivetechnologies.com

This is a 10 page report.  Read below or download the PDF.

Contents 

Abstract
Case Study: Direct Marketing Customization
Case Study: Identifying Cross Sell Opportunities
About Market Basket Analyzer

Abstract

Analytical Customer Relationship Management is a fast growing marketing tactic. The promise of Analytical CRM is compelling: spend marketing dollars more effectively by sending the right offers to the right customers at the right point in their purchasing lifecycles. Increase marketing ROI by improving customer adoption, and cross-selling / up-selling effectiveness.

But as marketing organizations have begun to explore this approach, they have run into significant obstacles in sifting through gigabytes of transaction data, and aggregating across millions of transactions, to accurately determine which customer behavior patterns can be influenced by marketing tactics and which cannot. Moreover, they are facing internal debate around the proper balance between precise, customer-targeted programs and more traditional programs such as mass media, visual merchandising, and in-store promotions.

The APT Market Basket Analyzer is designed to help companies overcome these challenges and reach actionable insights. Leading retailers are using Market Basket Analyzer to bridge the gap between Analytical CRM’s highly promising concept and the challenges of implementation.

Case Study

Direct Marketing Customization

A leading retailer was in the midst of re-evaluating its direct marketing activities.

Historically, it had sent out over 10 million pieces per year, employing a standard set of catalogs and mailers. The company determined who received a catalog or a mailer based on its estimate of the lifetime value of each customer.

In recent periods, the company had begun to slightly tailor the mailings by customer segment, using a segmentation defined by product categories each customer had purchased in the past. Within the pieces, the advertised product assortment was selected based upon commonly purchased items, plus new product introductions and vendor-funded promotions.

This retailer worried that one-size-fits-all content did not create an adequate purchase stimulus for many mailed customers. They believed that tailoring the pieces by product category would make the mailings more compelling for recipients and drive higher response.

While this theory looked good on paper, the retailer had never been able to prove that it truly generated incremental sales. Furthermore, selecting the product assortment had always been a difficult task. The company had a process to profile individual items, but the analysis was time-consuming and shed light only on those specific products, not related items.

APT Market Basket Analyzer was applied to provide a fully integrated view of customers and the items they included in their baskets. In Market Basket Analyzer, the retailer analyzed the cross-category purchasing behaviors of customers, both within baskets and over time. This company saw that most customers tended to be loyal buyers within their initial product category. Relatively few customers tended to shop across product categories. The targeted mailer content strategy was essentially correct.

An additional finding from the analysis demonstrated a clear upward trend in the price point of customer purchases over time. First-time customers bought almost exclusively entry-level items. With each subsequent visit, they tended towards higher price-point products. The company concluded that, while the mix of advertised products generally matched the mix of what mailed customers had bought in the past, it did not sufficiently cover items that customers wanted to buy next. Significant gains could be realized by advertising more higher-priced (and higher-margin!) items. (Fig. 1)

Fig. 1: Cross-Vertical Purchasing Behaviors (Illustrative)
 

Using insights from Market Basket Analyzer, this retailer improved its direct marketing effectiveness by:

  • Further differentiating mailing materials by customer segments, defined by past categories of purchase
  • Shifting the mix of advertised products to higher-priced items

Case Study

Identifying Cross Sell Opportunities

Another APT client wanted to begin targeted outreach to qualified customers in order to encourage follow-on purchases.

However, this retailer was uncertain of where to begin. In order to pilot this initiative, they would have to understand:

  • Which high-value product(s) to promote
  • Which customers would be most likely to respond

To answer the first question, the client sought to develop a profile of the total value attributable to a product – including not only the value of the product itself, but the stream of subsequent purchases that would be unlocked after the initial purchase. Using Market Basket Analyzer, the retailer developed a “repeat customer” profile for each of several thousand products. Customers were categorized according to their basket composition during transactions in one timeframe. Then, each customer was profiled for subsequent behavior, including what items they bought, the overall frequency of those follow-on purchases, and the time between purchases.

Through this analysis, the company was able to identify a set of “driver” products that led to outsized follow-on customer purchases. The retailer learned that customers who initially bought these driver products were likely to buy significantly more after that initial purchase.

Customers were also profiled to see what they had purchased prior to making the “driver” category purchase. With this information, the company was able to develop a qualified list of customers who had not yet purchased the “driver” product, to be targeted for cross-sell efforts. (Fig. 2)

Fig. 2: Identifying Driver Products

Using insights from Market Basket Analyzer, this retailer was able to design a pilot for targeted outreach to potential qualified customers of these high-value “driver” products.

The retailer also instituted an ongoing Test & Learn process to measure the success of this pilot program and improve the program design over time, expected to be worth millions of dollars per year.

About Market Basket Analyzer

In these case studies, retailers used APT Market Basket Analyzer to inform mailer content and targeting decisions. More broadly, Market Basket Analyzer enables retailers to efficiently generate actionable product attachment insights to guide many high-stakes in-store and direct marketing decisions. Leading retailers are using Market Basket Analyzer to answer questions including:


1. What products are sold together?

Understanding how products attach to each other can inform merchandising, bundling, pricing, promotion, associate training, and direct marketing decisions. This is the core analytical output of any market basket analysis. But too often, the findings are so common-sense that they fail to create new insights or drive incremental value.

APT Market Basket Analyzer augments this core analysis with the ability to segment by attachment rates, at any level of the product hierarchy. So, in addition to learning which razor blades are sold with razors, analysts understand what types of stores or what segments of customers tend to attach products more or less frequently.

This allows the retailer to be more targeted in its response, and significantly increase the effectiveness of the marketing spend

2. When a customer buys a certain item, what will she buy next?

Commonly, attachment analysis treats the transaction as the fundamental unit of analysis.

However, a customer lifecycle viewpoint can reveal how a purchase of an item may unlock a future stream of purchases. Think camera-to-printer-to-photographic paper, pet fish-to-fish tank-to-fish food, etc. – common examples of dynamic customer behavior stimulated by an original purchase. Viewing purchases only at the transaction level misses these relationships. The immense size of a SKU-level transaction log makes transaction-level analysis daunting enough - sifting through customer patterns only magnifies the scale of the analytical challenge.

With APT Market Basket Analyzer, retailers can seamlessly analyze individual customers’ transaction patterns over time.

3. Which sales, marketing, and operational programs actually change customers’ attachment behavior?

Market basket analysis can highlight key attachments and repeat customer purchase behaviors. But identifying these behaviors is not enough to make money. Retailers must also determine how to influence attachments.

Linking APT Market Basket Analyzer with APT’s Test & Learn for Sites and Test & Learn for Customers toolsets, retailers can test programs designed to influence purchase behaviors. Market level programs (e.g., insert ads), store level programs (e.g., improved adjacencies) and customer level programs (e.g., targeted coupons) can be tested, to understand which levers best influence behavior. By measuring changes in attachment patterns after the introduction of a program, relative to a matched set of control stores or customers, retailers can easily understand which programs truly drive incremental sales and profits.

About Applied Predictive Technologies
APT is the industry leader in helping large-scale consumer-focused companies institutionalize a world-class Test & Learn capability. Through the combination of APT’s proprietary software and capability-building consulting support, APT has helped some of the world’s largest and most successful companies achieve significant bottom-line improvement.