In working with data from Fortune 100s to dot-com start-ups nationwide and abroad, two things are clear. First, though the old toolset of data analysis techniques remains viable, innovation must be constant in a world where change happens at a record-breaking pace. Second, companies, wittingly and unwittingly, are awash in data that they do not make good strategic use of.
Take the ever-capable RFM (recency, frequency, monetary), the tried-and-true analysis that all direct marketers begin with. It’s a great tool that businesses often forget to employ, but it still focuses on customer behaviors in just three strict dimensions, which often means missing some strategic marketing insight.
As in most cases of innovation, the innovative component is not something new, but a new way of combining something old that no one has thought of before. By combining traditional and slightly modified analyses such as RFA (average order instead of monetary) and C&RT (classification and regression tree), a multidimensional scoring process called ABT (advocate, buyer, tryer) was developed. ABT is not a new analysis, but a new way to pull together the results of multiple analyses to make them easily actionable and understandable to an enterprise.
Consider a dot-com, business-to-business enterprise. RFM can quickly tell who buys the most, who buys often and who bought recently. But what about who buys from multiple product categories, which client has one versus several contacts that have purchased and which customers have untapped subsidiaries that have never bought? Those are typical dimensions that are pertinent in identifying a “best” customer in a BTB market.
In a recent case study, a dot-com asked an all-too-common question: Where should we deploy sales and marketing to get a high return on investment? An RFA analysis was run, and it quickly showed a fundamental flaw in the way the company viewed its best customers and went to market.
RFA is a twist on traditional RFM analysis that can reveal crucial customer insight missed in a straight RFM. Consider a jewelry customer who purchased 11 times over his lifetime. The accumulated monetary is $6,000, which ranks him high on an RFM analysis. Using this rank, a company might send the customer a high-value offer on a high-dollar item, such as $200 off a $1,500 purchase. But this customer bought one item at $4,000 and 10 items at $200. The average order is $545, which means that the offer above would miss the mark.
Additionally, in most businesses, frequent low-margin purchases cost more to fulfill than less-frequent, high-margin purchases. This typically means that low average orders correlate to low average margin, something companies often are not cognizant of.
Using RFA for this dot-com allowed for more relevant offers and showed where to invest higher-dollar sales talent (larger average orders) and where to strive to move low-margin customers to higher-margin products via less-costly telemarketing and direct marketing efforts.
But stopping here would have left a lot of insight on the table about other behavior traits that are important, but usually not revealed in an RFM or RFA analysis. For example, are customers that purchased multiple product lines more profitable, and less likely to defect? How about customers that have multiple contacts buying versus a single point of contact? And there are many other dimensions to consider, such as functional title and its relationship to profitability. Are you talking to the vice president of sales or the mail clerk with $50 invoice authority?
In this case study, these and other dimensions were isolated. Three core behavior traits were found crucial to retention and profitability. Basically, any customer that had transacted four times, bought across four product lines and had at least two contacts buying had a defection rate of zero. These customers also had a very high average order.
The challenge was to put these discoveries in terms that marketing and sales could easily understand. Since everyone intuitively understood the difference among an advocate, a buyer and a tryer, the term “ABT” was born. Using all of the knowledge about what behaviors described a best customer, an ABT scoring methodology was developed that was unique to that business.
What makes ABT especially significant is that it can combine insights from as many or as few analyses as deemed relevant, making the scoring complex but the output simple. For example, a customer with high recency, frequency and average order who bought two product lines and had two contacts buying might score an 8. Overall scores of 8-10 were advocates, 5-7 were buyers and scores below 5 were tryers.
The sales force then targeted the advocates, many of which were companies not targeted previously. The buyers were hit with a mix of direct and indirect sales efforts to keep their cost of retention profitable, and about 30 percent of the tryers that had a significant chance of becoming buyers were hit with an indirect sales message.
It’s still early for a full accounting of results, but this dot-com has survived and received continued investment while many of its peers have not. Since then, ABT has been used successfully at other large BTB firms across a number of industries.
No matter how complex the process of data analysis to insight is becoming, the results still have to be simple to understand across an enterprise and quickly executable.
Today, when everyone is rebuffing the black box approach and customer relationship management’s promise has been found lacking, innovating with analyses like ABT is another tool for database marketers to steer sales and marketing strategy to profitable results.
