Tag Archive for 'Customer Profiling'

Cross-selling equals customer retention…but who’s buying?

Published on November 20, 2009 in Brian Diepold. 0 Comments Tags: , , , , .

Brian Diepold, Pitney Bowes Business Insight

It’s pretty well recognized that the deeper the relationship with the customer, the more valuable it is for the bank and the customer.  For years we have been preaching cross-sell as the way to build relationships, gain lasting customers, and run a profitable bank.  It’s also generally regarded that cross-sell leads to customer retention.  There may have been a bit of chicken and the egg with this part, but I think that has been resolved by looking at the timing of most cross-selling.  As we all know very well by now, the great majority of cross-sell takes place within the first 90 days.

I’m willing to take that as evidence of causality – cross-sell does in fact lead to customer retention, not the other way around.  The luxury of having millions of customer data records is that one cannot find themselves bored.  It seemed like a worthy endeavor to quantify the actual impact of cross-sell on customer retention.  And with that, enter 80MM customer account records.  By quantifying this relationship, we can see the direct impact of a firm’s strategic initiative and diagnose the types of households where cross-sell is more important to strengthening the relationship and increasing the probability of retaining the household.  The analysis uses a logistic model to determine the impact of cross-sell on the probability of retaining a household.

The bottom line is that regardless of how you measure cross-sell ratios, you can find the intuitive relationship that more products per household lead to a higher probability of retaining that household.  We also find a diminishing marginal return – i.e., the first cross-sell is the most important to establishing the relationship.  While each additional product does increase the likelihood of retention, the biggest jump is with the first cross-sell.  Rather than rewrite the research paper, I’ll just highlight some of the key findings:

  • Moving from one product to two for the household yields the greatest increase in the probability of retention.
  • Households with low deposit dollars yield an 8 – 10% increase in the likelihood of retention through cross-sell.
  • Households with mortgages or interest-bearing deposits as the lead product also yield an 8 – 10% increase in the likelihood of retention through cross-sell.

If you have the luxury of focusing your resources on certain types of households, these are the ones where you can make the biggest impact.

Consumer Demographic Profiling: Does Distance-weighting Make a Difference?

Published on November 20, 2009 in Nat Evans. 0 Comments Tags: , , , .

Nat Evans, Pitney Bowes Business Insight

It is a standard market research practice to use psychographic segmentation as a primary tool for discerning a company’s target customer.  This “customer profile” creation is a primary means by which customer behavior is bucketed into distinct groups that reflect differing customer characteristics, shopping behavior and loyalty to a retail, restaurant, or consumer package brand.

Over time, the Strategy & Analytics statistical modeling team at PBBI has given a lot of thought to the idea that customer profiles, and their use in sales forecast models, may be enhanced by weighting the customer source survey data by distance.  It makes sense.  The farther away target (or non-) customers are, the more pronounced the profile scores may be.

For instance, a typical customer profile for a high-end department store, with a specific high income, suburban customer segment or “cluster” (Corporate Clout, say, from Acxiom’s PersonicX lifestage segmentation system) may have a score of 200, meaning that people within the segment spend two times what an average customer spends for the concept.  A low-income segment (Single City Stress, for example) may have an index score of 40, or the segment spends 40 cents for every dollar that an average customer spends. Weighting the profiles by distance, however, may yield a more intensified result.  You may expect the high-income cluster to go from an index score of 200 for customers only within 0 to 3 miles, to 225 for those same customers beyond 6 miles from any given store.  Perhaps the Single City Stress cluster would go from 40 to 25, meaning that the farther away the cluster is from the store, the less they are willing to patronize and spend at the department store relative to other customer segments at the same distance.

In theory, it seems to be a reasonably insightful approach.  In practice, the S&A modeling team created just such an analysis, and found that among several clients’ customer databases, distance has no significant bearing on the relative spend of psychographic segments at the same distance.  The following box plot will give a sense for one sample profile’s distribution:

Distance Weighted
As shown, 50% of all scores’ distributions fall between approximately 50 and 110.  A couple of outlying clusters find themselves floating outside the distributions (the “1”, “2”, or “3” above the whisker for each plot), but statistically, no significant difference was found between scores at different distance increments.  The distributions’ medians were roughly the same; variance was the same.  This pattern was consistent among several customer files we tested, and the application of several distance weighting methods yielded no statistically significant enhancement whatsoever.  Goes against the hypothesis.  To be sure, there exists a multitude of ways to carve up customer data, and this analysis is by no means definitive, but from a macro level, it seems the proof is in the data. This analysis also does not mean that distance in its myriad forms (straight-line, drive time, drive distance) has no influence.  Obviously, it does.  Distance decay portrayed on a sales per capita, or other relevant, basis is very important indeed, and is an extremely well documented predictor of consumer behavior.  It’s just a matter of proper application, and what any one retailer’s customer data is truly telling an analytics researcher.

Are There Limits to “Profiling” Your Customer ?

Published on September 16, 2009 in Gary Faitler. 0 Comments Tags: , , , , , , , , .

Gary Faitler, Pitney Bowes Business Insight

In the arena of retail site selection research, we have been a vigorous proponent of the need for customer profiling via segmentation. Our approach has emphasized the bucketing of customers into distinct segments, reflective of differing levels of customer engagement and spend. This segmentation, based on variable demographic, behavioral and psychographic characteristics, ferrets out your most productive customers (i.e. the 10% that generate 40% of sales) and equally important, identifies your least productive. Understanding who these customers are, and where they are concentrated geographically is one of the pillars supporting an effective retail business, as it contributes to critical decisions ranging from store site location (or retrenchment) to the prioritization of investment in marketing dollars.

Pitney Bowes Business Insight has also pioneered the concept of “effective population” in which the segmentation profile becomes the means of adjusting actual population to create a foundational construct, influenced by shopping behavior. This allows us, in one quantitative statistic, to wed the variables of mass and profile. The approach has been extremely useful in rationalizing distance decay patterns across divergent segments.

We have noted recently, however, that a number of our clients are pursuing a “silver bullet” in their approach to customer profiling. We suspect that, as a result of the difficult times confronting retail generally, decision makers have turned to new market research professionals, hoping to ensure the best approach for optimizing sales in a climate of “down” markets. For some of our clients, these professionals, who tend not to think spatially, may, at times, be fostering an inflated view of customer profile – possibly to the exclusion of other real estate fundamentals. As noted above, we view segmentation as one of the pillars supporting effective retail execution and site selection, but others are equally critical.

As champions of the concept of segmentation and customer profiling, we must also emphasize its appropriate use. The importance of customer profiling is variable based on the level of differentiation of the customer. For example, a high-end purveyor of fitness and athletic equipment would represent an extremely segmented profile, while a ubiquitous quick service restaurant would represent an equally extreme non-segmented profile. Both concepts, in the end, would be highly dependant on what we call the fundamentals of sound real estate site selection: convenience to critical population mass, retail synergy and visibility/market exposure. However, for the highly segmented concept, any application would additionally be extremely sensitized to store positioning vis-à-vis the most in-profile customer base.

For some of our retail and restaurant clients that have mass appeal (i.e. minimal differentiation within the customer base), we have recently observed exaggerated expectations from the customer profile. They appear fixated on deriving some insight into their customer base that will drive a deeper penetration than realized to date. As much as we’d like to glean from their customer data some new profile-driven formula for uncovering optimal store placement, in these instances, it is primarily the real estate fundamentals, on which we have traditionally focused, that remain the best predictors of sales potential.

It may be helpful for marketers to think of the brick and mortar unit as a “touch point” itself, with physical convenience and outstanding execution the unique features driving customer response.

[More insights on site selection research from our Predictive Analytic consultants].