RSS Twitter LinkedIn
The Global Resource for Loyalty Marketers

Insight: How to avoid the mirage of data

In an article published on LinkedIn, airline consultant David Feldman uses several case studies to outline the ways in which "big data" can betray marketers. The biggest danger, Feldman argues, is that data betrays marketers when marketers become convinced that all of the answers they need can be found in the data, when in fact the data contains significant blind spots - a phenomenon that marketing consultant Gerardo A. Dada calls the "mirage of data." How can marketers leverage data without falling victim to its limitations?

By Rick Ferguson
 
Feldman (who has fast become one of my favorite loyalty commentators) higlights three loyalty case studies that demonstrate the limitations of a "big data" approach to customer loyalty:
 
Woolworths: We have on this site documented at length the struggles of Australian grocer Woolworths with its loyalty program. Feldman tells the story of how Woolworth's loyalty data analysis told them that their program members weren't redeeming Qantas points earned while shopping, and therefore wouldn't miss them. Woolworths ended the Qantas partnership, suffered a firestorm of negative customer reaction, and was quickly forced to reinstate it.
 
Starbucks: As Feldman points out, Starbucks is held up - including by us - as the "Gold Standard" of retail loyalty programs. But Starbucks was betrayed by their loyalty data when they failed to see that changing their reward structure to reward big spenders at the expense of frequent customers would result in a backlash. Money quote from Starbucks Global Chief Strategy Officer Matthew Ryan: 
 
"What we are seeing is basically how we record the transactions of customers in our stores and what we had in the past, but we didn't necessarily have visibility to, are people or parties of people coming in and buying multiple items ... And, of course, there would have been no way for us to see that before. But what we saw were multiple beverages being put on Starbucks Rewards accounts where they hadn't been put on there before."
American Airlines: The airline eliminated an aspirational reward option for which its data showed that AAdvantage members rarely redeemed. The inevitable backlash ensued and American backtracked - while still believing that their interpretation of the data was correct.
 
Feldman then quotes at length the "Mirage of Data" post by Gerardo A. Dada, which is also worth summarizing in this space. Here are Dada's "Seven Hazards" of a Big Data approach to marketing:
 
  1. Data only looks at the past. In other words, heed the warnings of every stock prospectus: "Past performance is not an indicator of future results."
  2. Data does not reflect emotion. Transactional data can tell you everything about a customer's purchase behavior except for "why."
  3. Data is always biased. As most US Presidential election pollsters can tell you in hindsight, data interpretation is always subject to "confirmation bias"- we focus on data that confirms our preconceptions while ignoring data that contradicts them.
  4. We focus on the data we can collect, not the data we need. Because data collection is difficult, we tend to focus our analysis in the data we can collect, even if it isn't useful to our analysis.
  5. We confuse correlation with causation. This is the classic conundrum of statistical analysis, and one that extends far beyond the loyalty space. Confirming that incremental behavior is driven by a loyalty stimulus, rather than merely correlated with it, is the toughest job in loyalty marketing.
  6. The delusion of a single explanation. As marketers, we tend to look for a single cause for customer behavior - a loyalty promotion, for instance - when the actual reasons for that behavior are far more complex.
  7. More data isn't better data. We marketers tend to "drink from the firehose" by collecting every scrap of customer data we can get our hands on - which, contrary to allowing for deeper analysis, actually increases the danger of wrong interpretation.
So how can we overcome these inherent hazards of a Big Data approach to marketing? Both Feldman and Dada advocate a renewed focus on listening to customers and grounding our marketing efforts in a deep understanding of human psychology. Here's Dada:
 
"The best insights come from customers, not spreadsheets or analytics. Sometimes we look for technology, sophisticated data models, complex analytics and expert advice to tell us more about customer behavior, when it would be easier to simply get out of the office and listen to them. The best insights come from observing and listening to customers. If you make it a habit to speak with a customer or three, every day, you will gain insights that no computer technology can give you."
And here's Feldman:
 
"It's important to understand that data is only one part of the equation and when it comes to loyalty programs - a deep understanding of customer emotions, behavior and psychology is a mandatory ingredient - to being able to accurately leverage the data you have."
True statements both. While neither commentator argues for abandoning a data-based approach to marketing, there is an equally likely danger inherent in their prescribed cure. Basing important marketing decisions on a Malcolm Gladwell-esque gut feeling, a few focus group sessions, your Twitter feed, or a few conversations with customers at the coffee shop can also distort your analysis. This distortion is often due to the "say-do gap" in behavioral analysis - what customers say they do in surveys and focus groups often does not resemble their actual behavior.
 
Listen to your customers, certainly, but test the validity of those conversations against the data. Rather than drink from the firehose, focus your data collection and analysis on that data most predictive of customer behavior. By focusing on data collection at each stage of the purchase cycle, and throughout the lifetime of the customer relationship, you can become very adept indeed at understanding how predictive those glimmers of behavior early in the relationship are of future customer value. Armed with this insight, you can proactively influence the value of those relationships through targeted application of reward and recognition.
 
Big Data isn't necessarily the answer - a "Small Data" approach to marketing, combined with the listening-based approach advocated by Feldman and Dada, can help you avoid the data mirage and locate, at long last, the loyalty oasis.

Read the Feldman article here.
 
Rick Ferguson is CEO and Editor in Chief at the Wise Marketer Group.
  • rDialogue: The Future is Now - Loyalty 3.0