We cannot solve our problems with the
same thinking we used when we created them.
I believe that making smart, fact-based decisions is absolutely essential to customer-centric leadership. I’m sure many of you reading this will say to yourselves, “We’re already doing that.”
Sorry, but you’re probably not. As you’ll learn, many decisions are made out of habit, not conscious thought. And others are influenced by biases we all have. But analytics technology alone is not the answer, either. Not every insight can be put into a spreadsheet, analytic model or computer system.
Here are five ways that that data analytics can help improve the customer experience, build loyalty and boost company profitability.
BIG IDEA 1: Practice Better Data Science
How can we expect to make rational decisions about customers, when customers themselves are not rational? In his provocative 2008 book Predictably IrrationaI, behavioral economist Dan Ariely argues that people don’t always make decisions by rational choice, carefully weighing benefits of a potential action against the costs. Instead, we are easily swayed by our habits, emotions and distractions.
Nobel Memorial Prize winner Daniel Kahneman, author of the 2011 book, Thinking, Fast and Slow, found that we humans use two systems for thinking:
- System 1: Fast, automatic, frequent, emotional, stereotypic, subconscious
- System 2: Slow, effortful, infrequent, logical, calculating, conscious
Even in conscious “System 2” thinking people struggle to think statistically. We are easily swayed by a small sample of readily available data that supports something we want to do, and ignore other factors that should be considered. This “optimistic bias” may explain why a good rule of thumb in IT project planning is to take the most conservative (highest) estimate of time and costs, and then double it!
Unfortunately, statistics can be just as misleading as human judgment. As economist Ronald Coase famously said, “If you torture the data long enough, it will confess.”
Excellence in Customer Experience (CX) is correlated with business performance. But I haven’t seen any studies that show that an improvement in CX is followed by improved business performance or that CX is the only factor involved in driving success. In fact, other loyalty studies show that product quality and price/deals continue to have a big influence on customer buying and retention.
This is an example of confusing correlation with causation. Other common mistakes include over extrapolating results to a larger population, drawing conclusions from insignificant differences, and failure to use a control group.
The expanding world of Big Data has elevated the role of Data Scientists, who should have a combination of technical skills (computer science, math, modeling and statistics) along with the business acumen to identify problems worth solving and influence business leaders to take action. That’s a tall order. McKinsey estimates the U.S. faces a shortage of 140,000 to 190,000 such analysts.
BIG IDEA 2: Optimize Marketing Spend
Business leaders are turning to analytics to uncover insights in so-called Big Data, an IT industry buzzword to put a spotlight on the increasing volume, velocity, and variety of digital information. And marketing is ground zero for many of the applications. Marketers are charged with figuring out the right combination of products and services that will appeal to customers, and optimizing the ROI on marketing spend.
Macy’s is a great example of a major retailer competing for the loyalty of omichannel shoppers—those using multiple channels such as retail stores, web sites, mobile devices and even social media. Five years ago, the company began a shift from product- to customer-focus, led by Julie Bernard, Group VP of Customer Centricity.
Speaking at a Forrester conference, Bernard said her goal was to “put the customer at the center of all decisions.” Sounds good, but old habits die hard in a 150-year-old brand where data was organized around products. The retailer used POS data to analyze product sales, but couldn’t figure out what individual consumers were doing. By also looking at data from loyalty program, credit cards and other sources, Macy’s was able create a more complete understanding of the products, pricing and experiences that move “loyals”—those consumers already buying regularly.
One of my favorite retailers, Nordstrom is an old company that is embracing new technologies. At a 2012 analytics conference, James Steck of Nordstom’s Advanced Analytics group discussed how the retailer used analytics to understand product and brand relationships. The idea is simple: figure how to promote the right products and brands to the right customers, maximizing revenue in the process.
Perhaps a simple idea, but not a simple problem to solve when you’ve got a busy website (www.nordstrom.com) along with 225 stores doing around $10 billion in sales annually. Nordstom analyzed 2,000 consumers over a one year period, covering 164 brands. They found 50% of shopper bought brand “A.” How then, can the retailer find those more likely to buy brand “B?” Turns out that analytics could identify a group of customers spending more than $187 had a greater likelihood to buy brand B. Armed with that info, marketers could make more effective merchandising and promotion decisions to increase sales.
BIG IDEA 3: Personalize the E-User Experience
Put yourself in the shoes of Joe Megibow, VP and General Manager of Expedia US, which serves millions of travel shoppers each month. Let’s say you want to present shoppers with hotel options in the New York Area. Megibow says most users won’t do a complex search of the roughly 800 hotels, so it’s critical that Expedia put the “best” options at the top of the list.
If your instincts told you to present the cheapest or more popular hotels first, Expedia would frustrate a lot of shoppers and lose bookings. That’s because the options most likely to meet customer demand depend on a number of factors, like real-time availability, inventory by class, rate deals, reviews, purchase frequency and more.
Using technology from an analytics software vendor, Megibow says they built a predictive analytics model based on the handful of factors that really mattered, out of about two dozen possibilities. Then they operationalized the model using their own proprietary technology. Result: when a consumers searches on NY hotels, they’re more likely to get the hotels that they really want.
Another interesting example is LinkedIn, which accumulates a massive amount of data as its users interact. Speaking at the marketing conference in 2012, Senior Data Scientist Scott Nicholson said they can use that data for the benefit of users (e.g. find future job opportunities) or for LinkedIn (e.g. present the right ads for monetization). What complicates matters is that LinkedIn offers lots of choices—40 different actions according to Nicholson. Using analytic techniques too complex to discuss here, LinkedIn can serve up experiences that are more personalized to the user, or ads that are more likely to be clicked on.
Analytics can also be used to optimize email marketing. Gilt Group’s Tamara Gruzbarg says the retailer is using analytics to influence their merchandising and promotion strategies. The heritage of the retailer is upscale “urban fashionistas,” but as the company has grown and expanded it has become more challenging to make smart decisions. Email remains a critical promotion channel even as users adopt mobile devices. One predictive model that paid off helped Gilt tune the email frequency based on engagement and age indicators, to maximize revenue while minimize unsubscribe rates.
BIG IDEA 4: Mine Insights from Unstructured Text and Speech
If you’ve eaten at a restaurant or shopped at a store lately, you may have seen an invitation on your receipt to call a toll-free number and respond to a survey about your recent experience. Surveys are now commonplace tools for companies to solicit customer feedback. Most surveys allow customers to add their comments. Other sources of text feedback include call center agent logs, transcripts from recorded phone calls, mobile (SMS) text messages, email messages, chat sessions and posts on discussion forums or blogs.
Text mining can be used to 1) determine what the original author was trying to say or 2) learn something completely new. Typically the goal is to identify the topics or categories in the text, such as products, service problems, etc. Taking it a step further, it’s also possible to use “sentiment analysis” to determine whether written text has a positive, negative, or neutral tone.
Garden Fresh Restaurant Corp., operating a chain of over 100 buffet-style restaurants, wanted to take better advantage of customer feedback in transcribed phone survey comments. It was cost prohibitive to manually process about 10,000 pieces of unstructured text a month, so Garden Fresh worked with a text mining vendor to find meaning in all that feedback. As a result, they created new monthly “praise” and “complaint” reports, showing important trends and drilling down to individual customer comments. One insight mined from customer comments resulted in an expansion of the soup varieties offered.
At JetBlue, text mining was introduced as a result of the infamous New York ice storm of 2007. After being overwhelmed with 15,000 emails in just two day, a text mining vendor helped the airline learn that customers were upset about the delays and cancellations, and disappointed that JetBlue didn’t have a backup plan. Since that crisis, JetBlue has worked to more systematically mine customer sentiment, as well as providing “tangible data around how to augment JetBlue services,” according to Bryan Jeppsen, the airline’s customer feedback analyst. By tying feedback data to a specific aircraft or even a seat number, they can find and fix problems that have a direct impact on the customer experience.
If you have called a customer service department for help, you probably heard a message like this: “Your call may be recorded for quality assurance.” Recorded speech is yet another form of unstructured information with valuable insights.
For example, Blue Cross of Northeastern Pennsylvania used speech analytics to understand the reasons for extremely high call volumes. According to Bob McDonald, Customer Service Director, they were able to validate that, in one case calls were the result of something anticipated—a recent system change. The data gave him “ammunition that the problem really needed fixing.” In another case, McDonald discovered, much to his surprise, that customers were circumventing processes to get faster service. Armed with this insight, McDonald changed the call flow and improved agent training.
BIG IDEA 5: Harmonize the Cross-Channel Service Experience
One of the more advanced uses of analytics is optimizing cross-channel customer journey. Interaction data is usually managed in silos, and you can’t easily get a complete picture of what’s happening as a customer navigates multiple channels. CustomerThink’s 2009 study found that consumers exposed to companies suffering from “touchpoint amnesia” (requiring customers to repeat information during a multi-touchpoint experience) were 50% less likely to recommend that company.
In 2007, the mobile telecom company Sprint achieved unwanted notoriety by firing its unprofitable customers for making excessive support calls. Unfortunately, leaders failed to account for the media backlash. Worse, firing “bad” customers didn’t address the core issues of why those customers were calling and therefore unprofitable.
Well, Sprint engineered a turnaround by systematically uncovering and fixing customer service problems. The process took a couple of years and required top management to get serious about improving the customer experience. On a CustomerThink webinar, Lance Williams of Sprint Nextel explained that in 2008 Sprint had the worst IVR customer satisfaction in the industry. They used cross-channel analytics to understand why customers were abandoning the IVR to call the agents—a frustrating experience for customers and a very expensive issue for Sprint. After improving customer usability, by Q4 2009 Sprint’s IVR CSAT was leading the industry. That helped the contact center to “contain” (complete interactions in the IVR) “tens of millions more calls” in 2009 as compared to 2008. Translation: huge cost savings.
Sprint’s improvement has been impressive. In 2008, Sprint’s ACSI score (a measure of overall satisfaction and loyalty) was a dismal 56 vs. an industry average of 68. By 2012, Sprint’s score had improved to 71, a point above the industry average and competitive with other major mobile operators.
The use of cross-channel analytics is becoming more accepted now, but justification can be challenging when managers fight over budgets. One large technology firm told me it took about three years for management to get fully on board and finally commit some real budget. Once they did, analytics showed the true cost of bad cross-channel experiences in agent call time savings. Similar to Sprint, poorly designed IVR systems forced customers to call an agent for help. This degraded the customer’s experience and wasted agent resources.
Towards Better Decisions
What makes Big Data interesting is the new types of information such as website clickstream data, social media posts, video surveillance feeds and even sensor data from consumer products. These new forms of data definitely pump up the volume, requiring new data storage techniques such as Hadoop. And, there’s probably a tool that can analyze any data you can collect.
However, more data (“big” or otherwise) doesn’t necessarily mean better decisions. The key is picking right decisions, says James Taylor of Decision Management Solutions. The biggest mistake is to start with the data or the technology, rather than the decision. “Big Data projects should focus on how to improve how we run the company,” advises Taylor.
A recurring theme from industry experts is the importance of knowing what’s possible. While so-called data scientists are emerging high-impact positions designed to mine Big Data effectively, I believe the real leverage is in data strategists. These are business leaders like Julie Bernard at Macy’s and Joe Megibow at LinkedIn who focus on key decisions that improve the customer experience and increase profitable revenue.
Further reading: The Origins of ‘Big Data’: An Etymological Detective Story (NY Times)
This article is based on Bob Thompson’s e-book How Customer-Centric Leaders THINK Clearly with Analytics and Big Data, available for free download to CustomerThink members.