BrainTrust Query: Does Big Data Help Retailers Really Know Their Customers?

Discussion
Mar 29, 2012

For years, retailers have worked to collect “Big Data” about their customers. From transaction history, purchase frequency, click-throughs and annual spend to demographic, geographic, and weather related appends, retailers have spent millions trying to create a more complete picture of what drives their best customers to shop. This data is harnessed in multiple ways — from complicated predictive analytic modeling to display retargeting — all applied to a frenzy of marketing activity designed to drive incremental sales.

But does all this data really help retailers know their customers in a meaningful way?

In a recent (private) RetailWire BrainTrust discussion, I asked my fellow panelists to come up with what they considered to be the biggest retail myths being propagated today. Several members replied that the idea that retailers know their customers is a big myth. To quote Roger Saunders, “Retailers know the how, who, what, where, why, and when of their customers’ behavior in their store, but they do not know the complete customer.” Ryan Matthews concurred that knowing purchase patterns and behaviors is “not the same as knowing people.”

It appears that several prominent and successful retailers agree. According to last week’s Wall Street Journal, Lululemon, the hugely profitable and wildly successful purveyor of yoga wear, eschews “Big Data” in favor of old fashioned techniques such as walking the store, talking to customers and eavesdropping on dressing room conversations to figure out what customers want. Lululemon has built a billion dollar plus retail brand where inventory is scarce, discounting is rare and 95 percent of its stock is sold at full price.

Costco could use its membership card to track customer purchases and distinguish between loyal and occasional shoppers. Instead Costco makes its membership card a profit center, ignores complex customer data analytics and focuses instead on having a unique customer experience built on great brands at great prices, clean wide aisles, well trained store associates, a cheap cafeteria and treasures buried in different aisles.

Discussion Questions: When it comes to Big Data, are retailers fooling themselves into thinking they know their customers? What is the right mix of quantitative data, qualitative feel (through social media, in store visits, observation) and simple gut retail instinct for running a retailer today? Is there such a thing as “too much data” or is all data good data?

Please practice The RetailWire Golden Rule when submitting your comments.

Join the Discussion!

38 Comments on "BrainTrust Query: Does Big Data Help Retailers Really Know Their Customers?"


Sort by:   newest | oldest | most voted
Bob Phibbs
Guest
10 years 5 months ago

The best retailers still know it is about you first; technology second. While an algorithm can show that the person who purchases canned tuna fish is likely to buy a toothbrush and tends to buy bananas, how much of that is truly useful to moving the needle? I would offer much less than BIG DATA would have you believe, becoming more of an anecdote for those who wade through it all than “knowing their customers.”

Paula Rosenblum
Guest
10 years 5 months ago

Oh, I think retailers know they don’t know their customers. And they’ve had “big data” forever. Retailers have always been data junkies.

What they haven’t been able to do is put the data together in a meaningful way to draw reasonable conclusions about the customer, rather than the product. Or as my partner Brian puts it, they’ve used product movement data as a proxy for information about the customer.

So with multichannel retailing, customer patterns have become better exposed, and thanks to Moore’s Law, we now have the capacity to actually aggregate it up into something meaningful.

It’s very early. I would say our ability as retailers to absorb the rate of technological change is far slower than its actual advances. Hence, probably in about 5 years, more retailers will be actually using big data.

Nikki Baird
Guest
Nikki Baird
10 years 5 months ago

I think it’s possible to spend a whole discussion on what “know” really means. I do think that retailers in some ways fall into the economic trap of sunk costs — they’ve been collecting customer data for years and they have a ton of it. It’s got to be worth something right? Or even worse, there’s that corporate myth that the secret to purchase behavior could be unlocked by plumbing the depths of that data — that the bytes and bytes of disc space that customer data occupies actually hides treasures of untold value.

I’m not convinced. What does a retailer really need to “know” in order to “know” customers? Behavior, definitely. Intention would be great. But why do they have to be so sneaky about divining customer intention? Why not just ask them what they want? That’s one data trade for value that I suspect most consumers will be happy to make: If I tell you what I’m trying to accomplish, will you actually make it easier for me?

Dr. Stephen Needel
Guest
10 years 5 months ago

All data is good and more data is better, until you reach the point where you don’t know what to do with it or how to leverage it. Modeling Big Data only takes you so far — the next step is to test the ideas it should generate. That’s what you don’t see much of — testing new ideas generated by the data/models.

David Dorf
Guest
10 years 5 months ago

With scale, retailers have lost intimacy. Expecting store staff to know their shoppers is tough, so retailers have tried segmentation, loyalty programs, focus groups, etc., all of which help but aren’t quite as good as truly knowing customers.

I believe so-called Big Data can help paint a better picture when both demographics and psychographics are combined, but accessing the necessary data can be tricky. Retailers need their customers to volunteer the information so there are no privacy issues.

Personal recommendations, offers, and pricing are already in the marketplace being powered by Big Data.

John Boccuzzi, Jr.
Guest
John Boccuzzi, Jr.
10 years 5 months ago
If you don’t know the questions you want to ask, neither “Big Data” nor any data is very useful. You may accidentally come across an insight by accident, but that is 1) rare and 2) when you find the insight how do you react to it? If you don’t plan on reacting to what you learned from data, stop spending time collecting and analyzing data. I like the Lululemon approach to collecting insights and it sounds like they take action from what they learn. Tough to beat old fashioned, in-person conversations to learn more about your company, brand and service. 101 on how to use Big Data (defined by me): 1) Know the questions you are looking to answer before you hit the request key; 2) Have a commitment from the organization that they will take action when that answer is discovered; 3) Don’t count on just 1 data source; 4) Verify your answers with customer interviews; 5) Test your answer; 6) If the test proves to work deploy before others in the market mimic… Read more »
David Biernbaum
Guest
10 years 5 months ago

In too many instances, the data is used to look through the rear view mirror. In other words, the process they use lacks vision.

Ryan Mathews
Guest
10 years 5 months ago
The value of a tool depends on three variables: the inherent quality and characteristics of the tool itself; the skill of the tool user; and, finally, the task for which the tool is being used. One could have the world’s finest screwdriver and place it in the hand of the world’s finest mechanic but it wouldn’t do them much good if you were asking him/her to use it to pound steel into shape. The same is true with data. I suspect the food industry isn’t using the “best tool” — a balanced constantly evolving blend of POS data, social media feedback, EFFECTIVE customer communication and a half dozen other variables. I further suspect that most people in the industry are skilled enough at the kind of robust, deep analytics needed to put an effective human face on the data. And, finally, I think retailers persist in misdefining their mission — defining it as the movement of inventory rather than the satisfaction of customer needs, both simple and complex. Just accruing more numbers in and of… Read more »
Liz Crawford
Guest
10 years 5 months ago

Analyzing big data is no substitute for customer intimacy. But I don’t think it is intended for that.

Instead, I think Big Data will help retailers more effectively offer personalized products and deals to shoppers online. The integration of this data to in-store shopping is still far off.

Ted Hurlbut
Guest
Ted Hurlbut
10 years 5 months ago

Successful retailing requires a skillful blending of art, craft and science. All the data in the world is just numbers if there’s no context. The challenge for Big Retail has always been converting Big Data into concise, meaningful and actionable Big Knowledge.

Distinguishing between what’s important and what’s not has always been a challenge. All too often, less is more. All that data can ever tell you is what customers have DONE. It can’t tell you what customers are likely to DO. Past performance is not necessarily an indicator of future results.

Which is where the art and craft of retailing comes into play. There’s no substitute for a merchant’s accumulated, applied wisdom and experience. There’s a lot to be said for Lululemon’s and Costco’s approach.

Ed Dunn
Guest
10 years 5 months ago

For retailers, the best use of big data is transaction centric, not customer centric. I think it is a misconception in the industry to focus on profiling customers and collecting data on customers.

Walmart does not use big data for customers and does not have a “loyalty card.” Instead, Walmart appears to focus more on using big data to move inventory to its stores, transition seasonal products to the highest visibility traffic areas, and create loss-lead scenarios such as selling soda below cost during the summer months.

Creating the right environment to maximize transactions should be the focus of big data, not profiling 20% of customers.

Ian Percy
Guest
10 years 5 months ago
There is math and there is mystery. In the 17th century, Newton convinced everyone the world and everything in it is all mechanistic, all controlled by mathematical formulae. No doubt about it, math explains lots of things. But we have to recognize the mystery too. There is another dimension where the highest possibilities live that is accessed a different way. Its secrets are given to us in dreams, intuitions, imagination, serendipities, feelings and so on. Look no farther than the stock market. Do you think its ups and downs are driven by the math or by mystery? Most of us know that a single word by someone in authority, or even their mood on a given day, can send the market wild, never mind what the facts are. Even MIT admits that well over 80% of all discoveries and innovations happen OUTSIDE of formal structures, academic analysis and so on. The big breakthrough insights — even those applied to customer service — come from the mystery, from another dimension. If it was ALL about the… Read more »
Cathy Hotka
Guest
10 years 5 months ago

I was once on stage and asked a large retail audience if they shopped in their own stores. Many hands went up. I then asked them to lower their hands if it was their spouse who did the actual shopping…and many hands came down.

Data is no substitute for hands-on observation, and not just during the holiday season. That said, as the technology improves and new practices are put in motion, retailers will get better at leveraging the data they already have.

Bill Bittner
Guest
Bill Bittner
10 years 5 months ago
There are a million quotations that buttress the tenor of this question. One of my favorites is that “Big Data provides us a plethora of transactions and a dearth of wisdom.” The challenge is that like most generalizations (including this one), Big Data is not always useless. Like any other tool, big data can be useful when it is put into the hands of a skilled practitioner who knows how to use it and has the additional tools necessary to be effective. For years I had a frequent shopper card whose registration was an old street address and did not include an email address. The advantage for me was that I never had to put up with mailings that meant nothing to me. The disadvantage to the retailer was that they had no way to reach me. Information about the buyers is one thing retailers can still monopolize. The foolish thing is to give this information away to manufacturers or market analysts. Even if there is a smaller payment, I would only provide third parties… Read more »
Frank Wagman
Guest
Frank Wagman
10 years 5 months ago

There is a saying in the computer world — “garbage in, garbage out.” Whether it is big data or little data, if the process of collecting it has no integrity, then it won’t do you any good. However, that shouldn’t dissuade companies from investing in the technology that captures and collects the data, because you can’t improve anything unless you can measure it.

Tony Orlando
Guest
10 years 5 months ago

I don’t need a 50 page report to know who my core customers are. Shopping habits have changed so much in the last 5 years, and it is no secret that consumers shop anywhere and everywhere. Online sales have also taken away from brick and mortar stores which skews the data as well. Tracking someone’s personal shopping habits in your store can be done, but what’s done with it is up to each retailer. I do know that true hardcore customers are less than 20% of my base, and we do take care of their needs quite well. Building loyalty with the remaining 80% requires the same amount of great service, but most of them will support your store if the deals you have drive the sales at the front end.

Adrian Weidmann
Guest
10 years 5 months ago

There is only too much data if you don’t take the time to understand what it is telling you. It is also true that that you have too much data (and wasted expense) unless you act on what the data is telling you!

We designed, measured and analyzed a comprehensive in-store digital media initiative for a Fortune 50 retailer. We had extensive quantitative and qualitative data which clearly showed trends and provided guidance for what worked and what didn’t. The data strongly suggested that existing processes and methods were simply not working and changes had to be made. The data, when translated into actions, clearly showed that the familiar and culturally accepted workflows and retail-centric ‘best-practices’ were simply not resonating nor engaging the shoppers in a meaningful way. The suggested changes dictated by quantitative data based on shopper behavior were deemed to be too disruptive to the ‘status quo’. Was it a matter of too much data or too much bureaucracy and ‘not enough pain’ to facilitate a change?

David Zahn
Guest
10 years 5 months ago
My thoughts are similar to John Boccuzzi’s on this one. Ask questions that matter or the dominos will fall in unanticipated directions. The efforts that support the “actions” of a poorly implemented initiative will amount to being “the emperor’s new clothes” at best. The reliance on data, analysis, number crunching, etc., has a place in retail — but NOT at the expense of the following: 1) Standing for something or differentiating yourself from others (how few retailers or manufacturers for that matter are TRULY distinctive from competition?) 2) Being able to communicate it and demonstrate it in meaningful ways 3) Reacting to pressures, competition, changing customer dynamics, etc. in a way that is consistent with the company’s mission, strategy, and identity (versus a “whoops, market share is down, let’s lower prices!”) 4) Enabling staff to execute against the strategy(ies) of the company in a consistent with the corporate strategy, but autonomous way. The problem is not big data — it is a tool (as Ryan Mathews correctly points out). It is in the environment and… Read more »
Dan Frechtling
Guest
10 years 5 months ago

I agree with the consensus that divining patterns through big data isn’t a substitute for simply asking customers what they want. Bob, Paula, Nikki, Ryan, Cathy, Bill and others have made a good case here.

“Data mining” succeeds in huge batches and works best through digital interfaces. “Asking customers” succeeds at the individual level and works best through associates in-store, as Liz suggests. It also provides good service.

Analytics has the power to centralize things that were decentralized. It gives the capacity and sometimes the illusion of control. The Achilles’ Heel is that each retailer only has a sliver of information about a customer’s purchase history — those purchases made in its own stores. That does not provide a complete picture for analytics, and is another reason analytics can’t replace asking.

Doug Garnett
Guest
Doug Garnett
10 years 5 months ago

A set of observational facts always appears important, but reveals little about people. Mobile data promises to make this problem even worse.

So in reality, it’s a myth driven by the most common research error — let’s call it the Einstein for his articulation: Much that can be measured does not matter and much that matters cannot be measured.

Anne Howe
Guest
10 years 5 months ago
I’m not certain there is one right answer here or one ideal mix of data. But knowing what you want/need to know by having a knowledge map and a clear sense of where your insight gaps are certainly goes a long way to help use data efficiently. I’ve also had a very long-held rule that looking for the SoWhat? is only a good thing if you’re prepared to follow it up with the DoWhat. That’s called Insight into Action! For a long while, retailers really haven’t wanted to change their long-term ways of doing business with shoppers. But, through data, a deep understanding of what shoppers are really doing, especially with technology enabling the mirroring of offline behavior to online, has caused retailers to take notice. Data has helped them face the future reality, and many are now using data in very smart and productive ways. Suppliers and partners who are willing to help them ask the right questions and use data to fill insight gaps will be better off going forward than not. Guts… Read more »