Co.Design

How Companies Like Amazon Use Big Data To Make You Love Them

Businesses now sit on data goldmines, but very few leverage the data to improve customer service. Ziba’s Sean Madden suggests 3 ways forward.

Last month, I talked to Amazon customer service about my malfunctioning Kindle, and it was great. Thirty seconds after putting in a service request on Amazon’s website, my phone rang, and the woman on the other end--let’s call her Barbara--greeted me by name and said, "I understand that you have a problem with your Kindle." We resolved my problem in under two minutes, we got to skip the part where I carefully spell out my last name and address, and she didn’t try to upsell me on anything. After nearly a decade of ordering stuff from Amazon, I never loved the company as much as I did at that moment.

Remember, this was a customer-service call, so I was fully prepared for it to suck. Like most American consumers, my experience with service interactions is largely negative, whether it’s on the phone, in the murky depths of a commerce site, or in the aisles of an electronics store. I’m accustomed to the company being in control, and for our communication to be cold, scripted, and inhumane. Barbara’s congenial but no-nonsense approach was part of what made this experience different, but more important, she had access to exactly the right data about me, and that made the favorable exchange possible. The fact is, Amazon has been collecting my information for years--not just addresses and payment information but the identity of everything I’ve ever bought or even looked at. And while dozens of other companies do that, too, Amazon’s doing something remarkable with theirs. They’re using that data to build our relationship.

The Most Useful Data Set in the World

Big Data has gotten a lot of attention over the past 18 months as retail, manufacturing, and technology companies realize the gold mines they’re sitting on and rush to scour them for competitive advantage. Nearly all of this discussion, though, revolves around consumer trends, marketing guidance, new product planning, and other market-level insights. When McKinsey wrote its omnibus report on Big Data last year, the consulting company identified five different ways it can be used to create value, but only one of those methods mentions customers at all, and then only in terms of improved segmentation. The Wall Street Journal outlines several business success stories in its Big Data blog series, but it focuses almost entirely on smarter market visualization, better process maps, and other efficiency enhancers. Efficiency is a worthwhile goal, but from a customer’s perspective, data has far more power at the personal level.

Perhaps the only business and marketing topic that’s been talked about more than Big Data recently is the evolution of brand relationships into two-way conversations. Now that consumers have seen what social media and mass customization are capable of, they increasingly expect this kind of personalization in their communication with favored brands, not just a passive role absorbing marketing messages. Combine this insight with the rise of Big Data, and you have a clear mandate: In order for interactions to feel individualized and human, they must be well informed. That makes data about the customer you’re talking to right now the most useful data of all.

Technically, this is hard to do. Amazon has grown large while staying fairly consistent as an organization, but most big companies got big through acquisition, and that makes synchronizing data a massive chore. Getting targeted information in front of the person who’s dealing with an individual customer, or designing for one, is still a low priority. Customer service in its various forms is still treated as an expense to be minimized, not an opportunity to be developed.

Service designers know that the opposite is true. When a customer calls the support number, sends an email, or talks to a store employee, he is initiating a conversation. You have his undivided attention, even if he’s annoyed, and that makes it a crucial brand-defining moment. He’s hoping for a conversation, but bracing for an ordeal. He knows you’ve collected information on him for your own purposes and wondering why you don’t do something useful with it. Not useful to you--useful to him.

Synchronized data is worth the expense because it’s a hallmark of human interactions. If I talk to a friend and they keep asking me for information I know they already have, I have a right to get irritated. In the age of Big Data, I hold brands to the same standards. The few that meet those standards earn my trust and loyalty. But if you’re hoping to use personal data successfully, there are a few things you have to get right.

1. Give your employees the right tools.

I have no idea what Barbara was looking at on her screen when she called me up, but it gave her the information she needed about me in a matter of seconds. Someone designed the tool that delivered it and made sure she had access to it. Despite your internal divisions, I as a customer have only one relationship with your brand, and it has to be seamless. That’s what makes information tools so vital. They transfer data that’s been collected automatically or through form-filling into the personal realm, allowing us to get the awkward, impersonal, corporate conversation out of the way, and make way for the human one. The rise of portable platforms makes this possible for designers and store employees, too, not just the headset-wearing call-center folks.

2. Let the customer know you know. Then listen.

When I meet an old acquaintance at a party, she remembers my name and asks one or two questions about things we discussed last time we spoke. The fact that she remembers establishes rapport; the fact that she doesn’t list out every bit of information she possesses makes me feel comfortable. Without even thinking about it, humans are very good at conveying just the right amount of information in personal conversation.

Companies need to do the same. When I spoke with Barbara at Amazon, she had access to plenty of data, but only referenced what was necessary, starting with my name and the problem I was trying to solve. It quickly disarmed my self-defense instinct and made me comfortable referencing facts we knew in common but hadn’t explicitly stated. "Can you send it to the Northeast Ninth Avenue address?" I asked when we got to shipping options, even though I hadn’t asked if she had it on file. "Sure," she said, and I smiled.

3. Give the customer a sense of control.

Many of us have read the story of Target’s uncanny ability to recognize a customer’s pregnancy based on her purchasing habits. At first frightening, this revelation sounds reasonable on further review, but no less creepy. Target quickly learned to get nuanced about using this insight. To avoid upsetting these customers (and their parents), they now send them flyers customized to include just a few coupons for prenatal necessities, mixed in with a random assortment of others.

That’s a partial solution at best. In the future, smart retailers will be more transparent about their data-gathering efforts and use the results more appropriately. They’ll give customers more options for controlling how much they share and how that information gets applied. Regardless of who gathered it, customers still see it as their data. They expect to be treated like the owners.

The power of being known

There’s a quiet race going on right now among brands to form customer relationships that earn loyalty in the face of increasing competition, and personal data is the surest way there. Brands like Zappos, Netflix, and Amazon are already showing the power of such an approach. Not only does smart data use empower you to treat customers as individuals, it does so without invoking many of the fixed expenses associated with improved service. Good data support doesn’t require a vastly expanded workforce, or even a new type of employee--these are conversations that people already know how to have.

But imagine the benefits if you get it right. An auto mechanic who’s smart about data could tell you that your fan belt is due for a change in 2,500 miles and suggest doing it today to save future labor costs. An airline that knows more than just your frequent flier number could propose a seat based on your past selections, offer discounted upgrades tailored to your preferences, and let flight attendants know you prefer tomato juice to orange juice in the morning--even if you’re just flying coach. If they’re really paying attention, they could even learn whether or not to offer you an upsell, and in which categories to do it.

As long as they’re given transparency and control, consumers are becoming quite comfortable with these kinds of interactions. In the future, they’ll expect them. When that happens, the question won’t be "How much do you know about me?" but "What are you going to do with what you’ve found?"

[Images: almagami and Everett Collection via Shutterstock]

Add New Comment

23 Comments

  • GeorgeWoodman

    I typed it already.
    Again?
    This is refreshing to read. I have concluded that the main problem with the economy was "customer ervice."

  • BenD

    This article does a great job at turning the attention to service. Amazon's customer sevice quality is standout and memorable precisely because the low competition, and as the American economy turns increasingly towards the service sector, their peers will proliferate. 

    Other than this point, though, I feel that this article largely misses the important points. It seems to me that a "service designer" like the author would be much more disposed to wonder "why is this company asking me for information that I already have?" as this thought has rarely crossed my mind. And takeaways like "provide them with the right information" and "don't dwell on too many points of previous coversations" are no-brainer, and do not advertise very well the author's expertise.

    A much more effective article should provide some hard data on why customer service is important and some ideas for how a company can adjust internal resources and workflows to emphasize service.

  • Tom George

    I just found this article from a curator who posted it on Internet Billboards. Amazon is an amazing company, but like any large company that has access to large amounts of data, there should be some concerns as to how the information will eventually be used.

  • matt819

    On the B-2-B side of things, kudos go to U-Line for one very simple tool they use to provide top-notch customer service.  U-Line provides shipping supplies - boxes, mailers, etc.  When I call, my name, company, and purchase history pops up on their end.  So I jump right in, tell them what I need, and instantaneously they confirm the current price, verify quantity, and move on.  For payment info, all they need is the last 4 digits of my cc and the 3-digit security code, and I'm done.  I don't think I've spent more than 30 seconds on an order.  Granted, this is not rocket science or data mining in the true sense of the term, but it is first-rate customer service.

    In contrast, I seem to spend an inordinate amount of time with other companies just confirming who I am before we can get to the substasnce of the call.

  • Emily


    @customerss:twitter 

    Great article. Amazon is a pretty unique retailer in that it sells such a wide variety of goods, so they should be able to build up a pretty good picture of regular shoppers. 

    In general, I'm always amazed how much data a company doesn't have on me when I try to get problems resolved. 

  • Anita Loomba

    Great article! Customers are connected now more than ever, which is why they are becoming more open to these types of interactions. It will be exciting to see how technology influences these types of changes for retailers. 

  • Drew Mohebbi

    Amazon.com's purchase of Zappos in 2009 should have been metioned (unless I missed it?) That phone call you experienced was almost certainly influenced directly, and definitely influenced indirectly by Tony Hsieh. Tony almost lost all his money (from his sale of Link Exchange to Microsoft back in the late 90s) when he put almost all of his personal assets into to keep it from going bankrupt. Once Amazon didn't spend ~1.2 billion on Zappos for shoes. It was because they were a real threat to Amazon and that was at least partially because of their customer service.

  • Schetikos

    I wasted $17.00 on flowers at Kroger's grocery store, because the person who works in the flower section did not know the types of flowers not healthy for cats if they ingest, which they tend to do.  I did not have my phone to call my wife, so got the flowers based on my recollection.  Sure enough, when I got home, my wife informed me I got the wrong flowers.  Now, if the sales person "who works in the flower section full-time" had known, they could have sold me a more expensive bouquet, and I would rely on this relationship in the future.  Unfortunate for Kroger, I will no longer purchase flowers because this is the type of information/data they should know.  

  • Kam

    Great Article! My team and I are developing a software tool to provide full customer profiles as part of the features. This article is an inspiration and says exactly what we thought about customer service and the need for our solution. Check out www.dropifi.com

  • Bob Jacobson

    Amazon is an amazingly well-tuned service machine.  So is a thermonuclear bomb. The better it gets, the bomb can destroy ever larger cities.  The better it gets, Amazon can more effectively destroy creative authorship and publishing as we know it.  Service like technology can be used for good and ill.  Its quality is more complex than discussions about the effectiveness of filling orders allow.

  • SLJonesDigital

    For the past 10 years most of the artifacts of my work and play float down the Amazon ... and never arrive wet.

  • Eileen M Thornton

    This piece makes a solid argument about using the data to make the customer happy, and paints a pretty anecdote about the author's Amazon interaction, but my Amazon customer service interactions have been really rocky until very recently. They are working on it, but far from getting it right. They still ask for order numbers in chat conversations where the order has been specified, etc.

    Although I order almost everything I buy through Amazon, I still know it's going to be a little bit of a hassle and time suck, and at least a minor frustration, to interact with customer service. There are easy ways to improve further.

    As regards Target, I was surprised (though I understood the teen's situation) to read the customer reactions in the initial NYT piece that describes how they populated the direct mail offers. I think it's brilliant and helpful when a retailer sends me something that meets a need.

  • C. David

    The difference between Amazon and a brick-and-mortar retailer is that Amazon invests *only* in the back end: their warehouses, pick & pack system, iVR/phones, and website(s). They don't have physical stores with merch and a sales staff to worry about, which adds to overhead.

    More often than not, any retailer with a physical presence worries first about that presence ("It's all about presentation" one of my bosses used to tell me) while dealing with greenscreen terminals on the other end of the phone line that have 17 legacy divisions patched together by the three employees that were there since punchcard programming was popular running things because they refuse to document anything as a form of job security.

    I'm only slightly exaggerating in the last paragraph, but I think you get my point; unless the company is TRULY going to invest in the backend processes that are the engine running the business (this could be in-house, outsourced, or whatnot), you'll see that defensive reaction when you deal with customer service, and the continuation of the horror stories of lame support from companies when trying to make a purchase, let alone solve a problem.

  • C. David

    Reply to Michael Fox:

    "Innovations such as "customers who bought this, also bought that" need to be accurately populated from the back-end, but presented in a timely and compelling way to the customer, while the continuous presentation of new, relevant products is very much a shop window - all front-end activities."

    Understand why you would consider their website to be their 'front-end", but there are a number of fundamental differences between a website with suggestion shopping auto-populating vs. a physical store with merch has has to be fluffed/folded, put on mannequins/display, asset tagged, recovered from the fitting rooms, etc.
    Same would be for a physical bookstore...say a Borders (let's ignore their poor financial decisons; I'm just talking about the manpower required for showcasing/presentment/recovery). It takes time to unpack, display, recover, stage all the "stuff" in a bookstore, which Amazon doesn't have to worry about.

    There is a cost for the code monkeys/web designers, of course, but they can change up the website in a billionth of the time it would take a traditional retailer to turnover their selling floor...a digital front-end by its very nature could respond to customer feedback and selling before a B&M store could mock up the plan-o-grams and get the staff moving.
    But, the selling floor is where the "show" is for traditional retailers vs. the website for an online retailer. Even when the physical retailer is online with their merch, they now need to build 2x the infrastructure to fulfill purchases, even if they can lean on internal shipping (Store-to-Door and inventory levelling) process.