Comments on Securing the Long Tail

by Eric K. Clemons, Fellow, Cutter Consortium

David Lineman makes a number of points in his wonderful Cutter IT Journal E-Mail Advisor "Securing the Long Tail." Some of the points, while well known, are still true and important to state. Some of the points are even more important to state, because they are based on popular misconceptions that are newly demonstrably incorrect, for reasons that reflect changes in consumer behavior as well as changes in corporate strategy they have produced.

Let's begin with the first part: Mr. Lineman's restatement of the need to secure the long tail. There are now many opportunities to capture information during every interaction with customers, many opportunities to use that information to benefit the customer for your own competitive advantage, and many opportunities to offend the customer by misusing this information or merely by admitting having it. There are also unique opportunities to destroy customer trust and lose valuable customer relationships as a result of hackers gaining access to and misusing customer history information.

This is a real concern. A report on Google's privacy policies, rating Google as among the worst top Internet destinations, was released on Saturday, 9 June [1]; by 11 June, the story had been cited on close to 15,000 Web pages.¹ This problem will not be easy for Google to address; tracking searches, retaining users' search history, and then mining this history for patterns and trends for sale is the basis of Google's business model. While Google may appear to enjoy insurmountable advantages, old-timers like myself remember when Gopher and then Mosaic also seemed to be untouchable -- until they were bumped off by Netscape. If Google cannot address consumer concerns, it too can fail.

My earliest conversations with service providers over customers' privacy were two decades ago, with the customer service representatives in guest relations at the InterContinental Hotel in London. The staff learned that I ate the red grapes in the fruit basket and that when my young daughter was in London with me, we would eat all the strawberries they gave us. They learned to put the welcome bottle of champagne in a bag with a handle to be in the room when I arrived alone but to transfer it to an ice bucket if I showed up with my wife; they noticed that I always brought the champagne home to my wife if I was there alone but that my wife drank it if she joined me in London. The guest relations woman eventually explained the pattern laughingly to my wife. But what if the women in the room drinking the champagne with me on early visits had not been my wife? I discussed this with guest relations later, in private, and I was reassured that they would never have had this discussion with my wife if I had had other women in the room. I then realized that the staff knew exactly who in the hotel was sleeping with whom, and when. This provided an interesting insight into just how much hotels know about their guests, which was confirmed with discussions with guest relations staff at other hotels with frequent guests and high staff-to-guest ratios.

I later asked guest relations how it determined what data to collect and how to use it. The general rules were:

1. Only collect data on guests who have joined the hotel chain's Six Continents Club and have been informed that their preferences will be monitored to improve service.

2. Only collect data that could actually be used to improve service.

3. Never share data outside the hotel or even outside the small number of hotel service staff who actually need access.

In his piece, Mr. Lineman makes three points on the importance of policies for the capture, use, and sharing of data, and a fourth point on the need to explain this policy to customers. These are also the basis of an EU policy described approvingly by Mr. Lineman. The points stressed by the InterContinental staff 20 years ago are the same as the first three points made by Mr. Lineman and by the EU. The hotel did not explain to me how it collected its data until I asked, and hotels in general do not describe their data management policies; the reason is simple -- explaining how the hotel performs its magic reduces some of the impact of this magic. Staff members are given a list of frequent and VIP guests each morning, along with a photograph and details relevant to their stay, enabling everyone to greet an important visitor by name; this is much less impressive if you know the doorman sneaked a peek at his crib sheet while you were getting out of your taxi. Still, given how much data is now stored electronically, available both to insensitive automatic systems and hackers, Mr. Lineman's fourth point -- "Tell the customer exactly how you are going to do steps 1-3" -- is more important than ever before.

Informational mistakes can potentially be very embarrassing. A friend stayed alone in a great hotel on a business trip; the hotel accidentally recorded that he drank champagne with two glasses and ordered candles in the room on this trip. He found out only because he was billed for the champagne and the candles, and he made sure that the hotel removed this mistake from his billing records. Imagine if he had not; the last thing he would want from a hotel that tracked his every whim is to show up later with his wife and find champagne and candles on the table in his suite, with a note saying, "As you required the last time you were with us!"

What about a financial services firm that notices you are charging more and more auto repairs on your credit card, realizes that you are a great credit risk, and offers you a preferred rate on a car loan for a used car? This may be really helpful and valuable to you -- but do you want them knowing this much about you? What if your credit card issuer (Citi) were also your life insurance carrier (Travelers),² and it noticed that you were charging a lot of alcoholic beverages on your Visa card and thus increased the cost of your insurance? While this data sharing may be legal, it is neither helpful nor valuable to you.

So I'm sure I speak for all the authors in the Cutter IT Journal issue on the long tail (Vol. 20, No. 4) when I agree fully with Mr. Lineman's four points. When using customer history information to customize the service experience, enormous care must be taken on what data is captured, how the data is used to help the customer, and how the data is protected from misuse. The policy must be clear. And to avoid mistakes and even innocent surprises, the policy must be clear to the customer.

And yet I know that the authors in the issue would also join me in disagreeing with the blanket statement that implementing long tail strategies will "require an unprecedented combination of data collection and data processing power" as claimed by Mr. Lineman. In the series of articles in the issue, most of the authors make a distinction between long tail strategies on the one hand and hyperdifferentiation and resonance marketing strategies on the other. The long tail says stock all the digital goods you want and all the varieties and tuned versions of them you want; they can be sold easily online, the cost of an additional CD or DVD or new parameterization of a game is almost zero, and they don't spoil. Hyperdifferentiation says yes, even in the physical world you can make almost anything, but differentiation needs to be guided by resonance marketing. What does the customer really want? What is the marketplace currently offering, and where are the major gaps and unserved segments? What is missing? What can we offer that will represent a near monopoly in matching the preferences of some customer, will delight that customer, and will command premium prices?

Resonance marketing does not require getting too personal with the consumer, learning too much about him, and sharing with him explicitly or implicitly what you have learned. Resonance marketing does not require that you chase down the consumer with a product that you know he wants. Indeed, the biggest changes behind resonance marketing are:

• The consumer can find whatever she wants. There is enough information available, from rating sites and other consumer-driven information sites, to enable the consumer to locate whatever she wants, even in a product-rich hyperdifferentiated environment.

• These informational sources are no longer under the control of companies or their advertising agencies. Users on Ratebeer (http://ratebeer.com) say exactly what they think about the beers they drink. One of the top raters on Ratebeer describes Victory HopDevil by saying, "The hoppy bitterness stands out in a class of its own -- all I can say, try it if you have the chance." Other reviewers are less generous, saying things like, "Didn't enjoy this beer, too bitter for me, not a fan at all of this."

  Travelers are also particularly detailed and can be especially brutal; for instance, although the following hotel does quite well when it is not fully booked, the reviews during a large convention or conference can be quite detailed, overwhelmingly brutal, and damagingly convincing, as seen in this TripAdvisor review:

  I stayed at the [hotel] for three nights for a conference and had the worst experience I have ever had at a hotel. They put me on a smoking floor despite my confirmation listing a nonsmoking room. There was no shade on my window -- I reported this immediately, and every day I stayed there, and yet nothing was done to cover the shade. They offered me a breakfast voucher for this and said it would be delivered to my room; however, it never showed up and I had to go to the front desk to get it. Finally, the front desk staff were extremely rude and incompetent throughout my stay. I waited in line at one point at the front desk, and all the other guests in line were complaining about their experiences as well.

  Firms cannot control what their customers think or say, and they no longer determine what potential customers learn about their products.

Manufacturers don't chase the car buyer or the beer drinker in ways that invade privacy; they don't advertise to the driver or beer drinker how much they know about them; and they really can't do much to influence how their goods and services are seen in the marketplace other than by through the execution of their goods and services. They do use information to find out what to offer and to determine how to design their product and service offerings. Indeed, we might argue that it is the brewer or hotel operator, not the beer drinker or the guest himself, who no longer has any privacy!

When Steve Barnett led Toyota's Research and Strategy team that designed the US entry strategy for the Lexus, they did not go through the history of millions of Toyota owners. They assumed that Lexus would be bought by the segment of Mercedes owners who wanted spectacularly well-engineered cars at a price that signaled they were savvy enough to get quality for less than Mercedes prices. These consumers got their information from Consumer Reports (the 1981 equivalent of TripAdvisor), so Steve and his team reverse-engineered every automobile report that Consumer Reports had ever published. They designed a car to get the highest possible reviews from the source that their prospective buyers would consult. Inspired? Probably. Manipulative? Possibly. Invasion of privacy? Not at all.

Likewise, Snickers' remarkably successful repositioning as food is a result of Barnett's ethnographic observations. He studied people who picked up a Snickers bar but then bought something else, determined that they bought something like a granola bar instead, and then helped Snickers reposition itself as food rather than merely candy. This later led to Snickers' introduction of its own line of energy bars. No candy-buyer had to participate in Barnett's study, and no data was captured that allowed Snickers to identify individual shoppers or their responses. Once again, inspired? Probably. Manipulative? Possibly. Invasion of privacy? Not at all.

In summary:

• Resonance marketing and hyperdifferentiation are not the same as customization and tailoring. They do not always imply use of private and privileged information to design custom offerings and do not always raise sensitive issues of privacy.

• Very few manufacturers base their resonance marketing strategies on the use of privileged information. The possibilities for abuse of information in these market segments are quite limited.

• While there is little data to suggest that collaborative filtering at Amazon.com works well enough to constitute invasion of privacy, many high-end service providers, especially in travel and financial services, do use customer service information. The possibilities to be abusive are numerous, the possibilities for loss of control are more numerous still, and the possibilities to be seen as offensively intrusive are almost without limit. Google's recent experience suggests that abuses, even potential abuses, will become very visible very quickly. Mr. Lineman's points here cannot be stressed enough.

Our thanks to David Lineman for his timely and well-considered comments.

NOTES

¹See, for example, www.washingtonpost.com/wp-dyn/content/article/ 2007/06/09/AR2007060900840.html.

²Citi and Travelers merged in 1998, and Citibank Visa and Travelers Life Insurance are, technically, the same company.

REFERENCE

1. Privacy International. "A Race to the Bottom: Privacy Ranking of Internet Service Companies." 9 June 2007 (www.privacyinternational.org/article.shtml?cmd%5B347%5D=x-347-553961).

ABOUT THE AUTHOR