Personalizing Your Service
By Wendy M. Grossman, Tue Feb 05 00:00:00 GMT 2002

What content needs to be delivered to whom and under what circumstances?


Every couple of years there’s a new key to success. One of the current one seems to be personalization.

As long ago as mid-2000, Ovum published a report on wireless portals (for which it projected $42 billion in revenues in 2005) saying that personalization was vital to their success. More recently, in May 2001 Cyber Dialogue conducted a survey on behalf of the Personalization Consortium that found that 87 percent of Web users are annoyed when a site asks for the same information more than once and that users of personalized services spent more and were more willing to pay for subscriptions.

However, the survey also noted that upwards of 80 percent of users think privacy policies are important, and have at times refused to provide information if they were not sure how it would be used. Also last year, Cahners In-Stat Group pinned down privacy issues as the main hindrance to the adoption of personalization among businesses.

So, is personalization different from other apparently "hot, new" ideas that cost a lot of money and that ultimately didn’t pan out? Or is it actually an important approach that will help make the wireless world succeed and help tie it into the existing channels of delivery (physical world, Web, telephone, post) that we already have?

Dueling visions


The disparity between what users experience as personal and companies think is personal becomes much clearer when you listen to how companies propose to use personalization and contrast it with what users mean by the same term.

Users tend to think in functional terms. To mobile phone owners, it may mean that the memory holds the numbers of family and friends, the case is a favorite color, and the ring tone is unique. To consumers outside of the wireless context, it might mean that the insurance company’s Web site know their history, or that they talk to the same customer service representative every time. Marketers, however, tend to dream in terms of being able to create advertising so tightly targeted that the recipients think of it as information someone miraculously knew we’d want. Larger companies see it as part of customer relationship management, another buzzphrase.

This disparity of images is why it’s so hard to pin down what analysts mean when they say that personalization is the key to success. They mean SMS discount coupons. You want your phone to tell you your train is halted by a security alert and offer you alternative routes. Yet that proposition involves not only access to multiple complex databases but the kind of predictive intelligence that computers have always been supremely bad at. Amazon.com’s recommendations system is one of the most widely touted examples of such intelligence, and yet it constantly fails. Partly, this is because it can’t distinguish between gifts and items bought for personal use. Partly, the computer simply doesn’t understand how humans explore their culture.

An expensive gamble


In fact, personalization won’t come cheap. The research company Datamonitor predicts that global investment in personalization technologies will grow from $500 million in 2001 to $2.1 billion in 2006, with North America accounting for 67 percent of the market and Europe 25 percent. The two biggest areas for personalization is said to be financial services and retail industry.

While the report itself focused primarily on the Web, Datamonitor says that anyone adopting personalization technologies is likely to do so over all channels. This is logical since a company’s goal is an ongoing, profitable relationship with customers that’s independent of the communications method or device they choose at any one time.

The question is why it’s so expensive. Evan Kirchheimer, the analyst who wrote the Datamonitor report, says that, "In order to do personalization right across all channels – Web, personal delivery, telephone, shop – really implies a great degree of data integration and clean data about your customers." Adding to the cost is the need to integrate marketing, products, and sales databases. This is especially expensive, he notes, in Europe, where ensuring that customers have given permission is vital for compliance with privacy laws and therefore will have to be built into software.

Other problems exist with one type of personalization a lot of retailers would like to try: individualized pricing. When Amazon.com briefly tried offering discounts on carefully selected items to new visitors it was hoping to entice into buying, existing customers were enraged. It’s a great example of what happens when a new-tech world chooses the wrong older world for its analogies. Amazon.com thought it could behave like a magazine trial subscription. But what it had built (by design) was an online community. In such a case, differential pricing will only work if the company announces its plan up-front, so customers don’t feel deceived.

The adoption of personalization technologies has slowed rapidly because of the economic downturn. With 3G still on the horizon and the bills for new services mounting, marketers interested in personalization in the wireless world are turning to the simpler technology of SMS marketing campaigns. Such campaigns are, as Forrester analyst Michele de Lusenet says, cheap, and they currently show high response rates (as much as 11 percent), perhaps because they’re a novelty.

De Lusenet, who wrote Forrester’s December 2001 report on SMS marketing, believes it is possible for these campaigns to grow without becoming intrusive. It seems impossible: as the volumes grow, it’s not going to matter how personal the marketer pretends the communication is, people will adopt schemes to block messages – a far more potent version of personalization these days – from all but their friends.

The importance of context


"But it has to get funded somehow," says Mary Ann O’Loughlin, an analyst at Ovum, author of several white papers on wireless marketing. "So it’s about making user desires and marketing meet." For O’Loughlin, however, the key to personalization is context: services need to understand what device you’re using and what (relative) location you’re at. "If you have to personalize your Web site, and then your wireless Web site, and then TV access, it’s far too hard work for the user. It’s not a trivial problem technically." This, she believes, is why estimates of the investment needed are so high. While dealing with multiple devices is relatively simple, managing a database and adding in some intelligence and anticipation of a user’s desires is again not trivial.

O’Loughlin also notes that a lot of what’s holding back personalized services is that users have to do the personalizing , and they simply aren’t motivated to do it. Or, as John Brimacombe, CEO of the Web/WAP/interactive TV games company nGame, puts it, "Laying out menus is not a mass-market endeavor." Right now, he says, "Personalization is all about menu design and opting into various types of content streams." He believes that this approach faces an uncertain payback.

For one thing, the wireless world’s current ideas about personalization are reminiscent of the wired Web’s early ideas about the same thing: portals. Even the example de Lusenet uses of personalization success fits into this category: the Financial Times’s branded MVNO service, a logoed phone bundled with an FT mobile subscription that the user can customize as desired. The idea behind Web portals was that you would have a single-page gateway into the Web with everything laid out as you liked. But most users already have a desktop on their own machine; why do they need another?

Truly personal


For another thing, many attempts at personalization are no more personal than mail-merge. What’s personal about an SMS message offering you the chance to ask the Bridget Jones server for dating tips? On the other hand, when they try harder, they fail. One of Brimacombe’s favorite failures is Microsoft Word 2000’s attempt to adapt to user patterns by showing you only what you’ve used recently. But some of the most important functions may be ones you use only rarely – and you can’t find them.

To Brimacombe, what will make the wireless world truly personal is Java programs. Once you can get users past the hump of buying programs they can use on their phones (probably stored on a server they reach as needed), he believes phones will become much more personal. "You have ring tones, logos, covers...we feel that putting on Java objects is the next step in personalizing your device." Such a view is much more about adding personally chosen capabilities to a phone than it is about interacting with marketers.

Wendy M. Grossman is a freelance writer based in London, and author of net.wars. Her new book, From Anarchy to Power: The Net Comes of Age is out.