If you currently don't know much about your customers, then you might want to think about a collaborative-filtering approach since it allows you to use minimal information to infer that a customer is interested in a particular subject matter. However, collaborative filtering is a complicated process, and it normally takes an experienced hand to program such a system. Collaborative filtering is also a relatively expensive technique. Luckily, software vendors can do most of the work for you and thereby cut costs. Using these software packages can be a weeklong project or it can take several months.

As an example, AT&T WorldNet uses software that tracks customers coming to its Web site's entertainment and arts services. The program uses artificial intelligence to decipher visitors' preferences without them knowing it, and continually makes suggestions to the visitors based on their current preferences. For example, a user who views information on romance may suddenly develop a taste for science fiction a month later. The customization tool quickly "learns" those traits and points the viewer toward additional science fiction information that may be of interest to him.

Additionally, at online bookseller, referencing tools allow the site to automatically suggest books of similar subjects to the ones a customer is already looking at. Therefore, the possibility of impulse sales increases, and on the Web, impulse sales are important revenue boosters.

A slightly different customization approach is employed by online grocer Peapod. Their personalization tools are designed to get to know you so well that they do your food shopping for you — a very personal task, indeed. The site's system learns the items on your personal shopping list, then workers scour the aisles of real grocery stores to find the best price for your groceries, selecting products that meet your needs and delivering them to your home.

Furthermore, the site allows shoppers to store information on their favorite foods, favorite butcher, and so on, enabling the site to identify and deliver meat, cheese and produce just the way you like them continuously over time.

Finally, the site offers specialized features, such as electronic coupons and sales advertisements, through the use of custom-designed software. These features, when combined, contribute to the creation of an atmosphere of trust and increase the likelihood of increased purchases. If you're leaning toward a simpler strategy, rules-based programs may be a dollar better spent when starting a personalization plan.

Issues to Consider

Planning is important when dealing with personalization on the Web. Just as you wouldn't want to waste a retail customer's time with irrelevant products and services, you won't want to waste online viewers' time with time-consuming and ineffective personalization features and functions.

You'll also want to remember that personalization, if it is right for your company, should become an integral feature of your existing sales and marketing plan. This will help you discover additional important opportunities for reaching your target audience.

In order to identify whether your personalization plan is ready for the Web, ask yourself these questions:

1. Does the Internet normally provide the type of people that I need to reach?

2. What personalized products and services do my customers typically want offline that I could offer online?

3. What do I want my Web site to accomplish, in terms of marketing?

4. What standards do I want my plan to meet?

5. How can I measure the results of my online personalization plan?

6. Will I get more hits by marketing my site on a one-to-one basis?

7. Will I get more hits by offering information that is of interest to individual visitors on a regular basis?

8. What customizable attraction(s) do I have to offer?

9. What technology is available to help me accomplish this feat?

10. Can I do this in a cost-effective method, or will it require too many added costs?

Putting Online Personalization Into Action

On the Web, it's important to set up your personalization tools while keeping the mindset of a typical salesperson. As most salespeople know, helpful information can be gained from such simple information as the user's ZIP code, which allows you to localize information and product offerings that may be pertinent to the user's local area. For example, a sporting goods retailer may want to concentrate a South Florida user's offerings on water sports equipment, while opting for skis and snowboards for a user in Maine in December. However, a good salesperson — and a good personalized Web site — always delves deeper into a person's personality.

"It's easy to say something like, 'Anybody who likes Eric Clapton is going to like Johnny Lange.' But that's based on an arbitrary assumption," warns Porter. "You have to look at the whole database."

Also, keep in mind your sales approach when setting up your program. For example, some salespeople sell a complex product on its features and introduce the price late in the sales process. Another company selling the exact same product might compete on price, so they want to get that up front to pull the prospect in, then talk about how many customers they have and give a few features that separate them from the competition. Your Web site should be set up to reflect your personal style.
It is also important to remember that personalized Web sites should strive to retain serendipity. One of the big mistakes made by many sites is that they don't try to emulate the "on-the-fly" thinking that live salespeople do. For example, a salesperson can use the knowledge they gain to offer up other products and services that the customer might not even know that they want yet.

But how do e-tailers actually set up a personalized Web site and gain information in the first place? Most turn to help from one of the many software developers that specialize in the subject. But no matter what technology you choose, technically receiving the information usually requires one of at least two basic features: log-ins or "cookies."

Log-ins are simply input fields, user names and passwords that allow users to register with your site. Once they are registered, your site can then personalize information on the user and store it in a database. Then, any time a user visits, the software immediately recognizes the user and shows them relevant content and merchandise. Log-ins are simple and inexpensive, but often disrupt the seamless nature of your personalization strategy.

The second way to track identities online is through the use of cookies, which are a little more high-tech and a little more controversial. Cookies are files sent to the user's computer that secretly live on the customer's hard drive. These little buggers are often maligned because many question their moral standing in regard to privacy issues. But cookies only know as much about a user as the user tells them. What's more, this technology adds a seamless element to your personalization. And the server that sends out a cookie is the only thing that can read it, so security is not an issue. The cookie stores information on the user, and whenever a person visits your site, the artificial intelligence and learning agents in the cookies recognize the user and store useful information as he or she goes. Then, the information that the cookie produces can be analyzed by the system to reflect which type of personalization strategy (rules-based filtering or collaborative filtering) that you choose.

What Your Personalized Site Should Include

While there are no set standards for personalization, and everyone finds different ways to approach the process, there are a few things that your strategy should include to make it effective. First, Web sites should start people off with a generic version of their site and allow users to customize it gradually as they see fit. Second, they should watch what users are doing and actively recommend personalized ideas. Third, they should combine personalized services with a generalized segment where readers can keep in touch with the rest of the community, read about new ideas, and get recommendations from experts.

When Personalization Works … and When It Doesn't

Typically, most offline offerings can be personalized and customized in one way or another to increase the value of a shopper's experience. That could include customized service, colors, sizes, shapes, packaging, delivery and so on.

It is important to realize, however, that some products might not lend themselves well to the personalized online approach. For example, books and CDs, which are the best-selling items on the Web, are perfectly suited products for personalization because they are usually purchased as people are searching for a specific title, subject or author, or, in the case of music, a specific title, artist or genre. However, products such as calendars might not be a good fit, since people generally like to browse through them and usually do not know exactly what they want before they do their browsing. Although there are cases where someone may be looking for a specific calendar, in general, the only information people are certain of when they buy a calendar is the year. In this case, personalization can only take you so far.

It is also important to remember that creating a successful personalization plan is challenging, and you should not take the effort lightly. Just as any advertising plan relies on repetition for success, so, too, does personalization. When developing your personalization plan, it is important to implement features that will keep people coming back on a regular basis. You will need them to come back regularly in order to track their habits.

Web portals and news organizations have adopted this strategy very well. Excite, Yahoo! and all have designed strategies to keep regular visitors coming back day-to-day and, in some case, more than once a day. How? Although the basis for their revenues is advertising, and their main product is their search engine, these sites have turned to personalized news and information as one of their main draws for traffic. The information they supply isn't just daily news; instead, it is targeted, constantly updated news. And when a visitor registers with any of the three portals, he or she is able to personalize the page they view each time they log on. For example, the regular Excite home page is general in content, including top news articles, general links, and so on. However, the personalized My Excite home page asks users to select specific types of news, such as technology, science, sports and business articles. It also allows users to localize their news to reflect their home state or select specific sports teams that the site will track. Excite does the same with personalized television listings and weather, based on the user's ZIP code.

The benefit of personalization for each of these sites is that these users will return regularly for their personal updates and then use the site's search engine as opposed to one of the competitors'. With the added visits, the sites' hits rise dramatically and attract additional advertising revenues.

The strategy has paid off handsomely for these companies, as it can for yours if you do the proper research and implement the concept correctly.

Personalization works best when you either have a large number of different types of products and services to sell, or your customers come from many different walks of life. For example, personalization greatly aids a bookstore or a music store because their customers are both young and old, male and female, and they all have different tastes in genres, subjects and authors. Meanwhile, a company that sells highly specialized items to a distinctive audience may find the effort and expense of personalization less beneficial.

Take this test to find out if you should be using personalization:

1. Do you offer a large selection of products or services?
Yes No

2. Do you have multiple target audiences with diverse demographic profiles?
Yes No

3. Do your customers have multiple buying motives, and do you know what they are?
Yes No

4. Can you satisfy those motives with different approaches?
Yes No

5. Do different customers use your products for different purposes?
Yes No

6. Can you adapt your products to address those uses?
Yes No

7. Do you currently have — or could you develop — a means for gathering and analyzing personalized information about your customers?
Yes No

8. Do you currently have — or could you develop — a way to offer different products and services to different customer groups?
Yes No

9. Do you have a budget for personalizing your offerings?
Yes No

If you answered yes to most or all of the above questions, your company may be good candidate to implement a personalization strategy