In the BagMerging online and offline data is now the key to unlocking lifetime customer value. But where do you begin?
By Kwong Sai Continued from Page 1 In terms of the online data, you can identify customers through IP addresses, cookies, or even login authentication if available. However, if you only use an IP address as the default for identifying customers, you will be unable to link it with anything — the only value you will gain is the ability to group all the session information for that particular user to that specific session. If that same user were to come back for the next five days, each subsequent visit would be recorded as a separate user. This method of user identification does not allow you to identify unique users, and in essence, only gives you information based on visits. Without identifying unique visitors, it will be impossible to merge the two different sets of data. Capturing Data by Unique UserIf cookies and user id logins are used to define a user, you can identify unique visitors via the login or the cookie. The historical data of that specific user can then be captured regardless of when or how often they come back to the site. When the data of the unique user can be grouped accurately, you gain the opportunity to integrate that data with the offline data. There are some drawbacks to this method as well. For example, if no logins are required at your site, how do you identify the visitor? What happens when part of the site requires login, and others do not? As for cookies, you could also debate where cookies should be dropped. Should it be on the home page, or with a registration section of the site? What if you don't use cookies at all? These setbacks can be addressed, but the solution will vary from company to company. In some cases, a combination of all three methods will address the obstacles, but in the end, it comes down to capturing the data by unique user. When you have grouped and standardized the format of the online data, it really becomes like any other data file: The primary key has to be identified between the two sets of data. The most critical aspect of merging the information is selecting what the primary key should be. The key could be a name and address combination, account ids, credit card numbers, cookies, or even social security numbers. The key itself will be determined by the data structure, but whatever the key is, it has to be a unique relationship that will definitively link the offline and online data for that one particular visitor. Is this methodology 100 percent accurate? No; there will be cases where you will not be able to match all the information. But it will provide enough of a match to gain the insight you need. You must also contend with the challenge of maintaining and updating your site. Some sites are very static but others are extremely dynamic and may require changes on a daily basis. However, logic can be developed to automatically categorize how these changes should be recorded in the infrastructure of the data. For example, if you had a sports section on you site with the path www.acme.com/news/sports, you could add or delete as many sports pages as you like but the program would always recognize it as a sports article as long as the pathway is used. However, the simple fact of the matter is that you will always need a human factor to set the infrastructure because no computer program can ever anticipate what value the data has to your company. Using our previous example, suppose you created a special section called www.acme.com/news/winterolympics2002 to cover the Winter Olympic Games. The path doesn't fit the sports article criteria so the article would not be automatically captured under "sports article" by the program. Rather, someone would have to identify what path within www.acme.com/news/winterolympics2002 would fall under sports article and be designated as such. Data vs. KnowledgeOn last thing to keep in mind is that building the correct data infrastructure is a critical first step in enabling your organization to unlock the value stored in your data. However, it is not the only step: Within the mounds of information your company collects are vital nuggets of data that are extremely important in helping increase revenues and reduce costs. In the past, companies have struggled to make decisions based on the lack of data. But, in the current environment, more and more companies are struggling to overcome "information paralysis" — there is so much data available that it's hard to understand what is relevant. Thus, you should keep in mind that not every data element is important in addressing your objectives. There are many site analysis tools that provide hundreds of reports with reams of data, but what value or knowledge do they really provide? You need to have the right people who not only understand the data elements being collected, but how those data elements map to business objectives. Data in itself does not provide knowledge. It's the analysis of the data that ultimately leads to knowledge. These two factors are intertwined; whether data ultimately provides value to your organization depends on your success on both sides of the equation. Kwong Sai [ksai@agency.com] is vice president of performance analytics at Agency.com Kwong works with clients to identify the metrics and measurements to use when examining the value of their Web sites. RESOURCESIntelligentCRM.com: Intelligent Enterprise's online CRM community "Branding the Customer Experience," Oct. 24, 2001: www.intelligententerprise.com/011024/416feat1_1.jhtml "Eight Steps to Customer Identification," July 23, 2001: www.intelligententerprise.com/010723/411feat2_1.jhtml "Mining a Demographic Mother Lode," June 13, 2001: www.intelligententerprise.com/010613/decision1_1.jhtml "The Model Customer," Jan. 30, 2001: www.intelligententerprise.com/010130/feat2_1.jhtml
|
 |
Most Popular This Week
IE Weekly Newsletter
Subscribe to the newsletter
| |
| |||||||||||||||||||||||||||||||
