How to integrate clickstream data and catalog strategy
Might 2008 be the year that catalogers embrace online customer behavior? Our customers continue to shift behavior to the online channel. Each time a customer visits the Web site of a multichannel cataloger, the customer shares important facts that the cataloger records in a clickstream database.
However, the data is frequently summarized at a “detail” level, meaning each page or action is recorded in the database. It is uncommon for catalogers to translate these facts into summarized fields in the corporate customer warehouse. It is also uncommon for catalogers to link this information to customer name/address.
The level of summarization doesn't have to be complex. For instance, the cataloger can create summary fields that list the number of months since the last action. If the referring URL is from a paid search campaign in Google, the cataloger can simply record the most recent date this action occurred. Similarly, other variables are created for natural search, shopping comparison sites, affiliates, portal and general banner advertising, important blogs, and other sites that are important to categorize. In addition, it is important to record the date of the most recent visit, most recent visit to key landing pages, most recent item in a shopping cart, and to count the number of lifetime Web site visits.
Next, the cataloger uses this information as part of a RFM or modeling strategy. Maybe the cataloger is having a difficult time reactivating online buyers who haven't purchased in more than a year. Simply overlay the date of the last Web site visit over the RFM segmentation, and measure whether this type of activity increases or decreases multichannel catalog response.
If the cataloger learns that this information is valuable, it becomes important to understand which activity is important. This is where all of the other activity recorded in the customer database come into play. Did visiting a particular landing page in the past cause the customer to respond to a subsequent catalog? Do customers who arrived through a shopping comparison site respond at a lower rate? Do catalogs interact with online marketing strategies to increase response? Do catalogs and blogs interact to grow customer loyalty? All of these possibilities can be measured if the right summary variables are captured in a customer database.
This information becomes critically important when evaluating the difference between mail and holdout groups. For instance, if a segment of customers with a good RFM score visits the Web site often, the customer may not require a catalog mailing to visit the site. This hypothesis can be easily measured via mail/holdout samples, comparing response at a segment level with summarized clickstream data overlaying traditional categorizations.
In 2008, the “heavy lifting” necessary to properly summarize clickstream data should happen. This summarized information complements the catalog and e-mail contact strategy. Instead of relying upon matchback algorithms, the cataloger uses in-house clickstream data to better understand the interaction between catalog marketing and online customer behavior.
Kevin Hillstrom is president of MineThatData. Reach him email@example.com.