Use Google To Your Advantage

Looking for a very simple way to segment online visitors into actionable segments?

Take any visitor who arrived at your website within the past thirty days following a paid/natural search via Google.

Store this fact as a Yes/No attribute on your database.

Now analyze your catalog, e-mail, online marketing, and retail marketing activities against your usual set of analysis segments --- overlaying the "Google" factor on each segment.

I promise you will learn something about your customers, and your marketing activities, that you didn't previously know.

Go ahead and expand upon this idea using other referring URLs identified in your clickstream data. Again, I promise you'll learn something actionable about your customers you didn't previously know.

Some folks are really doing sophisticated stuff with this information. We can all benefit by starting with such a simple attribute, a yes/no indicator, and then work toward complexity from there.

Multichannel Forensics And Information Value

Friend of MineThatData Alan Rimm-Kaufman shares that Google will retain about eighteen months of clickstream data, going forward.

Multichannel Retail faces similar challenges, when looking at the value of customer information within the context of multichannel forensics.

Customer information ages differently in multichannel retail. While the relationships are different for each business, the following example helps illustrate the point.

• Catalog: A purchase twelve months ago is worth about 1/2 of what a purchase that occurred today is worth.
  
• E-Commerce: A purchase twelve months ago is worth about 1/4th of what a purchase that occurred today is worth.
• Retail: A purchase twelve months ago is worth about 1/8th of what a purchase that occurred today is worth.
• Clickstream: A visit twelve months ago is worth about 1/32nd of what a visit that occurred today is worth.

This topic becomes important when evaluating actual customer behavior. Most multichannel retailers would consider a customer who purchased via catalog twenty-four months ago, and purchased online today, to be a "multichannel" customer.

The reality is that this customer is heavily skewed toward the online channel.

Multichannel marketers have an opportunity to run a regression-style analysis, to determine the appropriate weight to use with older purchase information. The weights determine how customers are segmented, and consequently, determine how the multichannel retailer markets to the customer.

Multichannel CEOs and CMOs: On Friday morning, talk to your analytics staff about segmenting customers on the basis of the value of older purchase information. Have your staff apply a new technique that ultimately mimics the time-honored system of "RFM --- Recency, Frequency and Monetary".