Segmenting Visitors – without PII?

Recently, I attended a webinar where one of the discussion points was segmenting users and providing custom content to each group.  To me, this  makes perfect sense – there is a wealth of information at your (web analytics) fingertips, why not use it to understand your users and ultimately improve the customer experience?

Yet it did not make sense to everyone – during the Q&A, someone asked how it was possible to segment users without using PII (Personally Identifiable Information), such as name and account information.

Knowing who the person browsing your site is may be useful – but what he is doing on your site is also valuable.  Using the click-stream data, such as what pages he sees and what keywords he searches on, you can understand what he is doing on your site, and perhaps determine his intent.

Of course, integrating the two data sets make them more powerful (as I wrote in an earlier post), but the click-stream data alone can be very useful.  So, what kind of information do you have available to you in the click-stream data that allows you to segment users in a meaningful way?

First, you could segment users based on what users are doing right now:

• Users who entered via an affiliate
• Users who logged into an existing account
• Users who saved an item to their cart
• Users who visited the FAQ area

In a more advanced implementation, you could segment users based on what they are doing right now but also what they have done in the past.

• Users who have logged in in the past
• Users who have shown interest in a particular area of the site (say, searching in books versus movies or those who visit their accounts page but don’t do much browsing)
• Users who have visited multiple times

All this without knowing who the user is!

So, what next?  You could look at whether the users you’ve grouped together behave differently than other users on the site, and how you could optimize the site experience for that group of users.  You could tie the click-stream data back to an individual user’s data for a more complete picture of the user.

Advertisement