Web Analytics Are Four-Dimensional Now (Or How Many Do You See?)

Posted by Akin Arikan on May 15, 2007

3 comments | 2799 reads

How many dimensions do you see when you do web analytics for your business? Three or four? Or more? Don't stop short!

Back in the 90s web analytics were rather 1 dimensional, namely just traffic stats. Then web analytics became two dimensional. Do you remember "quadrant analysis"? Back then it seemed as if web pages could be described as being divided into four quadrants and web analysts wanted to know where to place links for greatest clickthroughs. A third dimension was added to web analytics, namely depth. We started referring to it as clickstream analysis. Funnel visualizations and tree-shaped visualizations of clickstream paths became available in web analytics solutions such as Omniture, Webtrends, and NetInsight, the solution provided by my employer, Unica.

Web analytics matured within these three dimensions when in 2005 the discussion switched over to key performance indicators (KPIs). You couldn't listen to a presentation without hearing the advice that KPIs were the single best method for preventing "analysis paralysis".

It is 2007 now, and web analytics have become four dimensional. What is the fourth dimension? The customer! We came to realize that web analytics is not just about the site after all, but it is really about the prospects and customers that visit our sites. We don't just want to optimize our web pages, but we want to optimize the interaction with prospects and customers. Oh yes, web designers have known this for a long while. But web analysts still focused on KPIs, paths, funnels, yada, yada. Now, the tides have shifted. Behavioral targeting is in site (ha ha). Coremetrics’ LIVE profiles have been around for a while already. But did you also notice Omniture’s acquisition of TouchClarity, or Unica’s Affinium Interact solution all of which focus on personalizing the site’s interaction with it's visitors based on previous behavior?

Speaking of previous behavior, it goes without saying that the 4th dimension needs to include the customer’s previous behavior in the offline space as much as possible. Otherwise, how could we hope to achieve customer-centricity and get the personalization right. Do you think?

Akin Arikan

Akin Arikan, the author of Multichannel Marketing: Metrics and Methods for On and Offline Success, is a director of product strategy for IBM's web analytics and interactive marketing solutions. Akin has been working with analytics practitioners since 1999.

About Akin Arikan |
Connect:

Categories:

2799 reads
Print
Email

3 comments »

Jessica

May 15, 2007

dimensions?

So when you say 4 dimensions you must be thinking as in 3-d / 4-d, right? Not as in multi-dimensional analysis / OLAP? Because the latter could be umpteen-dimensional, say Products x Customers x Referrers x campaign channel x bla bla.

Edwin F. Setzpfand

May 15, 2007

Counting the dimensions -- do dimensions count?

I am afraid Jessica is missing the point and Arikan's valuable and clear message about inclusion of the customer's (prospect's) previous behaviour has nothing to do with 3D, 4D or whatever you want to choose from that arena.

It's about customer behaviour, like he wrote, which in a way can be seen as multidimensional in itself.

However, the number of dimensions is not that relevant at all .. as long as our speaking of 4D-analytics doesn't warp the customer out of our B2B-space! ;)

Let 's keep our feet on the ground and speak language that convince customers that we know what we're talking about.

Akin Arikan

May 17, 2007

Counting the dimensions -- do dimensions count?

I think Edwin is right on. Both views of "dimensions" have merrits. They are just different discussions. This essay was really about the customer behavior in all its facets and online+offline as Edwin underlined.

Jessica is raising a very good point too thinking of OLAP style slicing and dicing of reports across "umpteen" dimensions. Such multi-dimensional segmentation of web analytics behavior is very useful to web analysts in order to find actionable insight. In fact my favorite web analytics solution is very good at that kind of right-click, drill-anywhere, slice by any dimension type analysis. 8-) .

But I guess Edwin and I are saying that you could slice & dice all day long -- if you didn't include the customer dimension in your point of view -- how good can the nuggets be that you will find in the end?

Right on Edwin!
Akin

Post new comment

Your name: *

E-mail: *

The content of this field is kept private and will not be shown publicly.

Homepage:

Subject:

Comment: *

Input format

Filtered HTML

Allowed HTML tags: <a> <img> <hr> <table> <tr> <td> <th><ul> <ol> <li> </li><blockquote> <colspan> <rowspan>
Lines and paragraphs break automatically.
Web page addresses and e-mail addresses turn into links automatically.
Images can be added to this post.

Post

Lines and paragraphs break automatically.
Web page addresses and e-mail addresses turn into links automatically.
Images can be added to this post.

More information about formatting options

Notify me when new comments are posted

All comments

Replies to my comment

CAPTCHA

No spam permitted! Moderator reviews ALL content before publication to ensure compliance with the CustomerThink terms of use.

To block automated spam submissions, please answer this question.

What code is in the image?: *

Enter the characters shown in the image.

MarketPlace

Drive customer loyalty, empower support teams, and reduce costs. Get social.

[Feb 22] Guest speakers from Forrester Research, Allscripts, and CustomerThink will discuss market trends and research on social customer service strategies, as well as proven tactics from the trenches. Join the live webcast on Feb 22 at 10am Pacific (1pm EST).

Global Customer Experience Management (CEM) Certification Program

[March 13-14, Paris] An internationally recognized program with proven track record of success - being run for 33 times in 13 cities with attendees from 50 countries, the program is developed based on the U.S. patent-pending Branded CEM Method which aims to drive customer loyalty and brand differentiation with quantifiable business results. Limited offer: USD300 early bird discount.

10 Steps to a Single Customer View

Linking customer data across department databases and business units improves business intelligence, customer profiling, and customer management. This paper outlines 10 steps to improve the quality of customer contact data, including physical mail, email, and telephone information.