Deprecated: Function set_magic_quotes_runtime() is deprecated in /home/mwexler/public_html/tp/textpattern/lib/txplib_db.php on line 14
The Net Takeaway: Analyze people, not sites


Danny Flamberg's Blog
Danny has been marketing for a while, and his articles and work reflect great understanding of data driven marketing.

Eric Peterson the Demystifier
Eric gets metrics, analytics, interactive, and the real world. His advice is worth taking...

Geeking with Greg
Greg Linden created Amazon's recommendation system, so imagine what can write about...

Ned Batchelder's Blog
Ned just finds and writes interesting things. I don't know how he does it.

R at LoyaltyMatrix
Jim Porzak tells of his real-life use of R for marketing analysis.







Analyze people, not sites · 10/06/2009 02:48 PM, Analysis

I am continually amazed at how ignorant current tools are of the people actually driving the behaviors we are looking at.

Recently, a quiet buzz rose around an analysis of Twitter usage by a fellow running a pretty cool company called RJMetrics (yes, the name sucks. Yes, the initials of the founders are R and J). At Twitter Data Analysis: An Investor’s Perspective at TechCrunch, Robert J Moore examines Twitter usage in a couple of different ways. I wasn’t all that impressed with most of the analysis; it was pretty basic stuff.

But one of the ways I was most excited to see him highlight is the Cohort analysis. This is one of the most simple segmentations you can do: just take everyone who, say, did their first purchase in July 2009 (we’ll call this Time 0), and see what else they did over time (each month, say Time 1, Time 2, etc.). Do the same for everyone who did their first purchase in, say, Oct 2009. Then line everyone up on a graph so that everyone’s Time 0 is at the left, and then Time 1, etc. This lets you compare behaviors of clumps of people to see if their “lifecycle” is consistent.

There is more detail at A VC’s blog The Cohort Analysis as well as more detail at RJMetric’s blog post Cohort Analysis in RJMetrics which is a recommended read.

But at the end of the day, beyond the value of this specific analysis, I admire that they are examining “people who” and then looking at “what they did”. So many analytic tools are stuck on “what they did” and forget the people part. So, you can get lists of most popular pages, but not who visited them. You can get lists of most often sold products… but can’t do anything to understand who bought them. And I don’t mean just getting a list of cookies; I mean actually having a group of people and comparing their behaviors to a different group of people.

Here are a couple of simple analyses; see if your web analytic tool can do them:

As you can see, almost every question starts with a segment. (BTW, this is kind of unfair; even if I dropped the “people who” part, most tools can’t answer the questions above. That, too, is sad.) But the current tools have all sorts of limitations that prevent us from looking at “people who”:

This is really sad. I’ve had the luck over the past few weeks to use a bunch of different tools, and I am shocked at how poorly they let me examine my business. And yes, if you read What Web Analytics is Missing which I wrote over a year ago, you’d see that there has been 0 progress.

So, try doing a Cohort analysis of your business. Try looking at how groups differ, or are similar. And try to put people first in your analyses.

At the end of the day, I am trying to get people to buy things. The things don’t sell themselves… but looking at most web analytic offerings, that’s what they want me to think. After all, that’s all they are measuring.

* * *


  Textile Help
Please note that your email will be obfuscated via entities, so its ok to put a real one if you feel like it...

powered by Textpattern 4.0.4 (r1956)