Using Regression Analysis for Web Metrics

So like I mentioned in a previous post, I wanted to try my hand at using some of the techniques that the authors used in “The Wages of Wins” to look at web traffic variables. At the time I had asked if anyone had any decent tools to use for this kind of analysis. Low and behold, the answer was already on my desktop. Microsoft Excel actually has the capabilities, you just need to install them. In Excel, go to Tools and select Add-Ins. This will give you some nifty analytical tools to use, I chose Analysis Toolpak. And then it shows up in the dropdown under tools as Data Analysis.

I only did a couple of simple studies. Of course, I’ll have to do a lot more tests with a little more sophisication and thought before declaring any revelations. That being said, I did a simple one where I looked at the correlation with web traffic (visits) to online commerce on one of our sites at Lenovo. I chose the web traffic to the commerce section specifically as the piece to look at. This is all to test the premise that if you get more traffic to your site, you should see more revenue. Right? Well, based on my simple study I got a correlation of .393, which isn’t all that high. Again, this was a simple one where i took weekly visits and weekly revenue for 2006. I decided to try it again with other parts of the site to see if one piece correlated better. The homepage actually came out a little ahead with .42.

What does this all mean to me? It says to me that sheer volume is not exactly the answer to driving revenue. There are obviously other factors. I’ve always said that getting more traffic is not the answer, but getting the “right” kind of traffic is. What is the “right” traffic? Not sure yet. I want to test different things like kinds of referrals, pricing, sales promotions, marketing spend, seasonality, etc to see if I can find the magic formula. Thought I’d share what I was working on, and if anyone else has done similar work please share.


7 thoughts on “Using Regression Analysis for Web Metrics

  1. Hey Jim,

    Although direct revenue is a major goal (who doesen’t like to make money!!) I think there are other interesting a Web like Lenovo’s has to focus on.

    For once support is quite important. Maybe determining what is a success in terms of support is not an easy task, at least in terms of metrics. Having people seeking for support (and finding what they need) is a major achievement.

    On the other hand a lot of people just want to find something about what you offer and at what prices. They might not have an immediate desire to buy, but having them finding that particular feature/product/price/whatever they are after is also a good objective.

    Since Lenovo’s web has many “targets” in terms of traffic, it also haves many different kinds of “right traffic”. Not having a “only-purchase-centered” site vistits to purchase numbers will allways be lower than those of other sites, but this is not necessaily bad.

    The god thing about having a complex site is that you have a lot of things going on and a lot of ways to be successfull in terms of user satisfaction. Neat challenge ain’t it?

    All in all I couldn’t agree more than more traffic doesen’t necesarilly mean “right” traffic… But lenovo has a lot of different “right traffics”.



  2. One of the best locations I’ve come across lately!!! Definately a permanent bookmark! Please also visit my site:

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s