Using Six Sigma & Statistics To Increase Adsense Revenue

Actually since most publishers, myself included, use more than just Adsense, a better title for this post would probably be “Using Six Sigma & Statistics To Increase PPC Revenue” – but putting Adsense in the title is more catchy’er so I’m going with it.

In my last post I mentioned that I was going to be doing more analysis and testing of my different PPC networks to try and find ways to optimize earnings. The method(s) I’m using to look at my data are DMAIC and Six Sigma strategies. These processes are new to me and I’m hardly even a green-belt myself – but I have a “Black Belt” mentoring me as I go through this process to guide me (and basically tell me what to do). The simple version of all these processes basically boil down to really looking at everything (stats) that impacts your desired result (earnings), then analyzing the data, making changes/experimenting, looking at the data again to validate your changes worked (or didnt), and applying the changes.

The first thing we had to “Define” was my “Big-Y” – this is the variable that I want to improve. In my case the “Y” is earnings. We drilled down through our “X‘s”, the variables that affect my Y, and we focused on PPC, and ultimately Clicks/CTR. There is an entire process involved in drilling-down through the X values to determine what to focus on – but even though I already knew that CTR would be my critical-X, my Black-Belt mentor made me jump through the hoops anyway – this kind discomfort is apparently one of the keys to Six Sigma.

I won’t go much more into the DMAIC and Six Sigma processes (they are incredibly complicated and boring) but I do want to share a few things that we discovered while beginning to do some statistical analysis.
The first thing we discovered is that with Adsense, more clicks = more money. This is something that we think we all know, but we proved this “statistically” and have demonstrated it with the following graph (note that actual values are not displayed to remain TOS compliant):

The next thing we discovered is that with Kontera, more clicks does NOT = more money! This was quite a surprise, and seems a bit ass-backwards, but the numbers prove it:

Obviously this does not mean that I should stop using Kontera – heck, Kontera earns more than Adsense on many days. But what it does tell me is that if I want to focus and spend time and energy on experimenting/improving CTR, that time would be better spent on improving Adsense CTR, because based on the data (see the Adsense chart), more clicks virtually always means higher earnings with Adsense. But, with Kontera, more clicks does not guarantee more earnings. So this reinforces that I should be focusing on how I can improve Adsense CTR to achieve maximum earnings.

We also partially busted the myth that Kontera clicks will reduce your Adsense CTR. As you can see in the chart below (actual values omitted to remain TOS compliant) there is NO correlation between Kontera clicks and Adsense clicks:

One of the tests that we will be doing over the next few months will be to compare CTR of Adsense & Kontera individually to either prove or bust the “Kontera reduces Adsense CTR” myth.

I will be continuing this process of collecting data, crunching numbers, making stupid graphs and figuring out what it all means over the next few months. As usual, as I learn or make interesting discoveries I will share them here.

3 thoughts on “Using Six Sigma & Statistics To Increase Adsense Revenue

  1. Well, this get’s my attention.

    I am a process engineer with year of experience in this sort of stuff — including Taguchi methods and Six Sigma. I’d say you are on the right track.

    I was a little confused by your AdSense vs. PPC statements at the beginning of the article. Are you talk about AdSense revenue from content that you have published, or are you talking about sales or products made via PPC?

    I assume each dot in your plots above represent a day?

    Anyway, I am very interested in seeing where you take this.

  2. [quote comment=”2294″]
    Are you talk about AdSense revenue from content that you have published, or are you talking about sales or products made via PPC?

    I assume each dot in your plots above represent a day?
    Yes – each dot represents one day of data..

    I initially started looking at only Adsense but after my “VOC” expanded to include Kontera, YPn and Microsoft pubCenter – so my entire project will be looking at all of my PPC providers on my main site instead of of just Adsense.

  3. I’m looking at ways to move some eggs around so am going to test Kontera on my biggest adsense earner and see what happens…. I think i’ll do a week on, a week off kind of thing and collect the data as I go along.

    @Mark, ahhh a familiar face, I found this article via google so always weird when you bump into someone you know 😉

Leave a Reply

Your email address will not be published.