Analyzing Adsense performance with Google Analytics (Part Two)

It's been 9 months since Google integrated Adsense reporting into Analytics. As I mentioned at Part One, the timing of this was a bit unfortunate in that it coincided with an economic downturn: Adsense revenues dipped sharply just as this tool became available.

However, we do have nine months now, and that's something. Because of the fluctuations and the fact that nine months is still not very much time, I'm hesitant to rely very much on what Analytics tells me, but some of it is at least interesting.

Channel Data vs. Page Data

Google has had channels for some time. With some effort, you could get an idea of how well certain sections or types of ads were doing, but unless your site was very small or you used third party tools like Asrep, you couldn't really tell how individual pages contributed to your income. Analytics will tell you how much income a specific page has earned. It won't break that down by individual ads - if you run a banner at the top and a skyscraper in the sidebar, Analytics won't tell you which of those ads was responsible for the income, but you can often guess by looking at channel data: if your skyscraper channel only earned $5.00 on a particular day and a page had $10.00 of income from one ad, it couldn't have been from the skyscraper... except...

Missing data

Analytics isn't perfect. Channel data isn't always accurate either. I've seen several cases where income shows up for a particular channel that happens to be limited to a small set of pages. Sometimes Analytics doesn't show any income for those pages! The reverse situation probably happens also, although I haven't noticed that.

What I have noticed is that income reported by Analytics is always less than normal Adsense reporting. That could be that I still have some old pages that lack the Analytics tracking code, but as I have been able to see definite omissions of channel data, I suspect that the tracking code just doesn't always work reliably. It could be someone clicking on an ad before the full page loads or some other problem, but it's not 100% accurate.

It's good enough.

Distortions of channel data

One lesson learned is that channel data distorts reality. For example, you may have a "Main Ad" channel that shows 30 clicks and $24.00 in income for today. From that, you might assume that each click was worth 80 cents, but when you look at Analytics, you may find that one click was worth $10.00 and the other 29 were all low paying.

Obviously you want more of the $10.00 clicks.

Still too early

Nine months really isn't enough, especially with the current confusion. Adsense income has picked up recently, but interestingly it has shifted for me - I used to make the most money Monday through Thursday but now it is often Friday through Tuesday. That might indicate more individual visits than corporate - or it might mean sad IT people forced to work weekends!

So, it looks like there will have to be a "Part Three" next year.