Adsense Ad Impressions vs. Actual Pageviews

While doing some counter-checks between the total Adsense ad impressions for one site against its recorded pageviews, I’ve noticed some significant amount of discrepancies. Adsense ad impression is counted if a certain page displays a certain single ad block. Theoretically, that means that if you have a single ad on a page, one pageview will generate one ad impression.

However, certain mitigating circumstances might prevent an ad impression on a page view and to some of you who have been observant enough, you will notice that your first couple of pageviews for a new blog entry would either result in a PSA or an empty ad space. PSAs still counts into ad impressions but obviously the blank ones don’t. This issue is factored into the discrepancy mentioned above, however I took 3 sample stats figures and compare them — 71.3%, 64.3%, 64.6%. These percentages represent the ratio between the Adsense ad impressions and the actual pageviews and there seems to be a huge difference.

I did a quick research and some more reading in forums on why or how this could have happened:

• Not all pageviews generate ad impressions. This is the major reason but what is the normal ratio for this is basically dependent on many factors.
• Not all stats tools are equal. What we used here to monitor the actual pageviews of the blog was WP-ShortStats and yes I know that it’s one of those unreliable (and non-standard) stats tools out there. I should have used another 3rd party monitor like Sitemeter or Statscounter to verify and calibrate this.
• Stats cycle — Adsense’s cycle starts and ends at around 3:00PM, Manila Standard time while our stats was doing it at exactly 12 midnight. So the numbers we are getting from the ad impressions will not really tally up because of the 9-hour time difference.
• Number of ad units on a page. In this case, there’s just one. However, it’s fairly common to see fewer or even blank ads on pages will 2 or 3 huge Adsense ads.

My concern now is that WP-Shortstats seems to be totally out of whack in terms of pageviews because of this huge difference. Will have to find out “by how much” it overshoots by comparing it to Sitemeter and Statscounter.

More on this in the next couple of days.

[tags]google adsense, statistics, ads, metrics[/tags]