| 5:56 pm on Jun 8, 2005 (gmt 0)|
| 6:24 pm on Jun 8, 2005 (gmt 0)|
Do you have AdSense blocks on every single page in your site? 'Cause that's how G does it's page loads, through how many times an AdSense block is loaded. If only 1/3 of your pages have AdSense, only 1/3 of the true pageviews for your site will be shown in your AdSense stats.
| 7:17 pm on Jun 8, 2005 (gmt 0)|
thanks rocky.... yes, the ad blocks are on every single page.
| 7:21 pm on Jun 8, 2005 (gmt 0)|
I have had quite a similar experience, but it actually behaves the same way in other ad agencies, so I think the reason must be that my own ad packages count things as page views, when they are other things, such as iframes, download files etc. That's the only explanation I've ever come up with anyhow.
| 7:32 pm on Jun 8, 2005 (gmt 0)|
Is it possible to google only serve about 1/4 of page views with ad blocks? - I've never experienced this though!
| 10:10 pm on Jun 8, 2005 (gmt 0)|
Where is your java/php checker at.
It should be the last piece of code on your page. otherwise it would count page views on pages that did not load all the way.
If you put a php script at the top and one at the bottom of a page there is often a large difference.
Usually adsense shows a little after page content which would mean there was even more discrepancy.
Also I don't think PSA's count as impressions
| 11:43 pm on Jun 8, 2005 (gmt 0)|
The problem your getting is probably due to a bad counter/tracker. My webalyzer tracker gives various results and is never very accurate. I used the free service from statcounter and the pageviews are pretty much the same as my adsense page views.
| 11:26 am on Jun 14, 2005 (gmt 0)|
The larger your site, in terms of pages, the more likely it is that bots account for large numbers of page views. These might be search engine bots but they are probably largely scraper site bots and other unhelpful ones.
| 11:55 am on Jun 14, 2005 (gmt 0)|
Webalizer is notoriously inaccurate in determining actual unique visitor numbers since it regards all visits as uniques including spiders, bots etc.
We experimented for quite a while with different counters and found Awstats to be one of the most accurate for ease of use however we do also use a couple of web log analysers.
The difference between Webalizer and Awstats showed that anywhere between 20% to 50% on even busy sites were actually spiders/bots.
Insofar as Adsense inaccurate page impressions are concerned I have been complaining about this since 5th May 2005, otherwise known as 05/05/05, and I have never read so much canned clap trap from Adsense support. They have clearly had an issue with our sites and, until yesterday, according to Adsense our traffic was down by some 40% over the past couple of months, remarkably, as I commented in this thread this morning, [webmasterworld.com...] our traffic has suddenly returned to normal even though all our statistics prove that they have not changed since 5th May!
We'll see how today pans out however I have fingers crossed that Adsense loves me again.
So maybe you have been subjected to the same problems we have had therefore how did your stats for yesterday look?
| 1:17 pm on Jun 14, 2005 (gmt 0)|
If you use Awstats it does count page IFrames as Pageviews. So a page with one IFrame will have 2 Pageviews indicated.
I download my site('s)logs to my home PC where then you can ad Awstat filters to avoid some of the miscounting.
I look at my own site alot, bet Adsense doesn't count that. I have difficulty filtering this from my logs at home, because my IP changes alot.
I've got a site uptime checker, I've made sure hits from that aren't in my logs, Adsense probably catches this too. Awstats might count these as Pageviews, depends on your setup.
The list goes on and on.
If you had multiple ad units on a single page, only after May 5th did the Adsense page views count come close to being accurate. With adlinks and another ad unit it was always 2X too high.
| 1:21 pm on Jun 14, 2005 (gmt 0)|
Also do not forget that all bots crawling the internet also add up to page views.
These suck up a lot of bandwidth and page views.
| 3:03 pm on Jun 14, 2005 (gmt 0)|
Another weird issue I don't like much and have begun to fight is Internet Service provider caches. A provider like AOL is copying all our pages into its cache. When the end user clicks on your site your server may not even get a hit!
This makes your impressions lower than Google's. Adsense figured this out real quick and put something in the HTTP response header to tell caching ISP software it better actually fetch the real Adsense code from the adsense server not from the ISP's cache. Here's a sample Adsense server response:
HTTP/1.x 200 OK
P3P: policyref="http://www.googleadservices.com/pagead/p3p.xml", CP="NOI DEV PSA PSD IVA PVD OTP OUR OTR IND OTC"
Content-Type: text/html; charset=UTF-8
Cache-Control: private, x-gzip-ok=""
Date: Tue, 14 Jun 2005 14:47:16 GMT
Notice the "Cache-Control: private, x-gzip-ok="" ", Google Adsense has put this in to make sure your ads are shown (and credited) even if your page is viewed out of an ISP's cache.
Unless you've taken similar measures you will see more and more hits from ISP caches in your logs than from actual users. I've notice lots of hits from AOL's caching servers in my logs so I've take steps to do the same thing Google has; I want to know when someone has viewed my page! (I've also taken steps, like GZIP compression, to make sure my pages are served up fast even to the poor abandoned 56K modem users out there; 50% of the web users by the way)
This is one way all these "web accelerators" are boosting speed by caching all of our pages. Your servers may never see the actual users hits!