Site stats and AdSense stats differ?

any suggestions?

[axlRose]

Hi everyone,

Do not know if this is the right question in this forum, but still i`ll ask it. To my opinion "Page impressions" in AdSense means - how many times adsense script was invoked - am i right? If yes it is the synonym for "Page Views" in a statistical program. But the numbers differ in my stat and in adsense differ for about 8.000 impressions - ex.: my stats shows that i had 20.000 hits july 1-st, but adsense shows only 13.000 - whay can this happen? Users not using JS? No - they do use it - the number of visitors who don`t use JS is very small on my site. I do not also count robot hits - 20.000 is the number of real users hits. Any comments on this?

Best regards,
/Alex.

[jino]

There was a recent post on this. Some of the cause for discrepancies are:

The logging of your stats for pages like SSI files, CSS style sheets, .js files, etc. If you exclude these files from your logs, you will find the discrepancies much less.

Also some badly behaved bots will be logged as page views even though it doesnt call the google ads.

I have found google ads page count to be extremely accurate after I remove these variables.

[axlRose]

Thanks jino. But u see i only count html pages where i have google adsense code. All robots are excluded and they are counted as robot hits. So 20.000 hits is actual users hits :(

[bnhall]

sometimes it takes the ads a while to load - so a page hit pay not be the same as an adsense impression.

[jino]

There are also some users with ad blockers like norton which will block out adsense ads. Hence, your html will show but the ads will not be called.

[jomaxx]

I doubt you are really excluding all robots, no matter what technique you are using. I get a LOT of robot traffic that tries very hard to look like a human surfer. Anyway, the best way to simulate what Google considers to be impressions would be to display a tracking image using a similar technique - i.e. using Javascript inside an iframe.

[jim_w]

I, like others, noticed a drop in our adsense revenue. I saw that it was attributed to some ads that were not being displayed any longer on some of our most viewed pages.

We have a log of ‘eyes on page’ so we know when a page actually gets viewed and not crawled by known bots, etc. So for a test I took one that looked really off and counted what our log indicated vs what (G) said. When I count the pages, we should have had 1001 page impressions, but (G) said we only had 856. So I thought, ok maybe there were a bunch of Code 206, partial content. But when I check the stats for the entire site, there were only 156 for the entire month. The delta of the discrepancy is 145 for just that one day. Now while it is possible that 145 of the 156 total 206’s came from the one day, I find it hard to believe that a high percentage of 201’s came from just one day.

Depending on the click price that could be a lot of money. Even if half of the 206’s came from that day, and I’ll bet it isn’t that high because the average over 12 days is 13, it could still be a lot of money, depending on what pages the clicks were not accounted for.

This is for just one day. Is there really a possibility that there were 145 hits to adsense pages that were bots reading JavaScript and/or browsers not using iframe, etc.? If some were ‘Programs that people may have written to grab website content ’, it would have been counted in our eye’s on log as well, if the bot was trying to look like a human surfer. We are talking about an 85% error here. That seems much too high to be attributed to the items (G) says can cause a discrepancy, +/- the 206’s. I have found with our banners stats that a lot of businesses, and we are a B2B, don't employ ad blocking at all.

jino
My Norton, Internet Security 2003, shows adsense ads. If an ad blocker kept the ads from displaying from my HTML page, wouldn’t that also keep them from being fetched from (G), thus nulling that issue out?