Forum Moderators: martinibuster
i've never really paid much attention to this in the past, because i've been in the game long enough to know that underreporting is normal. in fact, i probably hadn't bothered to compare the figures in six months.
but today i decided to have a look, and was shocked to find that adsense is reporting just 20% of what my server reckons i get.
now... i immediately read up on what adsense counts as a page impression, and it said no matter how many ads you have on the page, it will only count the page impression once per visitor. fine... that is what my server stats count it as too.
and it can't have anything to do with bots or people with javascript turned off, because that would imply that 80% off my traffic is coming from those. (and given the amount of traffic i get, that is impossible.)
i thought it might have something to do with that "Allowed sites" feature. Previously i only had 'example.com' in the box. so i'm thinking of adding 'www.example.com' too.
anyone know of anything else which might cut down on the difference a bit?
Read that thread and you'll see who's been writing your stats.
given the amount of traffic i get, that is impossible.
It's only impossible because you don't want to embrace the fact that bots often outnumber humans when it comes to generating traffic.
Any browser that can execute the iwebtrack or google analytics code, which is Javascript, should also be displaying the ad. So the whole "more bots than users" theory doesn't explain how both of these web analytics services show more page views than Google claims to have shown, in a login area where ONLY human-browsers can access. And I'm not talking about a site where 20 or 30 people a day log in here, this is a 36,000 member site with millions of page views per month. On any given day, Google's numbers are off by at LEAST 20%.
I've asked Google this question through the AdSense contact form twice. Both times I received an e-mail confirmation that my message had been received, and then within a day I received another email asking me a survey about how satisfied I was with my "recently answered question".
The bigger question is why Google refuses to answer. Google is marking the question as completed in their support system even though they aren't answering it (that's why I'm getting the same follow-up survey that I always get after they DO answer a question).
Their numbers ARE off. What are they hiding?
There are other answers besides scrapers which such as Norton Firewall, ad blockers, no javascript, PSA ads or the Google ad network simply didn't respond to the page request. Even behind a login page things such as Firefox or Google Web Accelerator doing PRE-FETCH of a bunch of pages that are never viewed will skew the results. Additionally, people using RSS feed locators will also skew the results, even with a LOGIN.
Unless your LOGIN page has a captcha or some other way of thwarting bot access, don't expect that all of your registered traffic is human either. If someone wants access to your content bad enough it may take a human to create the account yet bots login and crawl the account afterwards.
Trust me, if someone wants to crawl your site bad enough, they'll find a way to do it, including manually answering captcha's for their crawler which I've encountered more than a few times.
Anyway, if you have Google Analytics or a similar javascript traffic tracker installed and the results are way off from AdSense, then you have a valid reason to be upset.
Otherwise, forget the server logs as they'll just make you pull your hair out because they aren't a true reflection of human traffic.
[edited by: incrediBILL at 1:01 am (utc) on Aug. 3, 2008]
Way back I added Google Analytics to one of my big sites, but it slowed down page response
It doesn't have to be Google Analytics as there are many free javascript-based analytics programs but they don't tend to be as comprehensive.
Have you compared AdSense with Google Analytics?
Yes - the numbers are different, that's what I'm saying. Google analytics and iwebtrack give one number (usually not exactly the same number but within half a percent of each other). AdSense gives another, sometimes 20% lower. The analytics code and the adsense code are in the same include file, so the numbers should always be nearly identical - anything that executes one should be also executing the other.
I'm not new at this stuff, I know all of the ways a site can be accessed. This isn't what we're talking about here, it's not scrapers and bots in the login area. The numbers are different and every time I ask the AdSense team about it the question is never answered.
I have a theory as to why, and I shared it with Google (naturally they never responded), but I don't know what their thoughts were. There are plenty of "ad remover" programs out there that claim to block advertisements, and most of them do this by modifying the hosts file in system32\drivers\etc - they modify the list of domains here and change the IP addresses for popular ad services to 127.0.0.1. Perhaps the use of these types of programs are more widespread than people realize.
At least that's what I DID think... the fact that Google doesn't seem to want to talk about it makes me wonder if there is more to it than that though. Perhaps they're too busy driving around taking pictures of our houses to care.
Read this post regarding the use of the "google_alternate_ad_url" [webmasterworld.com] and then check your logs to see how many times that alternate URL is being used.
It that doesn't account for the discrepancy, then I'm kind of at a loss.
The only other thing I can think of is you have a page(s) that's not generating any ads whatsoever that you aren't aware of, not even using the google_alternate_ad_url.
I've seen this happen before but it seems you have to find out there the problem is before the AdSense people will help you rectify the issue.
In the last couple of instances where I had something like this the email came back with "we don't see any issue" yet the page(s) were suddenly and magically displaying ads again.
Go figure.
The only other thing I can think of is you have a page(s) that's not generating any ads whatsoever that you aren't aware of, not even using the google_alternate_ad_url.
That's possible, I hadn't considered that, but I would think they would be showing SOMETHING by now. Usually when AdSense sees a page for the first time it shows my alternate ad (I disabled PSA's and have an alt ad url) for a couple hours, then starts showing ads. I guess it's possible that there are some pages that just aren't targeted to anything at all, but it seems that with the header code being universal across all sites it would still show a lot of the same ads shown on the home page.
I'll give the alternate ad url a read and see where that goes. Do you think Google wouldn't want to acknowledge that third party ad blockers are blocking their ads?
I guess it's possible that there are some pages that just aren't targeted
I've seen glitches where previously well performing pages just suddenly went dark for no reason, no PSA, nada. Then they came back without any explanation whatsoever, so it's not just new pages you can have problems with.
Do you think Google wouldn't want to acknowledge that third party ad blockers are blocking their ads?
Not sure what their motivation would be.
Another thing you might want to try...
<script>
AdSense code here...
Some other counter here...
</script>
<no script>
Some image here...
</no script>
Then you can physically test how many times the adsense script should be loaded and how many times the NOSCRIPT image gets loaded.
Could be an amusing test to see what turns up.
In the past you could tell by the size of the AdSense element, because certain ad blockers set the size to zero. I had a script served other ads if the size = 0, but Google changed their code, and the script stopped working. That killed AdSense on my site for a few hours until I figured it out and disabled the script.
I.e. Google Analytics may have code that registers a page view even if people hit the "back" button. Adsense, on the other hand, may serve the ad from the cache and thus does not register a view. I have not looked into this issue closer (and I do not run GA), so this is just a guess. (And whether a 20% "back button" rate is realistic for your site, that's another question.)
I agree that Google may not be comfortable answering this question (as they never are comfortable answering any questions that go beyond being "basic"). Then again, I do not see why Google might hide views? Clicks - maybe, but views? Where's the point?
(Probably it's the fact that Google -per definition- can't be wrong, so they can't admit to be wrong in this case.)
I am using an ad server and it shows almost double the clicks Google shows. What could be the reason for such a discrepancy?
