Adsense Clicks Different from Actual

Big difference in the number of clicks

[dhaliwal]

Hello Everyone

Before i start the topic, i would like to tell that i am not new to Adsense and WebmasterWorld. I have been testing different ad formats, ad tracking, ad placements on our websites and have gained a lot of information from WebmasterWorld and DP on various topics around Adsense.

Now, the issue i am facing may have been discussed here for many times and there may be some canned responses from some members on the forum.

I signed up for a pro account with mybloglog

They give the details of where people clicked and from which page they clicked.

The results are listed below.

For February (1-7 Feb. 2008)
My Adsense account tells 4432 clicks
MyBloglog account tells 5830 clicks

Now there are some points which i would like to share here.

1. We are using only two ad units on a page
2. We are using only Ad units and not Adlinks
3. The stats shown by MyBloglog aren't having many clicks on the same page.
4. Out site is big (with over 15000 pages in total) and clicks were on different pages. So, we can't say that they were double-clicked or something which Google won't count.

Even if we say that there were some invalid clicks, the difference should not be that big.

(Please do not tell that i shouldn't share stats as this is not as per terms of Adsense)

I wanted to share this with members on this board.

Also, i will suggest some members to sign up for an account on Mybloglog.com (its a yahoo company now), if you haven't done it so far. They are also offering 3 day trail, so if you aren't interested in paying, you can still see all the stats.

Members with big sites may see a bigger difference.

I will ask Adsense team after a few days and once i see response from other members on this board.

Please do share your views.

Thanks
Dhaliwal

[tim222]

Yeah, this is a tough one to explain. But here are some of the typical responses:

Could've been bots
Could've been browsers blocking Javascript
Could've been browsers using ad blockers
Could've been PSAs

I'm sure there are more. But I think this is an issue that will drive you crazy if you keep trying to match up your web host page impressions with your AdSense page impressions. They seem to be two different things.

[dhaliwal]

Thanks for a response tim
I would say,

We are not talking about impressions, we are talking about Clicks.

The data i am getting is of the following format in details

Google AdSense (336x280)
http: / / pagead2.googlesyndication.com#336x280

Google AdSense (300x250)
http: / / pagead2.googlesyndication.com#300x250

Google AdSense (234x60)
http: / / pagead2.googlesyndication.com#234x60

Google AdSense (468x60)
http: / / pagead2.googlesyndication.com#468x60

can't be bots as most of them do not follow links in Adsense Javascript.

if a browser blocks Javascript, then Adsense ads won't even show up. So, there isn't a question of click in this case.

I am not talking about page impressions, there are the total number of clicks which i tracked on a regular basis and the data is for seven days.

[alephh]

Well, Google has tendency to ignore quite large percentage of clicks on various grounds.

Also, how reliable are stats by Mybloglog... Have you contacted them and asked about this, and asked if this is normal or exception?

Most Adsense-related stats are infamous for not matching with stats by other apps.

[gibbergibber]

--Most Adsense-related stats are infamous for not matching with stats by other apps. --

If you see the thread on total impressions, Adsense stats don't even match themselves! (At least on my site...)

[zett]

Starting September 2007, I see Adsense underreporting up to 30% on individual days (compared to an independent click tracker I am using).

I could not come to a final conclusion as to what is happening exactly, but I think it is related to the location of the click. In any case Google seems to decide that certain clicks are simply void, and instead of setting their values to $0.00 they just do not count them. (Accordingly, EPC has gone up slightly since September.)

But that's the beauty of Adsense: they don't need to tell you. Heck, they don't even need you (or me or anyone else). So, I'd hold back that mail to Adsense support. You never know what will be triggered once you contact them.

[gibbergibber]

--But that's the beauty of Adsense: they don't need to tell you. Heck, they don't even need you (or me or anyone else).--

Yeah, they could do with some competition but so far there isn't any in Europe.

[MyNewPC]

I just did a Google search and there is a lot of discussion on various blogs about this. From [jeremy.zawodny.com...]

Check your mybloglog for how many adsense clicks you have, then come back to your site, let your mouse hover over one of your adsense ads, then click backspace, or somehow use your keyboard to navigate away from the page. DONT CLICK THE AD. After you’ve navigated away, check your mybloglog and you’ll have another adsense click.

Mybloglog determines clicks based on where the pointer is as the browser navigates away from the page.

I also read on another blog that the same discrepancy appears for affiliate links (and likely any other link on your site). If these findings are true, that would make MyBlogLog useless for this purpose.

[dhaliwal]

What MYNewPC has said, may be true in a very few cases. But, we are talking about big numbers. Usually the site visitors won't do that to the extent of 25% of the total number of visitors.

[MyNewPC]

Total number of visitors? What does that have to do with the discrepancy in clicks reported by Google vs. MyBlogLog?

What you are reporting has nothing to do with the number of visitors to your site.

[gibbergibber]

I think they meant "total number of visitors who behave in that manner", ie the total number that hover over a link and press backspace.

[MyNewPC]

If that were the case, then the number of clicks he says show up in his Adsense reports would be way too high for that number of visitors. He hasn't said how many visitors he had over that 7 day period, but I suspect it's much higher than what you think he meant as it correlates to the discrepancy.

[dhaliwal]

i meant what "gibbergibber" said.

I meant that if most visitors (who click) were doing that process, things would have been much different from what i have noticed now.

[BigDave]

If you have a banner at the top of your page, or any sort of dropdown menus near an ad, or have forms on your page, 25% could be low.

A pointer could be left over a banner after selecting history/bookmark/favorite and leaving the page.

The same goes for a menu that leaves a pointer over an ad, though that is against adsense TOS

And advanced users start typing on a form, they tend to tab their way around the page to the submit button.

I don't completely trust Google's numbers, but I don't understand giving the benefit of the doubt to an obviously flawed tracker implementation either.

[jomaxx]

I'd have to agree with BigDave that Google's numbers are inherently more credible. And it's worth adding that I never used to have a noticeable problem when I used a different click tracker, but I've seen lots of complaints about discrepancies with MyBlogLog. Not sure what the significance of that is, but I'd start there.

What's the exact code that MyBlogLog uses to identify a click? Not the exact exact code, but a reasonably exact pseudo-code version so that we can see how it makes its guess.

Oh, and here's a theory... If they really do rely on "where" the mouse is when you leave the page, note that if the ad block is at the top or at either side of the page, it's very likely that the user moved the mouse over the ad block and kept going right out of the window. I'm no guru, but in my experience of writing mouseover code, Javascript can't detect that the cursor is outside the window, and thus if you're not careful it can erroneously assume the pointer is still over the div at the edge of the page.

[dhaliwal]

as if there are some issues with MyBlogLog, i will try using some other click tracker as well.

[jomaxx]

I don't know if there's a fundamentally better technique, though. The method I used involved checking for the onFocus() event, but the reason I don't use it any more is that this trick no longer works.

Actually, there's a second reason I don't use a click tracker any more -- it never provided useful, actionable information. If you don't know what advertisers are bidding and you can't control which ads will show anyway, it doesn't matter a whole lot which ads people are clicking on. It satisfied my voyeuristic curiosity, but that's about all.