| 3:59 pm on Aug 24, 2006 (gmt 0)|
One problem is that your AdSense stats are two steps away from your AdWords stats. It would be better to look at your logs for Google referrers, better yet to use a unique URL for your AdWords landing page so you can monitor your traffic even more closely.
| 4:46 pm on Aug 24, 2006 (gmt 0)|
it is a unique html page already, unless I don't understand what you mean;
All traffic on that page comes from Adwords, the landing page from Adwords is the page where there are 2 Adsense ad units.
It is a page specifically made for the traffic coming from Adwords.
| 4:57 pm on Aug 24, 2006 (gmt 0)|
| 5:02 pm on Aug 24, 2006 (gmt 0)|
I'm saying look at your raw site logs, or at the very least a statistics report (e.g. Google analytics). That's the only way you can know what's really happening. At the moment you're just guessing.
| 5:31 pm on Aug 24, 2006 (gmt 0)|
Just a quick one...if AdSense does not serve a new ad it does not show as an impression therefore if the same ads were repeated several times it could easily skew the figures?
That's always been my understanding, am I correct?
| 5:48 pm on Aug 24, 2006 (gmt 0)|
Alright I'm gonna go with Google Analytics and see what can be done.
| 5:49 pm on Aug 24, 2006 (gmt 0)|
Thanks BigDave, I didn't know that.
| 5:56 pm on Aug 24, 2006 (gmt 0)|
Google analytics is also JS based, so if they have JS turned off, their stats won't show up. If you have access to your raw logs or a log analysis program, that is the way to go. I recommend awstats as one of the best for people just getting started analysing their traffic.
| 6:21 pm on Aug 24, 2006 (gmt 0)|
Thanks, but I've been thinking more about this...
In Adwords I chose from the start that the ads only appear on content websites and not in Google's search;
So it might not be that.
Anyways, I'll see what Google Analytics will say about this all, and I might check the raw logs also.
| 6:50 pm on Aug 24, 2006 (gmt 0)|
|Just a quick one...if AdSense does not serve a new ad it does not show as an impression therefore if the same ads were repeated several times it could easily skew the figures? |
That's always been my understanding, am I correct?
I don't think we know exactly what G counts as an impression, but with MeUzik's one-page site and 809 Adwords clicks there should be somewhere around that number of unique impressions at least. But, as has been said, without checking the logs or analytics there really isn't enough to go on.
| 7:11 pm on Aug 24, 2006 (gmt 0)|
I will have to see what Google Analytics will reveal because I can't download the raw logs, and the hosting's log program doesn't say much.
| 7:26 pm on Aug 24, 2006 (gmt 0)|
In the Adsense FAQ from 2004 I saw that every ad unit counts as an impression.
"Page Impressions = Number of times an ad unit (collection of 1-4 ads/image) is displayed. (2 ad units on the page gives 2 impressions each time the page is loaded)."
But again the message is from 2004 so there's a very high probability it changed since, and that an impression is counted as one page load no matter how many ad units now.
I searched for a current faq on Adsense but didn't find any yet.
Then it could explain the "lost" impressions since I do not use such an "ad blocker disabler".
| 7:48 pm on Aug 24, 2006 (gmt 0)|
|each time the page is loaded |
Ah, good, that's what I wasn't sure of, whether by just page load or page load and unique viewer.
| 8:21 pm on Aug 24, 2006 (gmt 0)|
|But again the message is from 2004 so there's a very high probability it changed since, and that an impression is counted as one page load no matter how many ad units now. |
If I remember correctly this was the update of 05.05.05 that gave me an absolute hammering!
| 4:51 am on Aug 25, 2006 (gmt 0)|
Now with Google analytics I had to tell the time region I'm in (eastern time) so now the stats will be even more nonsense because Adsense will not be on the same time frame (and I didn't find where I could change it if it's possible), Adwords and Adsense used to be, however in the long run it should make sense...
I may update this post if I see something about this all.
Thanks for your answers so far.
I should've put "Something MIGHT be wrong with Adsense's stats" as the subject of this post, but at the time I wrote it I was a little upset and worried, because during these try-outs it continues to cost me money in Adwords... ;)
| 5:53 am on Aug 25, 2006 (gmt 0)|
You are right, MeUzik, and Google is wrong. Simply and plain.
If Google send you an invoice with 809 clicks, then you should have 809 page impresssions in your Adsense stats. As said in previous threads, this is the best way to demonstrate how innacurate are Adsense and Adwords stats and how vaste are the fields for scammers or, simply, thieves.
| 6:17 am on Aug 25, 2006 (gmt 0)|
|If Google send you an invoice with 809 clicks, then you should have 809 page impresssions in your Adsense stats. |
Um, no. Assuming you are interpreting your stats correctly then there are 2 separate possibilities here:
1. You are not getting the correct number of visitors from Google.
2. You ARE getting that many visitors, but for some reason Google is not displaying ads to everyone.
Your first step should really be figuring out which of these is the case. That's why you need access to accurate log stats.
- Poor server responsiveness could cause #1 and/or #2.
- Showing public service ads for whatever reason could cause #2.
- Fraudulent clicks COULD cause #1 and/or #2, but the ads should only take a second or so to load, so frankly I'd expect that fraudsters would usually generate impressions anyway.
- Google reporting could be incorrect, but in my experience the AdSense stats are always very accurate. I have no experience with the AdWords side.
| 7:08 am on Aug 25, 2006 (gmt 0)|
Add to that the following possibilities:
- people stop loading the pages before tha page is fully loaded.
- people hit the back button before the page is fully loaded.
- people close their browser before the page has fully loaded.
- people click on a link before the page has fully loaded.
If your pages do not load within 5-10 seconds, you can be certain that many of those will happen. If your ads don't show up till the end of the page rendering, it will be an even higher percentage.
There are also plenty of times where google delays the showing of impressions or clicks for quite a while. It would be nice if they kept the stats totally up to date, but I don't really sweat it because the clicks do seem to show up eventually.
| 7:58 am on Aug 25, 2006 (gmt 0)|
Those examples are what I was basically getting at when I suggested poor server responsiveness -- the server being either down or slower than the attention span of the average surfer.
| 8:19 am on Aug 25, 2006 (gmt 0)|
Has anyone considered spyware?
I have this same problem, checked my server logs & the numbers just don't match. Its always about 10-20% off.