Forum Moderators: martinibuster
Question is , how to find out where?
Thanks.
>Looking at my channels I have every site I have in there, but the clicks and impressions are
>higher, so obviously they are coming from somewhere else
Are you looking at stats for "today" or previous days.
If 'today', this might just be due to the fact that Adsense statistics are maintained on a number of different servers, and your report screen draws stats from different sources. I don't fully understand how this works, but as the servers try to keep sync with each other, validate clicks, put suspect ones on hold, etc., the different screens can report different numbers, and sometimes the summary stats can go backwards and forwards.
If it is for previous days, although the different servers should have been synced, I think the stats can still be distorted by clicks that have been held in reserve and then dumped, once validated.
I am sure i covered al the webpages and not missing one or more. Checked this en double checked.
I could imagine that page views via the Google cache wouldn't appear in any of your URL channels.
If the copied page has been indexed you should be able to find it by searching for unique text or using Copyscape.
If someone is running your ads without their being indexed you're just very lucky or we're back to "googley things" which I am finding more and more of the closer I examine my stats.
That's a euphemism for "we haven't established anything"! Imho, there is usually a 'rational' explanation for everything, and it's a question of deciding whether it is worth trying to get to the bottom of things. Sometimes it is, because there have in the past been errors in the Google reporting system.
However, I'm not experiencing the size of discrepancies that leonardp is reporting so, unless lots of others report the same thing, I don't think that is a Google error - it is more likely to be something about how his channels are set up.
I have a channel for my domain, for the index page, for each directory and some channels for some pages in the directories.
If I compare the channel for my domain with the channels for the directories plus the directory for the index page, the channel for the whole domain has more pageviews and clicks then the channels for the directories and the index together. That's strange right?
That's a euphemism for "we haven't established anything"!
We do resemble the proverbial blind men describing an elephant. I stand by my analysis above.
Agreed. Except, I assume you resemble a woman :-).
This puzzles me - is your channel set up for your index page, or for a directory that is accessed by the index page, or something else? The URL channel for your index page probably should be something like [yourdomain.com...] If not, you may be missing index page impressions from your channels.
Ergo, to count most 404s, you would need to look at your logs to see what mistypes are being used, and then set up a channel to track them, if that is possible (and it isn't always).
I have only done these types of analyses with server log data because traffic is the more important component for me. But I will eventually begin analyzing my AdSense data this way as well.
Agreed. We have been doing this over a long period.
>I don't think you can draw any valuable conclusions until you have a large enough sample.
Again, agreed. Although, the sample sizes we are dealing with are so large that they make a mockery of "statistical significance" - virtually everything becomes statistically significant when you are dealing with large numbers. Yet, to properly understand the (long term) data one needs to use appropriate statistical tests.
>The short term figures simply boggle the mind and distract you from more important issues.
That can be true, but there are two benefits of looking at short term figures.
One is that it provides confirmation that everything is working. If our stats get stuck, we check our website and, just occasionally, it has been down. On one occasion it transpired that our (reputable) ISP was taking the system down during prime business time for scheduled maintenance - they have politely been asked not to :-).
A second is that reconciling the figures can help to make sure that your long-term channel data is being recorded correctly. We experienced leonardp's problem some time ago, and it turned out that our channels weren't set up properly. Now, more or less, our data reconciles, so we know that the long term data we are capturing is accurate and therefore useful.
there are two benefits of looking at short term figures.I confess to checking my stats (and channels) many times a day, I have the Adsense plug-in for Google Desktop as well as checking a normal log-in and the google.co.uk log-in. No one is consistently ahead, they leap-frog.
This will catch any big problem quickly and I have gotten so sensitive to the flow that I noticed when just one of my pages (channels) inexplicably disappeared from the Google index for several hours. I never did figure out why, very typical of trying to grok Google.
My web host is so reliable that I have not experienced more than two hours of unscheduled downtime in three years. Watching adsense stats is a good monitor of that too.
But I'm afraid I've boggled my mind far too often trying to figure out the odd numbers and turning to WebmasterWorld forums and others to confirm that I'm not alone...
It's just a compulsion, but yes an effective monitor too.
Still, I try to follow my own advice and concentrate more on longer-term numbers. :)
Some time ago we discussed cached pages at google and yahoo? Publishers code is included in these but if channels recently added would not be unless crawled since then.
This is a good point and brings up a few other possibilities as well--though these only come up as a tiny fraction of the pages served, it seems unlikely to be more than 1% and probably not causing the discrepancy Fugazi was concerned about... Fugazi hasn't been back for a while.
Others: I noticed my pages on archive.org still have AdSense code on them.
I have also seen my pages reproduced and translated into other languages and served from other domains. Since the text is all different and they display my ads I like this.
Hello Everyone. My name is Blue. And I'm also a Stataholic. All we need now to make these Stataholics Anonymous meetings official is for someone to come up with a 12-point plan on how to give up our addiction to stats :-).
On a serious point, this thread has highlighted how difficult it is to capture comprehensive data using URL channels. I don't subscribe to incrediBILL & Ann's view (expressed elsewhere) that channels are garbage, because I've found them very, very useful. But they could be improved.
And one improvement would be to have a comprehensie URL report - listing impression data by unique URLs, rather than having to define channels. This could be analysed by CSV download. However, I recognise this might create problems, eg: gobbling up Adsense resources, especially for mega-sites.
Dont think so, but i'll look into that one...
