Forum Moderators: martinibuster
I run a computer PC hardware content site (news, reviews, listings).
Anyone take a guess?
What do you think they will be able to tell from that?
If a site has X% CTR for a PSA, and 5X% CTR for ads, then the site may simply have very targeted niche traffic (a tech blog, perhaps).
If a site has X% CTR for a PSA, and X% CTR for ads, then the site may have very broad general traffic - a portal site, for example.
If a site has 0% CTR for a PSA, and X% CTR for ads, then that would look fishy. That is the only obvious case for investigation. And surely no successful fraudster would silly enough to implement their scam so badly!
Edit: the only reason I can imagine Google would ever want to have these figures is to build a database of sites ranked by ad_CTR:PSA_CTR ratio, and work out what the leaders have in common. Then perhaps lead a project to build more sites with good ratios, thus increasing overall clicks to take full advantage of advertizer's adwords' budgets, and hence maximize revenue. Strange that they bought Blogger *before* any such analysis took place.
This fixation on extreme short-term fluctuation is simply a time wasting exercise.
Wise words, wise words. If only we could convince people to stop getting excited or depressed over how Adsense did in the last 6 hours.
corpuscle, if you have enough data it's amazing what a good mathematician can deduct from it. Serving PSA intentionally and watching to see if your CTR does indeed change is something that would work to detect fraudulent clicks. What if during the time Google was serving PSAs your CTR doesn't drop to PSA level of CTR but stays at "normal" level of CTR? You possibly then have some bot clicking away at the ads.
If a site has 0% CTR for a PSA, and X% CTR for ads, then that would look fishy. That is the only obvious case for investigation. And surely no successful fraudster would silly enough to implement their scam so badly!
Not necessarily a fraud. A site could have the type of visitor who's not very charitable and who only clicks on ads that will take them to the information they are urgently seeking.
On a separate matter, EFV, I'm glad it's picking up for you :-)
Tom
Is it the case then that Google doesn't know when someone clicks on one of our Alternative Ads?
Nope. They know.
JAG
What if during the time Google was serving PSAs your CTR doesn't drop to PSA level of CTR but stays at "normal" level of CTR? You possibly then have some bot clicking away at the ads.
I may be interpreting your message incorrectly, but I think you're suggesting there is a "normal", standard level of PSA CTR. I would think that PSA CTR is highly variable, depending upon the nature of a sites' visitors.
The Red Cross site, for example, may produce a better PSA CTR than a browser hijacker that redirects a dismayed user to an adsense stuffed page.
Thus it is only meaningful to compare a site's AD CTR with the *same* site's PSA CTR.
If Google did indeed measure the diffences between PSA and AD CTR for all adsense publishers, then they could calculate an average PSA:AD CTRR and standard deviations to know when a specific site's PSA:AD CTRR looked a bit strange.
If a sites's AD CTR = PSA CTR, then I highly suspect that:
a) this would fit within the norms of their numbers,
b) the site attracts broad, untargetted traffic OR AdSense ads are poorly matched with the content (which should not be that common) OR as you say, a bot is clicking randomly.
My personal feelings (which, mathematically speaking, are irrelevant) are that the latter explanations are far less likely. My opinion however is unecessary: without knowing the source and niche of the site's traffic, it is impossible to imply fraudulent clicks from PSA:AD CTR figures if they fall within the norm.
Not necessarily a fraud. A site could have the type of visitor who's not very charitable and who only clicks on ads that will take them to the information they are urgently seeking.
Good point. I agree with you and withdraw my suggestion that 0% CTR for PSA is fishy. I think 0% PSA CTR might also fall within the norm, in which case nothing fraudulent can be implied.
Thus it is only meaningful to compare a site's AD CTR with the *same* site's PSA CTR
Absolutely! That's exactly what I meant. I've long argued that "average CTR" across different sites is a completely useless metric for us webmasters. I will not change that position now :-)
If a sites's AD CTR = PSA CTR, then I highly suspect that...
What I'm saying is that Google knows what your CTR is and what your average CTR is even better than you do (because they may be carrying some figures forward/backward and the stats displayed to you may be lagging a bit. Also, then know CTR per page!). Google also knows what your PSA CTR is. So assuming that all your pages are serving properly targeted ads all the time Google could intentionally show PSAs on a particular page to see if the CTR for *that* page then drops to that page's normal PSA CTR. If the CTR doesn't drop then this could raise flags at the plex.
JAG
Even if they knew, there would be no way they could use the CTR of the alternative ads to detect fraud:
- alt ad CTR smaller than regular CTR? Maybe I just made an ad that's less interesting than Google's ads.
- alt ad CTR higher than regular CTR? Maybe I wrote something really targeted to my readers.
- changes in alt ad CTR? Maybe I just changed the ads.
What's left for the conspiration theorists then is to show that Google is running PSAs on pages that should be showing either targeted ads or alternate ads. I've seen no evidence of this. (Maybe "conspiration theorists" is a strong term :-) ... I meant those supporting the theory of Google serving PSAs just to check for automated fraudulent clicks)
<edit reason: added the last sentence>
[edited by: Macro at 6:10 pm (utc) on Jan. 25, 2004]
Actual users click PSAs less frequently than regular ads.
I'm not so sure one can assume that.
Don't get me wrong....I understand the logic behind thinking PSA's could be used but I think it's not a very good way to do it and can very inaccurate.
If one click occurs on a paying ad and one click occurs on a PSA and those are the only clicks that happen for a day does that mean a click bot was at work? They are exactly even so it could be but does Google want to punish a content site with the dreaded email because of that? I doubt it.
What if the ads are on an exotic animal site and PSA's for saving Zebra's are shown? That audience may click on more PSA's than ads for Zebra skin rugs. So now what does Google do?
It's just to volatile to make business assumptions on.
JAG
it and can very inaccurate
Yes, UNLESS it was done in a carefully planned way (see previous posts). However, at this point I don't suspect that Google are in fact using PSAs to monitor fraudulent clicks.... not because it's not possible but because the other (circumstantial) evidence suggests they are not.
UNLESS it was done in a carefully planned way (see previous posts)
I didn't see anything that could accurately detect it in any previous posts. No matter how well planned there are too many variables Google cannot control to get an accurate reading.
I agree though that it just isn't happening. At least I hope not for the sake of some unsuspecting content site owners who get nailed by something that can't be accurate.
JAG
Last I checked PSAs didn't use tracking URLs (which is how clicks of Adword advertiser ads are tracked), just direct URLs.
For the benefit of anyone searching the archives at a later date I thought I'd post this. It seems that PSAs now contain a tracking URL like their paid ad counterparts. I don't know when this change was made, but the change was made sometime after alternate ad capability was introduced by Google.
