Forum Moderators: martinibuster
I'm sorry to say that it appears every time CPM ads appear on my site(s) earnings drop, probably by half.
There is still something wrong with the algorithm and the many "experiments" are really cutting into earnings.
It appears duplicate ads are also purposefully run, perhaps to test various full size single text CPM ads “click through” performance versus full size CPM image ad “click through” performance and in the meantime Earnings performance drops in half. Intentional duplicate ads can never compete against the basic multi-ad units where the diverse suite of ads is likely to have a relatively high paying click every two or three days on a per page basis.
The CPM ads being run appear to mostly be untargeted. I have had communications with Adsense support, I won't be specific, but these communications strongly indicate quite a few experiments are going on. The only thing missing is the confirmation that these experiments can really hurt earnings. I do appreciate that Adsense support does intelligently respond to input.
Of course there's the TOS which allows all this experimentation.
But the evidence from a Webmaster perspective is certainly, untargeted CPM ads will never produce "click throughs", so inherently they make a given website appear less productive and profitable, which in turn may drive the sites income even lower. These experiments are also associating poor performance historical data to the website, but this poor performance data is unjustified. How will the historical misrepresentation of site performance be corrected? I'm guessing we’ll, just have to wait a couple months and earnings will slowly return.
Perhaps higher paying CPM advertisers will reject sites simply because historically adsense was running untargeted, duplicate, CPM ads on a given site. Don't get me wrong I have seen CPM ads that are appropriate to some of the sites content, but then they are splattered all over the site's pages in an untargeted fashion displacing targeted, non-duplicate, conventional ads that do pay well.
Perhaps some sites are doing well with CPM ads, but it would probably be because they were performing poorly, perhaps in CTR, before the introduction of CPM ads.
My suspicion is sites that had high CTR historically are now seeing earnings drop with the introduction of CPM and "single" ad targeting. It is very difficult to predict or measure random clicks on that infrequent very high paying ad that really boosts earnings. The CPM ads eliminate those high paying random clicks on unusual ad topics that are related to site content, ads that the rare visitor is interested in. The evidence indicates that the Adsense statistics which are probably simply averages are not properly taking into account rare statistical events (outliers) that actually produce significant income for the Publisher and the Advertiser as well, through actual conversions or future conversions; future conversions the publisher is probably not credited with.
Over the past 3 years, for this webmaster traffic is up, CTR is up, earnings per click are way down, but earnings are up some. So one could say, earnings based on quality content have actually steadily dropped. Based on historical data earnings should be substantially higher. So are we saying historically Adsense/Adwords was overcharging advertisers?
But these were advertisers willing to pay!
Adsense has not compensated cost reductions with volume increases. Adwords advertisers are complaining they can’t show their ads even though they were historically. That to me must be income loss for a publisher. I believe these were quality, low volume, low click rate, high paying ads that the enhanced Adwords system is basically rejecting with exorbitant pricing. ( I do use Adwords. )
Oh well enough of this rant. I hope this extremely complex system stabilizes and produces income reliably for all, sometime in the near future.
Look early in the morning at your ECPM on pages with no clicks, if you have ECPM and no clicks, you have CPM ads on those pages. Google may be setting cookies as well to limit CPM ad display. I really haven't investigated this much.
What I think is good right now is I do currently run 3 Adunits. Adsense is currently kind enough, or knowledgeable enough, to not run CPM ads in the most prominent "hot spot" Ad unit location. The CPM ads do seem to currently, and correctly in my opinion, appear in the second and third less prominent Adunit positions. But I maintain even here on higher CTR sites they may be a losing proposition, one that Adsense has not properly statistically accounted for.
CPM ads above or below the "fold" is certainly a big issue, they should have different rates. I haven't played with CPM on Adwords enough to know about this factor.
If a CPM ad unit appears in the first, most prominent Ad Unit location, it must pay well indeed!
I added all these units to facilatate a more flexible environment for Adsense to play with, even Adsense support recommended setting the second of 3 units to "text only" and they realized this would only limit duplicate image CPM ads. The experiments were clearly running different versions of the same ads probably to sample "click through" rate. perhaps even conversion rate, on these CPM ads.
This is the missing piece of data for Adsense, CTR's and conversions for CPM ads, they certainly know how much they cost, but not CTR or conversions! But in the meantime non-relevant performance data is being recorded for a given site. It is performance data that has nothing to do with the sites true potential.
Does adsense know the location of CPM ads on the page? They must parse CSS to know for sure. Anybody seen a CSS file read by the Adsense Media bot?
So some sites may be making out like bandits during the experiments while others are seeing earnings fall, so if Google just looks at averages, things may look better. And again I ask does this questionable historical site performance data possibly impact future site earnings.
Like I said I hope the is stabilized, and balanced, before the high traffic season that shortly will be upon us!