Welcome to WebmasterWorld Guest from 188.8.131.52
I've seen people in this thread saying that this has prompted them to try YPN, and it seems that YPN is even a viable alternative for some people, if they get good targeting. I haven't tried it yet, but incidents like this certainly push me in that direction.
Unless I learn otherwise, I consider this to have been a technical glitch and I understand that such glitches can happen to any company.
The communication could have been better, but then again, I always keep in mind that there are people out there waiting to exploit every little bit of information they can gain.
That said, I decided to give YPN another chance over the past few days. And I've been pleasantly surprised.
If I take ads off useless pages then page impressions reduce and the Page CTR improves. If I reduce the number of ad blocks per page, then Ad Block CTR and Ad CTR improves (assuming I can keep the same level of clicks). I also control ad impressions and therefore CTR by not showing ads to users who bounce in and out of my pages from an image search.
The exception is Link ads where the Ad CTR is unaffected by the number of Ad Link units.
I am working on the assumption (which could be horrendously wrong! :)) that the most important metric for smartpricing that I can control is CTR. And that would be Ad CTR, rather than Page or Ad Block CTR.
I would be interested to hear others views on this.
Several times, most recently over the past six months, I tested the notion that removing Adsense from marginal pages, thereby raising overall site-wide CTR, would improve site-wide eCPM or EPC.
In my experience, I see no evidence whatsoever that this is the case.
So, I have restored Adsense to my (formerly) low CTR pages. I am still judicious enough to limit Adsense to just one ad block, or just one link unit, on such low CTR pages. There are a few page categories with such a low Adsense CTR (0.10% or less) that I have opted instead to display low-performing Amazon ads (that presumably give me a better eCPM than Adsense, though I can't prove it).
Anyway, just my opinion, based on just my experience.
What's going on here?
I gave ypn one of my busiest pages to test. If that comes out okay I will be soon giving them a larger share.
Talking about boiling, sheesh!
Channels are somehow linked to smartpricing and ECPM
I think it was Ann who brought this up last year and I may have been one of the people who sneered. Yet, as I've commented in a couple of threads this past week, breaking up our strongest channel (for which all the pages had individual URL channels in any case) has resulted in the eCPM on those pages returning to levels unseen since early in the year.
Anecdotal? Perhaps, but I did offer the thesis that Google may be putting trust into the webmasters ability to group pages, and giving it enough weight in their ad targetting to result in lower earnings when the pages should be targetted with different ads.
For example, say we have a channel on foreign cars. Google might think the fact we lumped them into a channel is a reason to show Porsche ads on the Yugo page. Not an overpowering reason, but enough to impact eCPM in the long run, maybe through negative word matching for the channel or something similar.
If it's not the case, no biggie, not a lot of work to delete the channel code from a handfull of pages or to put it back in.
Is there any proof of this? Channel data would seem an odd thing to include in a smartpricing algo. People use channels to monitor so many different things: individual pages, page groups, ad block types, ad block position, links or content, etc. And they are changing all the time. I am working on the assumption (which could be horrendously wrong! :)) that the most important metric for smartpricing that I can control is CTR.
I think you may have already answered your own question.
CTR (Click Through Rate) is not an odd thing to use in determining Smart Pricing. In fact it is one of the most logical stats to use.
Why? Smart pricing is pricing based on conversions. Low-converting sites get ads with lower ("smart") prices. A click is the first stage of a conversion process. The visitor obviously has to click on the link before going to a site that attempts to get a sale/whatever ("conversion").
It's been discussed many times in this forum in the past, that a high CTR is likely to get you smartpriced very quickly. The premise is that site visitors click on the Adsense links not realizing the links are ads. Hence the clicks are not going to lead to conversions.
Whether or not you use channels, Google can still determine the CTR for your site. I don't think Google smartprices some pages only (yet); it seems to smartprice the entire site. Thus it does not need to "spy" on your channel data to smartprice you.
However, if Google now or later tries to make its smartpricing more smart, obviously it will look at conversion rates for pages in particular, not the site in general. To do that it could collect a pile of new data streams or borrow the stats from channels already stockpiled (which you created).
So my conclusion right now to explain why a number of people who've had stable data for months/years suddenly saw steep drops (around the time the channel data Google admitted had a glitch) is either:
1) a very severe smartpricing hit due to a radical change in the smartpricing algo; or,
2) a glitch that causes under reporting of income in our control panel (which will be adjusted upwards when the glitch is fixed); or,
3) Google is trying to pull a fast one by radically changing the payout structure (commissions) on some accounts.
I don't buy the trite, flippant theory that 'this is just some people having a bad month.' There's too much circumstantial evidence to believe that.
Due to the obvious glitch that: a) several independent users got data to prove, and, b) Google confirmed, I think it's more accurate to conclude it is Google that is having a bad month.
edit: No, it is probably this instead: [webmasterworld.com...]
Then the channel misreporting started. Since then, not only did that top position get slashed by 2/3rds, but my whole site has been seeing less earnings than any time in the past 3 months.
The only thing predictable about Adsense is it's unpredictability...
Fortunately (I guess) the two channels just happened to be separated by directory structure. One of the channels appears only in one subdirectory, so I put a URL channel on the whole domain and a URL channel on the subdirectory. In theory this should track very nicely (whole domain - subdirectory = 2nd channel.)
Well, the upshot is that custom channels are still broken. There are missing clicks and earnings from the custom channels.
When I go to "Advanced Reports" and I select all custom channels, the earnings are lower than when viewed in aggregate. I've never used an ad that didn't have a custom channel associated with it.
So yes frakilk, custom channels are still broken for me too.
I only hope that they didn't break custom publisher-IDs, too.
[edited by: RonS at 2:21 pm (utc) on Nov. 1, 2007]
I pondered the suggestion that the two events were connected and removed the 10 digit channel code from the channels that seemed most affected--no change.
Last night I got very frustrated and not only removed all the channel codes but went to the AdSense setup and deleted ALL the channels.
It is just a small data point so far being only nine hours into the day but eCPM has recovered to the September level. CTR seems a bit low still but is within the pre October range. I do hope this holds, money is a sweet consolation for having no channel data.
Did you delete them entirely or just dissable them?
AdSense Reports just updated and so far so good, happy days are back.
Based on a consistently bad eCPM from Oct 17-31, and a consistently normal (i.e. pre 10-17) eCPM over the past 14 hours my advice to anyone experiencing eCPM probems is to opt out of channels until they are guaranteed fixed.
I appreciate that ASA is grateful for my patience but frankly my patience ran out long ago. I was on the verge of dumping AdSense altogether until I decided to try this one last thing which seems to be working fine. It does disturb me that this problem has dragged on this long but even more so I am angry that AdSense has denied that there even was a problem until forced to admit it by the evidence.
In case you all didn't see our blog post a few minutes ago -- our engineers have been working to fill in all of the channels data that was misreported last week. You should see the correct data in your accounts by early next week.
Thanks again for your patience!
Now I cannot possibly know what others are experiencing, perhaps I'm just benefiting from the changes mentioned by ASA. But since the marked improvement began precisely coincident with the deletion of all the channels I suspect that is it. I'll begin replacing the channels when people stop complaining about eCPM issues on this and other forums. Yes, I know there is a constant background of complaints, these past two weeks are anomalous and I can adjust.
But the increased revenue from just today is worth the effort needed to replace the channels--when I do. I have all the past documentation and will update and redesign the channels when I replace them.