Forum Moderators: martinibuster
I'm mentioning this to belie any notion that Google has assigned any rational scoring component. While over-all eCPM is down its erratic nature is not at all like its more consistent pre-glitch behavior. Nor is it like any intentional ranking change unless there is a random factor as a part of the strategy, which would indicate some kind of deception.
While EFV does seem to have a predictable bias I think his posts are thoughtful, well considered and deserve respect, as does he.
I'm actually lost for words this time.
Mainstream corporate advertisers care more about where their ads appear (and not just in terms of ROI) than traditional CPC advertisers do, and advertiser tools such as unlimited domain filters and site-targeted CPC ads will benefit some publishers while hurting others.
When AdSense launched in 2003, anybody could slap AdSense code on a page and collect earnings from clicks that were priced the same as clicks from Google's own SERPs. Those early, easy glory days are gone forever.
My theory is that the AdSense publishers seeing lower eCPMs are being 'sacrificed' in order to make up for very high revshare deals Google has engaged in over the past year (Orange, Ask re-up, MySpace, etc).
It think shorebreak hit the nail on the head with that analysis.
Indeed, when TAC (traffic acquisition costs) have been nearly stable around 78% of Adsense from content partners, yet Google increasingly makes these private deals with, as suggested, as high as 90% payout to partners, the payout difference has to come from the non-private-deal pockets.
To me, it looks that Google has, for every financial quarter, a TAC goal, which is around 78% of content partner revenues.
They are contractually obliged to keep the private deals at those 90% payout levels, but nothing prevents them from adjusting the remaining revenues so that they do meet that percentage goal.
As you can imagine, they can adjust Adsense algo in several ways for the remaining accounts to hit the near 78% target every quarter. And this algo change does not have to hit every account, or every website/niche (I don't see it in all niches) to have the desired effect from Google perspective.
One theory would say that the latest changes have been made to meet those TAC goals for the Q4. When the financials from Adsense from content are balanced, we may see adjustments to the opposite, and threads that support that eCPM levels are returning to normal.
Is a Google TAC goal a plausible theory? Heck, that is exactly as I would do in Google's shoes. It's as important metric in the financials as any I've seen.
Because it's a "new economy" metric doesn't make it less important (I did my university thesis on the importance of these "new economy" metrics).
How can one prevent this from happening? Work your way up to 20 million page impressions per month and get into the club that has a contractual revenue share with Google.
How can one prevent this from happening? Work your way up to 20 million page impressions per month and get into the club that has a contractual revenue share with Google.
Why would they then bother to collect some small fries coming with the Google AdSense referrals to the system if the biggies can fill much of the inventory?
OnlyToday: do you have a lot of advertisers targeting your site? that appears to help.
Actually my forum name "OnlyToday" was inspired by a similar experience sixteen days ago, though not on a Friday.
Why would they then bother to collect some small fries coming with the Google AdSense referrals to the system if the biggies can fill much of the inventory?
Because the marginal acquisition costs for these publishers is near zero online, complemented with the fact that Google can adjust much of the terms of the deal for this group, not the biggies, as they please.
I think the long tail, or "small fries" as you put it, is as important advertising long tail as I've come across.
I'm primarily in travel so that's going to get hit pretty hard when the purse strings get tightened. Anyway at least there may be some schadenfreude to go around.
My advice. Buy gold, tins of beans and a gun and head for the hills!
FWIW, I'm a believer in diversification myself. Multiple revenue streams make it easier to avoid dying of thirst if one stream dries up or shrinks to a trickle. Relying solely on AdSense--or on any one revenue source--isn't a wise publishing strategy.
Conspiracy theories don't pay the bills.
Realization that Google actively manages its financials (the 18-20 adjustments to Adsense shorebreak mentioned) each quarter is not a conspiracy theory, it's a fact of life.
Why would exit strategy make immediate sense even when Google does adjustments to payouts to manage financials?
Every publisher manages their own financials, and if Google contributes less than alternative arrangements, then exiting would make sense.
Getting enough leverage (the 20 million pageviews) to cut a deal with Google directly is one way to enhance bottom line, through joint ventures, hard work, or whatever.
I've always known the advantages of diversification, but diversification has some downsides as well. When you spread yourself among several projects it is impossible to give any one of them your best and there's the risk that all of them will be mediocre as a result. New projects are always riskier and require more time than established, going concerns.
With AdSense diminishing and glitching the whole equation has changed for me. If I begin a new project the current site will necessarily get less attention. The loss of my contributions won't be that consequential, but in the aggregate I think the whole internet is diminished by the loss of smaller players.
I was in a steady decline for a while until the May clean out of arbitrage/MFAs.. In June, things were back to where they were a year or so earlier and getting better. I was starting to get excited again. In early october, I started to notice a decline to the point where checking my stats provides dark humor.
Is it possible that the 'get rich quick with adsense' group found some new formula? I know some people were saying that foreign language sites seemed to be getting hit harder. The search results are still filled with tons of useless 'directories'. My alerts for my domains are showing scrapers still grabbing my content daily. In other words, the 'get rich quick without working' crowd is alive and well.
I would think that the mfa er's would probably target the high value keywords first. That would affect some of the bigger member's sites. Little guys like me might feel some kind of secondary effect since my stuff probably would not be directly targeted.
If I get some time this weekend, I'll try to do some 'get rich with adsense' searches and see if there are any hints on new schemes(on the other hand, maybe I'd better not since google might think I joined the darkside and I'd lose my buck a day)
The one real downside to this theory is that too many people got hit with a statistically improbably 50% reduction which does defy logic
cg
The one real downside to this theory is that too many people got hit with a statistically improbably 50% reduction which does defy logic
But are you seeing a 50% reduction in earnings per click?
If you are not seeing it in EPC, but only in CTR and ECPM, then the tin hat theory would lead me to believe they are reducing the amount of clicks and/or gauging a valid quick differently.
I was saying to my Adsense rep that for the first time ever our banner advertising outstrips Adsense per unit.
One good thing about this is that it means Adsense is less important to us. The bad thing is that it means we have had to run faster to stand still...especially with the dollar dying a death and set to get worse.
It seems to me Google are too busy coming up with daft additions to Adsense instead of concentrating on the main event. I mean next up is advertisers being able to book your site ahead of time. The problem is agencies already do this...the difference is they try and get the max the client will go for not the minimum. If Google want us to open up the rest of our inventory to them then they need to be less opaque.
I'll be glad if we can extract ourselves from the Adsense dependency by the end of the year (revenue from ADsense is low enough now to try some alternatives). Not perhaps dropping them just not having it as an essential component of our revenue.
An oddity I'm seeing this week is that my ecpm drops down extremely low overnight sometimes and sometimes does not. When it's really low overnight, it never does very well the next day. When it's high overnight, it gets even better during the day.
During the Great Glitch of October, my overnight ecpm was was consistently dismal with a dismal daytime ecpm to match. So, maybe it was a shift in the pool of advertisers, a shift in the way my keywords were valued, or a shift in the way my site was valued. Maybe is was an algorithm glitch and some combination of factors triggered it. Since the problem seems to be working itself out for several of us at the same time, and since it hit such a diverse group of sites all at the same time, I tend to think it was a genuine glitch.
In any case, the whole situation has certainly made me more aware of how unstable Adsense can be. I'm devoting much more time to seriously managing my site and I'm going to I devote even more time to developing the non-Adsense areas of it.
[edited by: Forest_Dweller at 1:26 pm (utc) on Nov. 17, 2007]
Traffic is coming back after experiencing a drop in the serps for no apparent reason (I don't do black hat), ctr and ecpm about the same as when I first began adsense. I give them close to a half a million PVs
a month so no I am no 'biggie' and if I wasn't diversified I would be hurting, income wise.
That's my story and I'm sticking to it. :)
Ann
But I've been running over 20% ahead of 2006 all year, so that makes the November 07 numbers stand out even more.
I thought its impossible, but believe me it is.
Usually, this number of page impressions on Myspace site generate at least 50 (thats small number, but lets take minimum ) clicks, and then suddenly niether one.
Adsense click tracking is silly these days.
I just hope its a glitch, and they dont do that by purpose.
1) I began Adsense on a very low eCPM level ($1). For about ten months, the average monthly eCPM growth was an aggressive 7%-9%.
2) Then it suddenly changed to "moderate" growth (1%-2%). This second period lasted for another 8 months.
3) At that point I reached a long-term eCPM "roof" (a round value BTW), and the "growth" now changed into -1% to -4% per month, i.e. decreasing. And this has been my situation for 14 months. This long-term eCPM decrease looks like a straight line, with only slight deviations then and now.
Looking at these long-term charts, I conclude that for my sites eCPM seems to be pre-determined by Google. The eCPM values appear to be changed probably by "triggering" something (e.g. by reaching a pre-defined value).
I further believe that the real payout is based upon pre-defined eCPM and traffic, with payout spread across all clicks that occur. This would mean that the actual bids on the clicks by advertisers are not necessarily connected to the payouts we are seeing. (Up until recently I was thinking that "bids" actually reflect my earnings. But the dramatic daily EPC and eCPM changes could not be explained by such "market forces", especially when advertisers remain more or less the same. It's rather the attempt by an algo to direct your eCPM towards a pre-defined value.) - This would support the idea that Google is actively managing their TAC. Clearly, Google are buying traffic from their partner sites, and they control what they are willing to pay for this traffic. And no, this does not need to be connected to clicks.
By the way, this theory supports several claims individual members have made over time:
- "I seem to have hit a revenue ceiling"
- "When traffic suddenly rises, my EPC drops to unseen levels"
- "I see heavy daily swings in earnings, eCPM, EPC while traffic, clicks, and advertisers remain the same"
the individual eCPM values have been swinging wildly. The variance of the daily values on my sites is typically between 30% and 50%.
"So what?", I hear you say.
Well, just compare this to another global "auction based" market, with millions of buyers and sellers: Forex. Look at the variance of the USD-EUR exchange rate at closing - this daily data set has a variance between 0.6% and 2.0%
Now, what was the variance of Adsense eCPM again? Between 30% and 50%? C'mon, gimme a break.
I am convinced that Google is using eCPM to determine revenue and EPC. This is probably the reason why we are seeing such wild swings. It's not market forces, it's not conversions. It's Google's estimation of the value of the traffic.
Oh, and before anyone jumps in shouting - "conspiracy theory" - no, that's not it. I am just trying to explain how I believe Google determines payouts. I am fully aware of the fact that Google never claimed that plain publishers are getting a revenue share. So, there is no conspiracy.
IMO, it can only be a conspiracy theory if Google claimed that the system works in a certain way (but secretly works differently). But no, they are surprisingly clear about that:
How do I get paid?
You get paid whenever someone on your site clicks on one of the AdSense ads. Advertisers can also bid to appear on your site on a CPM (cost per thousand impressions) basis. Both CPC and CPM bids compete in our AdWords program to ensure that you optimize your ad revenue.How much do I get paid?
How much you earn depends on a number of factors including how much an advertiser bids on your site -- you'll receive a portion of what the advertiser pays. The best way to find out how much you'll earn is to sign up and start showing ads on your web pages.
(from their sign-up pages; highlighting by me)
Note how they try to avoid saying that you get a fixed portion of what the advertiser pays. Sure, you get a portion of the payment (which is logical, as Google does not want to lose money). But they also do not say that the portion you get is directly linked to the amount an advertiser bids. It's just one of the factors.
And I believe that the influence of this factor is rather small, at least for the average small publisher.
By the way, there's a thread featured on the Webmaster World home page that may be of interest to publishers who rely on e-commerce ads. It's about how the holiday season is shaping up for e-commerce sites, and it has implications for what vendors who don't have war chests can afford to spend on direct-response advertising:
[webmasterworld.com...]
You're suggesting that Google "predetermines" your eCPM, and that the other numbers are juggled to maintain that predetermined eCPM.
Yes, the analysis of my stats suggests this. I would want to avoid the term "juggling", because this indicates that something questionable is going on. I'd prefer to just say that revenue and EPC is determined based upon eCPM and not vice versa. But if you want to call it "juggling", I am fine with this.
How would Google determine what was a reasonable eCPM for each individual publisher? More to the point of this thread, how would Google decide which publishers to hit with a 50% or higher drop in eCPM in November and which ones to leave alone?
That is a very good question, EFV, and it is best answered by Google. (Clearly, that would be the easiest way.) However, I can not get rid of the feeling that answer will not come any time soon.
My take? I believe that "something" actually triggered the 50% reduction, and it was probably indeed a glitch that this reduction was executed with merciless accuracy. Was it a coding error? Or done to a specific group of publishers? A/B testing? We can't know. That's the beauty of a "black box" system.
Also I remember a thread that discussed a screenshot from an (supposedly) internal Google source where site specific information appeared next to SERPs. Along with $ amounts and industries. Could it be that this information is another factor in the formula to determine eCPM?
Anyway, I guess someone at The Plex must have a good snicker about this thread (and similar threads), probably thinking: "Fools! In fact, the system works entirely different! If only you KNEW what I know...!"
It's important to remember that contextual CPC text ads (meaning most AdSense ads) are like direct-response ads in magazines or on TV: They're a niche product, not the main event, and the supply of contextual ad inventory may well be exceeding advertiser demand for some topics, putting downward pressure on prices.
The fact that Google bought YouTube, which is much better suited to branding display ads than to contextual direct-response text ads, should send a strong warning message to anyone who thinks that publishing an "AdSense site" is a viable long-term business model. AdSense is a useful source of revenue, but it works better as a side dish than as the main (or only) course.
EFV writes:
What's the common factor among publishers who are seeing their AdSense revenues dry up?
There probably is one determined by earnings level, but there may not be. When this sophisticated diminishing payout algorithm (not a conspiracy) was developed years ago the existence of a thread(s) like this one was no doubt anticipated and plausible deniabilty was built right into it. It worked well for quite some time but probably recent revenue demands from on high, or other factors broke it. Here we are. Yes, I am certainly open to hearing other explanations or refinements to this one.
I might add that my earnings are now almost back to pre-glitch levels and holding nicely. There could be any number of explanations for this consistent with zett's analysis.
"Fools! In fact, the system works entirely different! If only you KNEW what I know...!"
ROFL:-))
Thoughtful insight guys, a pleasure to read.
I decided a couple of weeks ago to leave everything alone and just see what transpired and it has been most interesting.
Bear in mind before The Glitch my eCPM was just over USD 10.00 with a year average EPC of USD 0.30+.
After going down to an eCPM of less than 4% and an EPC of 0.16 (some channels were on 0.01) I have watched with interest a steady eCPM return to almost normal levels day by day in very small increments of between USD 0.10~0.50.
Saturday saw my EPC back to just over USD 0.27 and two consecutive days of "almost average" CTR, just a bit more to go.
Am I the only one returning to "normal" figures or is anyone else seeing the same?
The last time I complained vociferously direct to AdSense it took about two weeks to return to normal, do they have a little dial that they can tweak for individual sites or do they say "Heck, we thrashed that sector a bit too hard, let's wind it back a bit"?
EFV
but I will say that I've noticed much greater strength (including steady growth in CPMs) on the dislay-advertising side than with AdSense.
Do you organise this yourself or through a third party?
