Forum Moderators: martinibuster
The CPC then drops so much, that despite getting another 1000 or so clicks, the ad block is earning just a little less than it was before. Pisses me off a little.
It's not in a spot that may cause accidental clicks, it just has a nice spot.
This happens ALL the time.
When traffic increases, clicks increase, and so CPC goes down. It's bloody ridiculous. #*$! is going on with their damned algo's.
#*$! is going on with their damned algo's.
The one thing for certain is that they don't know!
If you send more visitors and make less money: undo the change, keep the visitor.
It's almost as though Google have a cap on your daily earnings
Isn't it possible that, when your traffic increases suddenly, you're fed a greater number of lower-bid ads? Presumably Google has some kind of allocation formula to spread the better-paying ads around--not to ensure fairness, but to protect advertisers and the diversity of the network.
Search Google "adsense earnings cap" or similar. Your not the first to theorize this.
I have noticed it too with my sites. I thought it was me being paranoid.
It's almost as though Google have a cap on your daily earnings
Isn't it possible that, when your traffic increases suddenly, you're fed a greater number of lower-bid ads? Presumably Google has some kind of allocation formula to spread the better-paying ads around--not to ensure fairness, but to protect advertisers and the diversity of the network.
Everybody is aware that Google AdSense program lacks transparency. All we can observe is our total earnings and dice that information up with channels. We can further analyze by observing contextual or placement type ads performance. Effectively, we have symptoms with very limited ability to analyze the root cause of our earnings symptoms.
Conjecture and speculation is unavoidable and normal for one with a healthy mind in this situation. Everybody could be right and wrong.
As a side note, without the ability to do a REAL root cause analysis on earnings declines or increases, publishers (businesses) cannot plan or anticipate the future. Without an accurate business vision, websites (like mine) cannot hire, project, or make other important decisions BASED ON ADSENSE REVENUE.
[edited by: Edge at 1:43 pm (utc) on July 29, 2008]
AdSense has never billed AdSense as being anything more than a way to "earn money from relevant ads on your site." The concept of "Stick our code on your pages, and we'll send you money every month" obviously works for a lot of Website owners, but it may not be attractive to publishers who demand a level of control and behind-the-scenes transparency that go beyond what the product is and what Google promises.
Side note: AdSense has worked pretty consistently for me over the last five years, so I continue using it. Several other ad networks and affiliate programs have been less productive or reliable, so I've abandoned them. Having the willingness to fire partners who don't meet your expectations can save a lot of frustration and improve overall RPM.
I don't think there's a daily cap; I have seasonal sites that will earn .30 one day and $300.00 on another day.
No, you can't plan or make important decisions based on AdSense revenue; the mistake would be ever thinking you could. You don't have enough control over it, and you don't know what either Google *or* the advertisers are doing, have planned, etc. If you don't consider AdSense to be incremental income, then you are operating at great risk of having it lowered or disappear (as we have seen over and over again) That's the unavoidable payback for the ease of a 'set-it-and-forget-it' solution, such as AdSense.
I guess Google tries to "smooth" earnings per day (over-earnings from one day get passed to low-earnings day).
which is exactly what happened yesterday to me. some call it "smoothing". others call it a "cap". whatever it's called, it's made of glass.
I guess Google tries to "smooth" earnings per day (over-earnings from one day get passed to low-earnings day). Just a theory.
Yes, that's what I think too. I just pulled up 3+ years of data into a new Excel sheet for further analysis. Interesting.
Saturday is my lowest traffic* day. -17% compared to weekly average. -14% when I look at the clicks. Yet, on Saturdays EPC is up +5%
Sunday is my average traffic day. -0.5% compared to weekly average. +5% clicks. EPC is +2.4%
Mondays and Tuesdays are pretty much identical, the strongest days of my week. About +6.5% compared to weekly average traffic each. Yet the main difference is clicks. Mon: +10%, Tue: +5%. But EPC is down -4% for both days. Huh?
On Wednesdays it gets slower (traffic still +5.6%, clicks +2% compared to weekly average), but the EPC depression goes away (-1%).
Thurdays - even less traffic (+4.5% compared to weekly average), average clicks (0%), and even less EPC depression (-0.5%).
Fridays - weekend begins, traffic is out-of-the-office (-5.4% compared to weekly average for traffic, -8% for clicks), and -hoorah!- EPC is back above average (+1.5%).
Probably, EPC, Clicks, and traffic are linked to each other, but not in a way we would have imagined. The more traffic you bring, the lower your EPC will be. To which degree, we'll have to do some more analysis. But from our data, it's clear to me that higher traffic does result in lower EPC, and vice versa.
Now on to the glass ceiling theory. By now, I do not believe that there is a static glass ceiling. I think Google pre-determines the amount you can earn on any given day based on historical data. For example, when your past 200 days average is X they say - this guy can earn (or even WILL earn) Y=f(X) -- That would explain why site outages seem to have no/very little effect on total earnings. Needs further research.
* traffic is "adviews", i.e. the data provided by Google, not my on site traffic from log files
Now on to the glass ceiling theory. By now, I do not believe that there is a static glass ceiling. I think Google pre-determines the amount you can earn on any given day based on historical data. For example, when your past 200 days average is X they say - this guy can earn (or even WILL earn) Y=f(X) -- That would explain why site outages seem to have no/very little effect on total earnings. Needs further research.
agree. which makes it all very frustrating.
As for the suggestion that higher traffic brings a lower EPC, my own numbers say otherwise, although I can see why it might occur in some cases. (See my earlier post about ad allocation: If there are only so many high-priced goodies to be shared, it's reasonable to assume that Google wouldn't let any one site absorb a disproportionate share of those goodies because of a traffic spike.)
On the one hand, there's "smoothing" where "Google predetermines the amount you can earn on any given day based on historical data," and on the other hand, there's "extreme volatility"? Which is it?
Please quote correctly. I wrote:
"extreme volatility of some of the Adsense metrics"
I see exactly that, while the resulting output (i.e. earnings) seems to be smoothed out. It looks to me as if the output is determined by other factors than just advertisers bid and click (and, say, geolocation of the user).
I know that some believe that Adsense is based on supply and demand. So let's compare it with one of the purest markets around, forex. Let's determine the volatility in this market by looking at the USD-EUR exchange rate.
The variance coefficient determines the randomness of a given data set (pls Google for further info). The lower the variance coefficient, the less random (i.e. the more consistent and predictable) is the data.
Between May 2005 and October 2007, I saw variance coefficients between 0.56% and 1.97%, based upon daily rates provided by Oanda. Not very volatile. This is a good and not too random data set.
Now, let's look at my total daily earnings: in the same period I saw variance coefficients between 25% and 50%. This data set appears to be RANDOM. (A statement that many Adsense publishers would sign right away.)
eCPM: 23% to 68%. Also RANDOM.
This is confirmed by a closer look into the daily data:
EPC
30/06 X
01/07 60% of X
02/07 X
That's random, man. One day a click is worth X, the next day it's worth 60% of that, and the following day X again? Com'on.
eCPM
26/07 X
27/07 175% of X
28/07 150% of X
Random.
My site did not change. My traffic sources did not change. All that changed were the metrics displayed by Adsense.
Question here - why does the daily EPC and eCPM data appear to be almost random, when the average total earnings (e.g. 30 day average) move smoothly in very predictable tight band? There seems to be no connection between the underlying click data and the amount of money earned.
And thanks to Google's beautiful black-box we have no chance to know what is causing this, or how to improve earnings to a point that the whole program actually makes sense. YOU CAN NOT BUILD A RELIABLE BUSINESS IF YOU ONLY RELY ON ADSENSE.
Well, at least I have no chance, because I do not want to build MFA sites, or scrapers, or other shady schemes.
When traffic increases, clicks increase, and so CPC goes down. It's bloody ridiculous. #*$! is going on with their damned algo's.
That's why I only display one ad on one page of my site. Adding more ad blocks is an exersise in futility for me. One adblock means less monitoring, less MFA culling and less loss of visitors. CPC is also a little more stable, most of the time.
That's random, man. One day a click is worth X, the next day it's worth 60% of that, and the following day X again? Com'on.
I'm enjoying this thread:-)
I must concur I'm seeing similar at present, as if you had not already guessed!
YOU CAN NOT BUILD A RELIABLE BUSINESS IF YOU ONLY RELY ON ADSENSE.
Have you considered the possibility that Google may not be interested in making life easy for people who build businesses around AdSense, and that the network's long-term success may require culling the herd through attrition, if not through overt action?
Have you considered the possibility that Google may not be interested in making life easy for people who build businesses around AdSense
Of course. That's the whole point of this thread.
On one end there is an intransparent service that pays whatever it wants to their ad space suppliers (we're not partners, I understand). On the other there are people who create web presences and fill that space with content. This costs money. If it is high quality content, then these people need to make decisions on how to produce the content and how to monetize.
This thread sheds some light on the question how reliable Adsense as a customer is (if we are suppliers, not partners, then Google is our customer, not a partner). And there seems to be signs of Adsense being a very unreliable partner. Yes, they pay their invoice on time. But the amount on the invoice is 100% up to them. If they tomorrow decide to pay half, then they do not have to communicate that. They do not have to ask. They just pay half.
Yes, yes, I know that anyone can leave Adsense at any time (heard that argument before here), but still this thread can act as a WARNING to those who think that they can build an honest content business around Adsense (not MFA, but honest web content). They can't. And this thread explores why.
I fully understand that Google has no interest whatsoever in "making life easy" for anyone. They have to earn money too, which means: they have to squeeze as much money out of publishers as possible. That's why the black box is so convenient for them.
I fully understand that Google has no interest whatsoever in "making life easy" for anyone. They have to earn money too, which means: they have to squeeze as much money out of publishers as possible. That's why the black box is so convenient for them.
They don't need to squeeze as much money out of publishers as possible, they need to attract as much advertising money as possible. That's why it makes sense for them to have a "black box": to ensure that AdSense works for advertisers, not just for publishers.
That said, I don't believe that this:
That's random, man. One day a click is worth X, the next day it's worth 60% of that, and the following day X again? Com'on.
is true at all. It's not random, it's just plain MATH. What that click is worth at the time it was clicked on depends on a multitude of circumstances that were occurring at the time of the click - who clicked on it, where they came from, where they clicked on it, what the competition was at the time, the advertiser's budget, the quality score in the advertiser's account, how well-themed his content campaign is set up, the vagaries of the way Google matches keywords with content, the placement of the ad - and you might as well say the phase of the moon, because there are probably at least half a hundred other elements we can't even guess at. So no, I don't believe it's random at all - after all, I'm an advertiser, and I can advertise high cost scientific equipment and in a single day pay 25 cents for one click and $7.00 for another click on the exact same keyword - it's just that the circumstances are different for every single click.
Heh - think of clicks like snowflakes - no two are alike.
just got a snowflake dump.... coool...
As far as the X's go, I see X then I see 10 times X. Then back to X or somewhere inbetween. No rhyme or reason to it given the data that we can muster out of the reports and channels. I haven't worried over it since, well I can't remember when...
Even if this is unintentional on Google's part it makes me angry and very eager to dump them for something better. Unfortunately there is nothing better and they know it.
