Forum Moderators: martinibuster
Get the stats for the whole 2006
A) Order by impressions, see the 10 days with highest impressions
b) Order by eCPM, see the 5 days with lowest eCPM
How many of b) you find in a)?
in my case, large-ish site and very static, it's 5/5
[edited by: frox at 8:38 am (utc) on Sep. 15, 2006]
2 out of 5 from (b) in the ten from (a). The sample was over a seven month period from a small-ish site with about 1,000 uniques a day.
But even if I had 5/5 'd still doubt it was due to an earnings cap.
More likely there is something else causing a significant increase in traffic, and the extra traffic doesn't convert the same way as the regular traffic.
Assume my regular 1000 visitors a 50% CTR for may usual Ads - I get my "usual" eCPM - varies a little according to the actual CTR and the value of the actual Ads on my page.
Then somethin happens causing, say an extra 1000 visitors - but these extra visitors have a different route to my site - they are looking for differnt things and only have a 10% CTR. Therefore my eCPM will reduce - because of the extra traffic behaving differntly.
What happens (IMO) is that there is something in the payout formula that smooths out daily earnings. This is the reason why (in my case anyways) eCPM is always highest in the morning and gradually falls throughout the day. But if the increase in traffic sticks, you would set a new baseline for daily impressions/clicks and the so-called cap would disappear.
In an attempt to understand how this all works, I've been keeping records of my reports several times a day. After looking them over, there have been 11 days this month that had a lower eCPM at the end of the day, and 3 days that ended with a higher eCPM than it was early in the day.
something in the payout formula that smooths out daily earnings
You got me curious Jomaxx, care to elaborate?
eCPM is always highest in the morning and gradually falls
I used to have it exactly this way when I started years back, now my eCPM grows though the (Google new day) gradually, reaches a peak and slopes slowly in the last 8 hours of the (Google day). The difference between now and then is less ad units per page + referrals where bad performing ad units used to be.
Added: I too now believe there is no "earnings cap"
[edited by: Hobbs at 4:56 pm (utc) on Sep. 15, 2006]
It's basically a zero-sum game in that you'll come out ahead on slower-than average days and pay a penalty on heavier-than-average days. I don't see any logical reason for Google to do this, except perhaps that they feel that smoothing out the peaks and valleys will keep publishers dumb and happy.
First: It is an unstable market place
Advertisers walk in and out with high frequency, and those that stay as carguy says cause waves of earnings not a steady income.. > instability for publishers, causing low loyalty and if publishers too are rocking the boat walking in and out with high frequency, the whole system might shake and become too unpredictable
Enter Jomaxx's theory, a stability for publishers, if I was Google I would do it through the balancing of the high and low paying ads based on past data of traffic, performance (conversion), earnings and many other factors.
End of the day more stability for the whole network.
But I'll add my own to oppose the above now: :-)
The very nature of stability of AdSense from Google's, Advertiser's or publisher's perspectives comes from the sheer large number of advertisers first, followed by a never ending supply of publishers, at high numbers, things get boringly predictable, I know that first hand from my own site.
The less intervention on Google's part, the more this large system remains stable, that is probably why you will not see any radical TOS modifications, sudden MFA crack downs, any rocking of the stable boat is unwelcome, but you will observe slow gradual careful cosmetic touch ups here and there, case and point the new ads QS rules. Public companies can't afford to do what collage students previously could do, I miss the good old crazy Google days.
[edited by: Hobbs at 6:08 pm (utc) on Sep. 15, 2006]
My highest impression days did have good ecpm, and my lowest ecpm days were the couple of weeks leading up to the removal of the ads following the QS algo driving most of the remaining ads in my niche away.
I'm not inclined to believe there is an earnings cap personally. One thing I did notice was that any sharp increase in traffic and clicks was always met with a very fast reduction of epc. This may be due to a variety of reasons of course, but as smartpricing exists to maximise Google's profits, a sharp increase in traffic and clicks is is an ideal opportunity for them to cash in by decreasing the payout whilst not increasing the smartpricing discount to advertisers.
I was just meaning that increases in traffic and clicks (sharp ones but imho gradual ones too) leads to a lower eCPM.
On this point, there seems a rather general agreement..
Actually the reverse is true: peak traffic days also resulted in peak eCPM.
But the less targeted visitors I was getting were much more click happy than my regular visitors.
I have been adding pages religously for years, and even with diversity, my earnings barely go up. I can put ads on a selected group (content catagory) of pages and watch my earnings tumble. If fact, I am working harder than ever to just keep my historic eCPM.
My lastest strategy is to change out the Adsense ads with YPN ads. My over-all earnings is rising as well as my ECPM.
My estimation of what is going on is that there is a number triggers which change our payout share.
[edited by: Edge at 5:38 pm (utc) on Sep. 17, 2006]
Guess your experiment is a flop.
Guess your experiment is a flop.
Guess not.
People should not be discouraged from posting about any patterns they observe and asking for feedback from others here.
Very often, by people posting patterns they can't account for, getting feedback from others and debate we do learn more about the intricacies of how adsense works. I suspect Google learn things they don't know too.
To give you a recent example, Ann started a thread where she discussed the results after she removed all channels from her site and the earning shot upwards. Yes, the thread sparked some debate but in the final analysis we may have learned something tangentical. Specifically that if you force Google bot into re-evaluating your site as a whole by major changes then there is the possibility of Google's valuation of your site going up - hence better EPC. Maybe even better paying ads getting placed.
This experiment may on the face of it not prove that there is an earnings cap, but it may prove there isn't. We may learn something from it, therefore discouraging people from posting by this type of comment is (in my view) wrong.
I agree.
