Google does not want me to get rich

Plotting adsense stats and CPC patterns

[Nonsense]

I have plotted some graphs of my Adsense stats, with all units removed to avoid breaking TOS too blatantly.

<snip>

A few notes.
1) The large band covers a period where I know I had external factors altering my stats, so that period should be considered less significant
2) Data is smoothed
3) Period covered is about 8 months.
4) total sample is in the order of millions of impressions

My thought is that one of the most important factors of the so-called smart pricing is keeping us from having too strong increases in earnings.

you have 4 graphs: Impressions, CTR, CPC, and Earnings.

You can see impression taking off after the initial site startup, you can see the CTR going up as a result of optimising and all the massaging being done to the site... and you can see the CPC going down and down and down , so that my earning are not changing that much.

In the first part, up to the first arrow, the impressions rise considerably, but CTR is almost constant. CPC does not suffer, and earnings grow.

But when CTR starts increasing, the CPC decreases almost in parallel. From the second arrow, it seems that any time the CTR has a small bump up, the CPC has a small bump down. Apart from the violet band, they seem almost mirrored...

I know, everybody will start speaking about conversion rates dropping when CTR grows, as a result of less motivated visitors, but I cannot take it out of my head: the smart part of smart pricing is smart in a different way from what we are told..

Also, ever wondered why G puts in our stats eCPM but not CPC?
They are both derived data (just a division), and CPC is far more interesting for us, but G does not want to give that much relevance to it...

Enough conspiracies now! I am going to rinse my brain.

[edited by: Jenstar at 4:27 pm (utc) on April 23, 2005]
[edit reason] Sorry, no links, even to 'free hosting' type pages [/edit]

[bjseiler]

I saw a post on Slashdot about a year ago where somebody had tried to scientifically figure out why it seemed that Netflix delivered slower, and less desireable DVD's (lower on your queue) the longer you were a member. It was actually pretty interesting and it made total sense. If you wow somebody with results in the first few months of a service, and then let things slide to just being good enough not to quit, then you will probably keep a high level of retention while saving money (by not having to pay out the high amounts all the time).

I am not saying this is happening as part of smartpricing, but it would not shock me if it was a factor.

[rbacal]

Unfortunately, too many unknown variables in the google blackbox to come to any conclusions about cause.

In any event, if you have all this data set up, any chance you could calculate correlation coefficients between your four variables?

At least then you/we would have a concrete number that, for example, shows the inverse relationship between, let's say CTR and CPC, and know how STRONG/LARGE that relationship is.

Now, the question is: what other kinds of causes can we come up with (excluding smart pricing, and any conspiracy junk) to explain an inverse relationship that you are describing?

[TheRookie]

All I can say is, who else can't wait for Ads from Yahooooooooo?

[Jon_King]

>>Unfortunately, too many unknown variables in the google blackbox to come to any conclusions

True save one. The majority of sites have declined in revenue thru the previous 6 months.

[ken_b]

The majority of sites have declined in revenue thru the previous 6 months.

Other than posts on forums like this, what source is there to substantiate this comment.

[Jon_King]

>>what source is there to substantiate this comment.

Yes indeed. Just posts here.

[europeforvisitors]

You can see impression taking off after the initial site startup, you can see the CTR going up as a result of optimising and all the massaging being done to the site... and you can see the CPC going down and down and down , so that my earning are not changing that much.

Stands to reason. Tweaking your layouts, ad placement, etc. is likely to boost ad clicks at the expense of conversion rates.

I know, everybody will start speaking about conversion rates dropping when CTR grows, as a result of less motivated visitors, but I cannot take it out of my head: the smart part of smart pricing is smart in a different way from what we are told..

Lower conversions (or lower anticipated conversions) = bigger discounts for advertisers. Sounds pretty smart to me.

Also, ever wondered why G puts in our stats eCPM but not CPC?

I assume that you mean EPC, not CPC. Google obviously isn't going to reveal CPC, because that's between Google and advertisers.

Google probably chooses not to display average EPC in your report because CPM and total revenues are more useful metrics for publishers: CPM is useful for judging the return from AdSense in comparison to other revenue streams; total revenues are useful for determining whether you're rich, poor, or somewhere in between.

[elsewhen]

let us assume that you did experience a decline in earnings specifically due to smart-pricing. why do you make the jump and assume that ALL publishers are experiencing a decline.

perhaps google just pays you by click until it receives statistically significant conversion data from the clicks that you deliver. once it determines that the visitors you send to advertisers do not convert too well, google starts to give you less.

cannot there be others that start to receive more money from google once their conversion statistics come in as favorable?

to really make the claim that you are making, i think you would need a significant number of randomly selected adsense publishers (in various fields) and crunch all the numbers together. only then, could such a claim be substantiated.

of course, collecting this amount of data is virtually impossible and against the TOS, so i doubt you will get many takers.

absent that, we have to rely on anecdotal info (like posts to this forum)... my EPC is down slightly from 6 months ago, but has been on a steady increase since the end of january.

[Oimachi2]

let us assume that you did experience a decline in earnings specifically due to smart-pricing. why do you make the jump and assume that ALL publishers are experiencing a decline.

I've never seen so many complaints about decreasing adsense earnings. This plummeting of revenues is of biblical proportion, not just your usual fluctuations.

I would say that these posts are a clear indication that there is a serious decrease of earnings from adsense. Almost factual actually.

Show me one forum that claims that adsense revenues are on the rise! And not one from 2003 either!

;)

[alika]

Errr ... our Adsense revenues have been on the rise, month after month. Does that makes us an aberration of the "factual truth" of declining incomes?

[DamonHD]

Hi,

Yes, modulo the usual wobbles, my AdSense revenues have been consistently (if slowly) rising.

Claims that AS revenue is falling *for all* are not true: I have a counter-example!

Rgds

Damon

[Visi]

Tries to get back on thread here....the initial post was is Google using CPM as the controlling factor for payment. This has been suggested earlier, as we also have seen this nice reverse relationship between CTR and EPC resulting in a flattning of CPM over a number of months.

[androidtech]

Another possibility is that, once you get past a certain number of clicks per day, you are beggining to "blow out" the advertising budgets of the high-paying AdWords advertisers that originally fed you're high CPC value.

[Powdork]

Lets not foreget supply versus demand. Suppose there are two sites, one with Adsense and one without. Your site with Adsense slowly starts to do better and better in the serps claiming some of the marketshare (impressions) of the site that does not run Adsense. As your impressions increase, the supply of advertising space for Adwords advertising increases. This can only drive the price down. If your clicks turn into impressions, this will increase your payout due to smart pricing and increase demand, which should cause your cpc to rise again. If your clicks do not convert then you face the triple whammy of
1.greater supply
2.lower payout %
3.decreased demand as advertisers back out

Clicks that convert are good for evryone.

[sailorjwd]

I do notice that the more visitors I drive to my site using adwords the more difficult it is to keep CPM up to same levels. I continuously add content and coincidently diversify advertisers in hopes of not blowing out budgets.