Does a higher average eCPM and CTR help combat smart pricing?

[JS_Harris]

Having written that title the answer seems a lot obvious, my question however is about making changes to very well established authority sites that have had adsense on them for years.

Channels indicates a couple of units that do earn money but at roughly 50% the site average of the other units. Remove them would reduce total income and increase the site average eCPM and CTR but would that provide ANY benefit at all... such as reducing the effects of smart pricing?

Advertisers would appreciate being in the prime site positions but I would lose money by removing units... tough call.

[JS_Harris]

Some numbers to illustrate what I mean, these aren't mine they are just examples.

Suppose that my stats today showed the following account wide average per channel:
1000 pageviews, 50 clicks, 5% CTR, $5 eCPM resulting in a $5.00 average per channel.

And I then removed one unit (one channel) which gives me the following figures:
750 pageviews, 45 clicks, 6% CTR, $6 eCPM resulting in a $4.50 average per channel.

As you can see a 1% increase in CTR and $1 increase in average eCPM is nothing to sneaze at even though the end result is 50 cents less in average earnings per channel.

Would the increased CTR and eCPM yield a bigger payout per click and possibly even offset the 50 cent loss?

P.S. I did check the stats both with and without a low performance channel i'm considering removing but the stats are the same either way. That's expected given that any smartpricing effects are per account and not per channel.

[swa66]

In your example, you lost 25% of pageviews and only lost 10% of revenue. It seems like a good move, but don't loose that much traffic :) .

Your CPC seems to stay the same in your scenario. That's something I don't think would happen in reality: you'll have less badly paying ads, hence the CPC would increase with having less ads.

I think there is a belief out here that "smart pricing" is a gradual gentle decrease in earnings, that can have many slow steps to "punish" a "bad" publisher.
I do not subscribe to that belief. I believe smart pricing is a big whip that when use decreases your revenue dramatically in huge steps.
I base this on both what I read here and reading between the lines in my interactions in the past with Google on the subject. But it is a belief, I can't proof it.

I'm also not sure if it is something that can be turned back all that easy.

To get your answer: to some A-B testing, over a long enough time and let the rest of us know the results ;)

I've done some testing some time ago and decided then that less is more.

[JS_Harris]

I've finalized plans to run some limited A-B testing over a period of at least 6 months, good or bad.

What I'm expecting is opposite to the example I mentioned above, as CTR average climbs eCPM average will likely fall and earnings will likely remain in the same ballpark.

I suspect that the trick to maximum earnings is to have an ideal CTR (which is likely fairly low and may fluctuate based on a gazillion factors) while churning out the pageviews.

Hopefully this doesn't cost me a fortune during the 6 months but there is only one way to find out. Unfortunately I won't be able to share whatever findings I get because of adsense TOS regarding the mention of CTR and eCPM values, I do wish I could.