I've seen similar behavior in the past: reducing the number of impressions improved things. I did monitor it like a hawk for a while to try to figure it out and all I came up with was a nearly impossible to proof theory that one can use up the good ads in their niche if you dominate the niche. Not using it up means you get less bottom feeders and can show better ads all day long.

Been a while since I played with it.

Living outside the US means no YPN as an alternative (I use affiliate links instead)

Something of that sort worked for me with Kontera. When I had Kontera along with AS, my revenue was way up, now when I removed Kontera it's lingering at the bottom. Strange!

It seems AdSense rewards change. When you reduce impressions, something triggers to give you more value. The question is whether it works as a long term strategy.

Over the past 3-4 weeks I've been experimenting

If you've done this experiment just once, you'll need to repeat it several times to exclude alternative reasons. There are too many variables and the system is far too complex to be able to judge from a single experiment. It could have been just a coincidence.

So if you've done it only once, I'd do the following to confirm the results:

May 1-14: AdSense only
May 15-30: YPN/AdSense
June 1-14: AdSense only
June 15-30: YPN/AdSense

So if you've done it only once, I'd do the following to confirm the results:

May 1-14: AdSense only
May 15-30: YPN/AdSense
June 1-14: AdSense only
June 15-30: YPN/AdSense

Who knows what constitutes once? Maybe it's rotating every day for a year, or rotating every week for 6 months, or rotating every 12 hours for 2 weeks, etc. Who knows?

I simply pointed something out that I thought others might find interesting and might want to experiment with. As I wrote before, YMMV.

FarmBoy

Who knows what constitutes once?

That should be easy to answer: If you've tried the YPN/AdSense combination for the first time and haven't switched back to AdSense-only since then.

I simply pointed something out that I thought others might find interesting

It is very interesting. But to make it more than interesting (i.e. useful) it needs to be verified using scientific methods (i.e. the experiment is repeated enough times so as to make alternative explanations unlikely).

AdSense appreciates their limited exposure more and shows better ads, etc.

No big mystery here as AdSense promotes the highest paying ads first to make the most income for themselves.

Therefore, showing fewer overall ads creates a backlog of higher paying ads that become a priority to display.

That brings up a question... are link units and ad units competing from a single source? I notice that ad units are usually *extremely* targeted (hair braiding sites on my Braiding Basics page), while the titles on the link units, while still targeted (at least to the site as a whole), seem to be more broad in relation to the page content (Prom hair, updos, etc., on my Braiding Basics page). If they come from two different sources, wouldn't you be less likely to exhaust the good ads by having both in a given number of ad slots on a page rather than one or the other? And if the opposite is true and they do come from a single source, do you compromise better ads in the ad units on the page by having a link unit at the very top above the fold?

No big mystery here as AdSense promotes the highest paying ads first to make the most income for themselves.

That depends on geo-targetting, performance history, and other factors. Sometimes it is wise instead to wait with displaying high-eCPM ads and display them later on during the day at the right time (for example, when people in the EU return from work).

I have noticed removing Adsense from pages that are ignored or opened often (e.g. inbox, writing a message), increases your CPM a lot. A good CPM draws more advertisers to your website and more good ads so you revenu goes up although you show less ads....intially. After 3-4 weeks the CPM goes down again. At least my observation.

display them later on during the day at the right time (for example, when people in the EU return from work).

Now that's an interesting idea. I'm going to have to do some thinking on that one.

A good CPM draws more advertisers to your website and more good ads

How do you know?

How's your experiment going now, farmboy?

I'm surprised that this isn't getting more attention. I consider removing AdSense from low CPM pages essential to maintaining decent CPM rates in AdSense.

It's not too difficult of a theory to prove, either. Every time I have a page that gets Digg'ed, not only does my CPM decrease due to the high number of non-ad-clicking visitors, my overall revenue goes down, indicating to me that low CPM on one page will hurt CPM rates on the entire site. Now When I have a page that gets Digg'ed, I remove all AdSense ads from that page.

Can anyone else report similar experiences?

How's your experiment going now, farmboy?

It's still going well.

And now there's something else to consider - over the past few days YPN has been outperforming AdSense.

FarmBoy

I like competency, when are we going to have YPN in Europe?

I'm surprised that this isn't getting more attention. I consider removing AdSense from low CPM pages essential to maintaining decent CPM rates in AdSense.

It's not too difficult of a theory to prove, either. Every time I have a page that gets Digg'ed, not only does my CPM decrease due to the high number of non-ad-clicking visitors, my overall revenue goes down, indicating to me that low CPM on one page will hurt CPM rates on the entire site. Now When I have a page that gets Digg'ed, I remove all AdSense ads from that page.

Great point dataguy.

I have not collected enough evidence of the relationship you describe, but I imagine this is taken into account in terms of dimishing a site's overall CPM, since all these "quick-hits" rarely produce clicks or conversions-- as people are in nanosecond browsing mode, just as many of those on social networking sites are in talk-about-myself mode... I think the best audience consists in those who are in "search mode" and that is easily demonstrated/analyzed by tracking...

farmboy

Are you using Google Ad Manager to rotate your Ads? There is an optimizer, sounds a little crude though.

Have you considered Ad Manager and ruled it out for some reason?

My guess is for now you've coded your own rotator server side?

My concern for Ad Manager is getting hooked, and then finding a price tag.