Questions about common Adsense tips

[silverbytes]

First of all, how do I determine what keywords are making money in my adsense sites? Is that something I can directly notice on adsense reports? I see there how much criterias pay or urls, but how do I know what page, keyword or even ad is making money?

Second, I've heard many times as a tip to avoid putting lots of ads in one page because that makes ads pay less money. Is that really true? 1 only vertical skycrapper showing 4 ads may earn as much as 3 vertical skycrappers showing 12 ads?

[ArtistMike]

First, you can measure your keyword success by putting a channel on pages that have one topic, and then watching that channel to see how it does.

Second, more ad blocks let in cheaper ads so they have the possibility of getting clicked instead of the more expensive ads.

There are page trackers that will tell you which ad was clicked on. Google "AddFree Stats" ... see what you get.

Mike

[edited by: ArtistMike at 1:02 am (utc) on Oct. 16, 2007]

[joelgreen]

1 only vertical skycrapper showing 4 ads may earn as much as 3 vertical skycrappers showing 12 ads?

If placed correctly. One correctly placed ad block may outperform 3 badly placed ad blocks.

[Scurramunga]

I've heard many times as a tip to avoid putting lots of ads in one page because that makes ads pay less money. Is that really true?

From my experience one ad on a page delivers best performance. I also lkeep the number of pages displaying ads to a minimum and make sure that the pages containing adsense are easily accessible from all other pages on my site.

[silverbytes]

I've noticed that one of my best sites about travel, that has few adsense ads is paying 0,05 per click. While others (about travel too) has maximum of allowed ads, and paying 0,04. There is not much difference so I wonder if worth take off adsense blocks in sites paying 0,04...

Is there any other way than experimenting taking off ads and checking again how much my ads are paying now? Is a week a reasonable period to make that kind of test?

BTW I think 0.04 and 0.05 is pretty low any ideas about how to improve that?

[netmeg]

Second, I've heard many times as a tip to avoid putting lots of ads in one page because that makes ads pay less money. Is that really true?

One thing I noticed is that if I had too many ad blocks on a page, some ads would be repeated, which really looked cheesy. As soon as I removed an ad block (or two) , that problem corrected itself (probably low inventory of ads that Google thought were relevant to the page) and earnings, if anything, went up.

[celgins]

BTW I think 0.04 and 0.05 is pretty low any ideas about how to improve that?

I think that's pretty low too, but there is no certifiable way to improve it.

If you follow the theory of "less ads per page", I think the quality of ads and their corresponding EPC will improve.

[europeforvisitors]

It probably depends on topic, too. Milan, Minnesota (pronounced "MY-lin," by the way) is likely to pay less per click than Milan, Italy.

[silverbytes]

Assuming the theory works (less ads per page) how about my testing questions:

Is there any other way than experimenting taking off ads and checking again how much my ads are paying now? Is a week a reasonable period to make that kind of test?

[europeforvisitors]

The length of time required for a valid result will depend on how much traffic (and how many clicks) you have. But even then, there are no guarantees of accuracy, because other variables may come into play: e.g., the weather, events, economic news, holidays that you may not know about (if you have a global audience), and so on.

The best way to test would be to do what magazines call a "split run," with half of your page views in a given period of time displaying single ad units and the other half showing multiple ad units. That may not be practical for your site, however (I know that it wouldn't be for mine).

[silverbytes]

As somebody said: 12 ads paying 0,03 may be better than 4 ads paying more, if user don't have option to see among those 4 anything he/she wants to click on.

But I guess I have no other way to know it, than taking off those and measure income. And put it back and measure again. Discarding all other variables and assuming the amount of ads makes difference...