8 {{ upvoteCount | shortNum }}
8 {{ upvoteCount | shortNum }}

Adsense 'Experiments'

by AdamJ85
7 replies
  • SEO
    • |
Hi all,

I have an adsense experiment (using the experiment tool within adsense) running on two ad units at the moment. I'm testing a grey url vs an orange url.

In one experiment the grey url is winning. In the other, the orange url is winning. The reason is that in one test the CPC is higher for grey and in the other it's higher for orange.

But surely url colour doesn't affect CPC? Only CTR.

So shouldn't these experiments base the winner on CTR rather than on RPM, which is influenced by CPC?
#adsense #experiments
  • Profile picture of the author yukon
    Banned
    Why are you looking at CPC If your testing CTR?
    {{ DiscussionBoard.errors[9014760].message }}
    • Profile picture of the author AdamJ85
      I'm not. I'm saying that adsense declares a "winner" - and that winner is based on RPM - which doesn't make much sense to me.

      I appreciate that I can look at CTR if I want to test CTR.

      My point is that I don't know why Google would choose to base the winner on RPM, when people are testing different colour schemes, etc, where the important variable is surely CTR.
      {{ DiscussionBoard.errors[9014769].message }}
  • Profile picture of the author MikeFriedman
    So just ignore it. Problem solved.
    {{ DiscussionBoard.errors[9014774].message }}
    • Profile picture of the author AdamJ85
      Indeed. However, I'm the kind of person that faced with the following options...

      a) Google have added a feature that is flawed in a very obvious way

      or

      b) I'm missing something

      ... will assume that b is more likely to be the case.

      I therefore ask other people for their opinions.

      I mean, if (a) is correct, then think how many thousands of people are currently testing colour or font variations and basing the winner on what google says (highest RPM), rather than what matters (highest CTR)!
      {{ DiscussionBoard.errors[9014784].message }}
  • Profile picture of the author MikeFriedman
    I think people are smart enough to figure it out.

    Your sample is probably too small. If you test 1000 clicks, the CPC will probably be almost identical, so the one with the better CTR will also have the better RPM.
    {{ DiscussionBoard.errors[9014789].message }}
    • Profile picture of the author Freeboi
      Originally Posted by MikeFriedman View Post

      I think people are smart enough to figure it out.

      Your sample is probably too small. If you test 1000 clicks, the CPC will probably be almost identical, so the one with the better CTR will also have the better RPM.
      I agree with Mike. You probably need more data to have more substantial results. How many clicks did both ads receive? Sometimes when the sample is too small the result often won't make any sense at all as you have already observed.
      {{ DiscussionBoard.errors[9014970].message }}
    • Profile picture of the author Mr Lim
      Originally Posted by MikeFriedman View Post

      I think people are smart enough to figure it out.

      Your sample is probably too small. If you test 1000 clicks, the CPC will probably be almost identical, so the one with the better CTR will also have the better RPM.
      Mike was right, you need a large number for testing.
      Also CPC is based on advertiser and the niche you're on, where CTR is under your control.
      {{ DiscussionBoard.errors[9020864].message }}

Trending Topics