Adsense automatic experiments - anyone explain? - Google AdSense - Display Ads forum at WebmasterWorld

Forum Moderators: martinibuster

Message Too Old, No Replies

Adsense automatic experiments - anyone explain?

paranoid android

2:49 am on Jun 21, 2016 (gmt 0)

I enabled the new automatic experiments feature a few weeks ago, but am slightly stumped. Do I need to manually do anything for this feature to work, or will Adsense eventually start giving me hints on tweaks I could make to the set-up in order to increate revenue?

I see the option of doing my own manual experiments (I haven't done any of these) and below it, "automatic experiments" with, so far, nothing in there.

Am i missing something?

Swanny007

6:11 am on Jun 21, 2016 (gmt 0)

You don't need to do anything. So far the only automatic experiments I'm seeing appear to be comparing the font I have picked with "Default" or "Optimal".

paranoid android

11:33 pm on Jul 1, 2016 (gmt 0)

How would you guys deal with this?

I set up a manual experiment. Currently one of my ad units is set to both text and display images. So I set-up a test where it would server text-only 50% of the time and both for the remainder.

At the moment the RPM has gone down significantly for text-only, but according to Google "quality score" for the text-only ads has improved.

See screenshot: [s32.postimg.org...]

Swanny007

6:29 am on Jul 3, 2016 (gmt 0)

That's a tough one. Personally I'd stick with the original settings.

Broadway

5:17 am on Jul 7, 2016 (gmt 0)

I didn't even realize that Adsense (publishers) had a Quality Score.
So, how do you interpret your experiment's results?
1) Evidently Adsense is saying just the act of showing their Display ads craps up your site by 100%.
2) The report tells you that RPM for the ad slot decreased by 49%. An income loss for that ad slot.
But then it also says an "Overall Change" of +59%. What does the "Overall Change" term mean? Does that somehow translate into money?

ChanandlerBong

8:03 pm on Jul 8, 2016 (gmt 0)

I'm running a text/text&display experiment.

RPM is up, but quality score kills it and overall score is down 66%! So I'm earning more money and doing what google is constantly pestering me to do: but it kills the quality score of my site?