A few weeks ago, Google quietly rolled out AdSense Experiments. Now you can do A/B split testing on your units without editing the code on your site.
So for example, if you want to see how changing the color of your ad title impacts your earnings, you can conduct this test right from the tool.
In the video below I’ll show you how to use it. And if you’d rather read, there are instructions below.
How to Use AdSense Experiments
To access the tool, login to your account and go to My Ads >> Experiments. Then click the New Experiment button.
Now complete the form…
Next create a variation of the ad. So for example, if your original ad was “Text Only” you can now choose “Display Ads Only” under “Ad Type” to see which unit performs best.
Or if your original unit had a red title, you can select a blue title for your variation.
It’s really best to change one attribute at a time in your variation ad so you will know what works. For example, if you change the color of the title and the URL, you won’t necessarily know which change impacted the performance of the ad.
And finally, decide if you want Google to control the traffic split for you or not.
If you check the box “Let Google optimize the traffic split” then Google will display each unit equally. As Google continues to collect data and determines which ad performs the best, it will adjust the impressions to maximize your earnings.
Google recommends you let them manage the traffic split, but as you can see in the screenshot above, you can control how often the variation unit is displayed.
If you’re trying to determine which ad is performing best, it just makes sense to let Google manage the traffic allocation.
To run the experiment click Start Experiment.
Your Experiment Reports
Once your experiment has been running for a few hours, you can come back to the Experiments homepage to view two reports.
If you just want to see how the variation is performing, then just click “View Report” from the homepage.
To view a comparison of each unit, click the title of the experiment.
This will reveal a chart so you can see which unit is performing the best.
Once there has been enough data collected, Google will display a Confidence Score that reveals how confident they are that a particular unit will earn you more revenue.
Google recommends you wait until either unit has a score of at least 95% or higher before opting to choose the unit you want to run indefinitely.
Once you choose the winner, Google will then run that version of your ad.
Just keep in mind, this tool is still in Beta so there may be kinks here and there.
Happy testing!
Hi Lisa, Thank you very much for your article. I try your tips for my adsense and now I could understand why my adsense earnings run slowly. I hope i will success using your tips next time.
Hi Lisa, Thank you very much.
I love you…
I try your tips for my adsense, with research compare first, and that increase my CPC, thank you
Thank you very much Lisa. Now, I could understand why my adsense earnings run slowly. I hope next time i wil reach more money using your tips.
Hi Lisa, I did A/B testing and I got responsive ads the best earning
thanks mba lisa
Thanks Lisa, I had not tried this yet. I am really interested now.
Thanks Lisa, I had not tried this yet. I am really interested now.
Thanks a lot Lisa, i hope these are great ways to maximise CTR… I started earning on my site with Adsense program and cam across your post here that helped me a lot..I am gonna to implement this new tool….Thanku very much for sharing such a helpful post….
Hi Lisa, I intend to learn incidentally Google adsense ads in preparation for the reception. I found your article fit in time, thank you you want to write this. I will learn slowly as best I can, I will only attach when it’s ready. Yours sincerely me from Indonesia.
Thanks sista Lisa for You have shared the nice info
Now I was testing the A/B experiments like above using the customized ads and already run 3 weeks. Now I’m waiting for 95% of confidence score. All the time I see the result of the A/B experiment seems pretty good if I use the big size of customized ad, because I got increased of adsense revenue for this temporary times.
I want ask, did the 95 % confidence score is effective to forever or just for a few months? Do I need to make the new experiment every 3 months once or how many months once? Although google tell to take confidence score for the long term.
Thanks again sista 🙂
Hi, Ms. Lisa, I am interested in your blog. Interesting tutorial about the world of blogging. But there is the most interesting thing for me about Adsense. I also wrote a few blogs, then I tried to register Adsense, but had rejected dozens of times, even, possibly, does not count anymore. Ms. Lisa, if you please, I ask for help on how to get Google Adsense account. Thank you Ms. Lisa.
Thanks for sharing video tutorial about this awesome feature introduced by google adsense. I am definitely gonna use this feature and hope it will definitely help me to increase my adsense earning.
Thanks again…!
I check Adsense daily and not once have I noticed that little experiment link sitting at the bottom! I had to go back into my account after reading the post just to see if it was really there lol. This is really cool as I am always doing different kinds of split tests. Thanks for pointing this out!
Im really surprised that google didnt publicise this feature much more and not many people know abut it . Thanks for giving a look into it , was helpful..might have to look into it now
Cheers
jj from digitalgrog
Me too. They announced it on their blog but I’m not sure if a lot of people follow the AdSense blog.
Lisa, just a question; how long you have to run the alternate ad before you will get some results? In your post you mention 12-24 hours and until the confidence score on one of them is at least 95%.
I have been running a split test for 3 days now, and for me the Comparison and Confidence boxes are both empty and I can’t choose a winner either. Is this normal? I checked both in IE and Firefox to see if maybe it is just a compatibility issue with my browser, but it is the same in both.
I had 23,000 impressions in total so far (that number does show), and also in the boxes I do have the total impressions and total RPMs for each version, but nothing else. I’m not sure why is this or what to think. Any insights? 😕
Yes it’s normal. Took about 4 days for my first one to show up. I guess they want to make sure they have enough data collected.
Thanks Lisa, I had not tried this yet. I am really interested now.
That’s a pretty nice feature! At least the adsense team is showing some love for the publishers :p.
Thanks Lisa!
Thanks Lisa for another interesting and helpful tutorial.
I think that I will be running an experiment or two right away because I am still trying to figure out what type of ads perform best on my website!
Thanks again! 😀
All the best!
Carlinton
You are just too AWESOME!!!!!! Thanks for all the wonderful and helpful information!
Very interesting! I’m definitely going to try this out! 😛
Hi Lisa
Thanks for highlighting this change in the Google Adwords world
I must admit I do not use it at the moment. As I don’t get enough clicks to really warrant using it. But it is nice to know what options exist should I decide to add it to any of my sites again
ashley
Thank you Lisa for posting this! I never noticed Experiments under My ads until you posted this.
I just created my first experiment and I look forward to finally finding out if the default color scheme of text ads perform really worse than a color scheme close to my wordpress theme.
I created my experiment in less than 5 minutes flat, thanks again Lisa!
OK, I will study this. However, my foray into Analytics A/B testing has almost scarred me for life. I hope this one is going to be better (read, slightly less user-unfriendly 🙂 )
LOL, I know what you mean. Analytics sort of overwhelms me too. I found this to be much easier to use.
I can assure you this way of doing A/B testing is way easier than doing so in G analytics. Tried to set up an experiment in G analytics myself and gave up within a few minutes. You should really try this as it is quite easy to set up.
Nice tutorial Lisa!
I do have a question though that I hope you or someone else can answer.
I’m running the experiments to see if different headline colors in my ads influence the results, turns out it does quite a lot.
My ads run asynchronously. Do you know if adsense changes this if I choose a winner in the experiment ? Does it stay async, or is the standard non async code picked when picking a winner ?
Great question! I would imagine it stays async or whatever you originally chose. Not sure though.
Appreciate the feedback even though you are not sure. Thanks
Hey Lisa,
I think this is a great way to maximize your chances of getting a higher CTR. Also with the responsive ads, I think that was also a great move.
Thanks for the update.
I love, love, the responsive units.
Thanks Lisa,
I had no idea this even existed. I guess I can be forgiven if its just been rolled out over the past few weeks, but still, I should be all over these changes. Its a pain going through posts and changing all the code, so this is definitely a winner.
Thank you very much for the revelation. I have liked the article and let me go try out my experiments.