AdWords is now Google Ads. Our new name reflects the full range of advertising options we offer across Search, Display, YouTube, and more. Learn more

Ads
4.4K members online now
4.4K members online now
Dive into advanced features like Remarketing, Flexible Bid Strategies, AdWords Editor, and AdWords Scripts
Guide Me
star_border
Reply

How to test multiple ads and choose the best

Follower ✭ ☆ ☆
# 1
Follower ✭ ☆ ☆

Hello everyone i am a bit confused on which would be the best way to do a/b testing on a set of ads that are on the same ad group.

Let's suppose i have two (or more) ads in the same ad group which differ only from the second description line:

 

------------AD 1---------

Hot Shoes For Men

Try Our New Shoes and

Run like a Pro

------------AD 2-----------

Hot Shoes For Men

Try Our New Shoes and

Run like Usain Bolt

 

And i want to know which could perform better.

At this point i have 2 options:

  1. Show them evenly over time
  2. Show only one of them for some time and then turn off the first one and enable the other one

What would be the best way to go in your opinion?

 

And how would you choose the winner between the two (or more) ads?

For example:

If AD1 had 55 % served and AD2 had 45% served would you turn off AD2 and let only AD1 as the only active ad on that ad group?

In other words: you suggest to keep only the best performing ad in one ad group? Will you give more importance to an ad that have a higher % served (and a lower CTR) or a higher CTR (but with a lower % served)?

 

Thank you very much!

2 Expert replyverified_user
2 ACCEPTED SOLUTIONS

Accepted Solutions
Marked as Best Answer.
Solution
Accepted by topic author Yari L
September 2015

Re: How to test multiple ads and choose the best

Participant ✭ ✭ ✭
# 14
Participant ✭ ✭ ✭
Hi Yari,
Correct, if the Keywords stay the same, do not create separate Ad Groups. Assuming targeting is the same for all Campaigns, you only want to have one instance of your Keywords in your Account.
In the instance you show above is the Conversion Rate the same for both Ads? If so, I'd just go with Ad 1 because of the better CTR - seems like searchers are engaging more with that call-to-action.
The only way I would not pause Ad 2 was if it had a better Conversion Rate. If Ad 2 has a much better Conversion Rate, I'd probably pause Ad 1.

Also, I think you may be interpreting some of your stats incorrectly - "So ad2 is globally performing worse than ad1 but it still works better (apparently) for 40% of the searches...correct?" --No, Ad 2 has been served on 40% of your eligible Impressions. Ad 1 was served on 60% of eligible Impressions. When Ad 1 was served, it had a CTR of 10%; when Ad 2 was served it had a CTR of 7%.
One last thing that Dave brought up originally - make sure your data is statistically significant. Don't base decisions off of one or two days worth of data. Typically you'll want to evaluate ~2 weeks worth of data, or enough Impressions to know that one Ad is always outperforming the other.

View solution in original post

Marked as Best Answer.
Solution
Accepted by topic author Yari L
September 2015

Re: How to test multiple ads and choose the best

Top Contributor
# 16
Top Contributor
Hi Yari L,

I wouldn't look at either metric. The bottom line is conversions and ROI. Let's say ad 1 has a 10% CTR with a 2% conversion rate and ad 2 has a 7% CTR with a 5% conversions, and I will assume for simplicity that both have the same or equivalent CPC. In this case, Ad 2 could be better performance.

And I would look at one other thing. Are both ad groups contributing positively to your ROI? If so, you might actually want to keep both ad groups running and try to optimize the lower performing ad. Are you profitable enough on your ads to afford to raise you CPC bid to get better impression share? Also, place at least 2 ad copies in each group to test against each other. Ad rotation settings only apply to an ad group, not to a campaign. When you are testing new ad copy, the conversion optimizer gets in the way. Turn it off and set your ads to rotate evenly. The length of time it takes to get enough activity to provide significant results should be considered whether you want to rotate indefinitely or for 90 days only.

That's my 2 cents worth. Best of Luck!

Pete
petebardo -- Deadhead doing AdWords

View solution in original post

Re: How to test multiple ads and choose the best

Top Contributor
# 2
Top Contributor
Evenly over time. If you're dictating the delivery, you're doomed to failure. You need to compare apples with apples, same time frame with same time frame.

It's not about % served either, it's about getting enough volume to make a statistically significant call.

Re: How to test multiple ads and choose the best

Follower ✭ ☆ ☆
# 3
Follower ✭ ☆ ☆

Ok Dave thanks for the quick answer.
But once i have them both running for some time (lets say a couple of weeks) the 2 main stats i will be looking at are:

  • the % served (how many times that particular ad has been shown compared to the total a mount of impressions)
  • the CTR

and based on these 2 i have to (eventually?) take a decision on which ad to keep and which one to delete or pause. This was part of my second question about the best way to choose which ad would be better than the other or eventually if it would be better to keep both ads active.

Re: How to test multiple ads and choose the best

Participant ✭ ✭ ✭
# 4
Participant ✭ ✭ ✭

Hi Yari,

To build on Dave's excellent answer - you may consider using AdWords Campaign Experiments.  Using AdWords Campaign Experiments you'll easily be able to compare two different Ad Copys, and their associated metrics.

 

To make the call on which one is best for you - you should use the Ad that delivers the best business results for you.  For my money, I usually prefer the Ad Copy that leads to the most Conversions.

Re: How to test multiple ads and choose the best

Explorer ✭ ☆ ☆
# 5
Explorer ✭ ☆ ☆

I think you can take advantage of following ad rotation option available on adwords :

 

"Optimize for conversions: Show ads expected to provide more conversions" - Considering your end goal of test is to improve conversions.This will not be a test but here you will be allwoing adword system to pick best performing ad for you.

You can choose other options too like:

"Rotate indefinitely: Show lower performing ads more evenly with higher performing ads, and do not optimize" - I will choose this when i want to make decisions on my own . Here you can run your ads for the time period you think is perfect for this test and take decisions on that. For more eveness in % served metrics i will suggest you to run for longer time period. 

Read more about this here : https://support.google.com/adwords/answer/112876?hl=en

Hope this helps

 

Re: How to test multiple ads and choose the best

Explorer ✭ ✭ ☆
# 6
Explorer ✭ ✭ ☆
Hi Yari,

Let's have a different view point. I would suggest creating two different ad groups under one campaign with different ad copies and keywords. Let's try this option for a limited period and based on the results, you can pause one of the ad groups. This would also give you insight on your keywords' performance and will put you under control of your ads.

I hope this helps.

Kapil M

Re: How to test multiple ads and choose the best

[ Edited ]
Follower ✭ ☆ ☆
# 7
Follower ✭ ☆ ☆

Ok, thanks everyone for your detailed answers.

But there is still one main questions that i can't figure out:

If an ad is performing better than the other one (in a metric that it's important for me: mainly ctr or conversions), is it a good idea to delete the lower performing ad?

I'm asking this questions because if (considering my example) 40% people search for "shoes to run like a pro" and 60% search for "shoes to run like usain bolt" AND this last ad is performing better than the other one (in terms of ctr or converisons) i will be deleting an ad that still will have a higher relevancy for the 40% of the searches.

Will i really benefit from the deletion of the lower perfoming 40% served ad?

Re: How to test multiple ads and choose the best

Explorer ✭ ✭ ☆
# 8
Explorer ✭ ✭ ☆
Hi Yari,

That's why I suggested you to keep both ads, but in different ad groups. This would help you create targeted keywords' strategy for both the ad copies. I would also suggest you to create two separate landing pages to get better idea. However, if you do not wish to create another landing page, I would suggest you to check other stats like bounce rate of your landing page in analytics for your ad copies, impressions and avg CPC. This would be helpful for you to make better decision.

I hope this helps.

Kapil M

Re: How to test multiple ads and choose the best

Follower ✭ ☆ ☆
# 9
Follower ✭ ☆ ☆

Thanks again Kapil but as Dave said You need to compare apples with apples, same time frame with same time frame. and i agree with that...if for example i want to test only a little difference in the ads (for example by inserting or not the price of something) it would not be useful to have 2 ad groups because the keywords would be exactly the same.

In my opinion, if i want to test 2 ads (or more), and only those, i must have the same "background conditions". I think that what you said is correct, but not in the case i am considering.

Re: How to test multiple ads and choose the best

Explorer ✭ ✭ ☆
# 10
Explorer ✭ ✭ ☆
Hi Yari,

No worries. However, I had suggested two different ad copies with different keywords, not with same keywords. Nevertheless, I wish you luck in your efforts.

Kapil M