AdWords
3K members online now
3K members online now
Dive into advanced features like Remarketing, Flexible Bid Strategies, AdWords Editor, and AdWords Scripts
Guide Me
star_border
Reply

Split testing ads in hundreds of groups

Follower ✭ ☆ ☆
# 1
Follower ✭ ☆ ☆
Hi,

I've hundreds of groups in a search campaign. I had the bad idea to multiple split test a new ad in hundreds of groups at once. Now i find myself tryng to manage the results of hundreds of ads in different groups.

What would it be the most effective way to test ads in a campaign with hundreds of groups? Because if i test ad B in group 1 and ad B turns to be a better ad, i want to try it in the other hundreds of groups, but then eveything turna very hard to handle even with editor and googd excel skills i have.

Please any help will be very appretiated.

Thanks very much from Buenos Aires, Argentina
1 Expert replyverified_user

Re: Split testing ads in hundreds of groups

Top Contributor
# 2
Top Contributor

Hi Juanintera, good question!

 

I think in most cases, when an Account has a large number of Ad Groups there is still a "common theme" across all those Groups - or at least across some sections of them.  If you're an eCommerce site your Groups are probably arranged by product but you probably have a range of products that are very similar and for which you essentially use the same Ad just with one or two different words for the specific product/group.

 

In these cases it's usually possible to just test variations in one group, then apply the best test to all the other groups.  You may not be able to do this with just one Group, but you should be able to reduce the number of Groups you need to experiment with.

 

Are you using CSV imports to the Editor?  I find this the best way to juggle large numbers of Groups.

 

Jon

AdWords Top Contributor Google+ Profile | Partner Profile | AdWords Audits

Re: Split testing ads in hundreds of groups

Follower ✭ ☆ ☆
# 3
Follower ✭ ☆ ☆

Thanks Jon!

 

After I posted my question I figured out to "group" my groups so I can have gather some statistical data. Because with hundreds of groups a big chance 95% of them don't gather enough data to make some serious ad tests. 

 

But if I group the groups by common denominators as you say, like geography groups. Then I made an Excel formula to test the same ad in all the groups with the geography variation. Like "Travel to Buenos Aires", "Travel to Patagonia"... in this case I'm testing "Travel to X". 

 

After a few days testing I download the data from the groups, sum them all and average them. This way I can make a decision with the aggregated data that in whole makes a sample size of its own.

 

By the way, I should label the "original" ads and the new ones so then it's an easy task to pause the group of ads that lost.

 

Do you think is OK?

Re: Split testing ads in hundreds of groups

Top Contributor
# 4
Top Contributor

Hi Juanintera, sounds like you've got it under control.

 

Yes, I'm a big fan of labels and don't use them often enough.  This is partly because (for reasons unknown) they're not shown in the Editor which makes them a lot less useful when most of your daily work is in the Editor!  I'd encourage you to use them though, let us know how you get on.

 

Jon

AdWords Top Contributor Google+ Profile | Partner Profile | AdWords Audits

Re: Split testing ads in hundreds of groups

Follower ✭ ☆ ☆
# 5
Follower ✭ ☆ ☆
Hi guys,

I've just wanted to add that I've tried some new ads in a CPA campaign and Google distributed impressions just it was split testing it. So it seems we don't need to take so much care of the split testing now. I will continue to look at it.

Best