Multiple Ad Variations, Wow, I should have done this a long time ago! Options

[EGOL]

If you use Adwords you probably know that you can provide multiple versions of your ads. I have been doing this since right after it became available but was not too aggressive with providing multiple versions - I put up two or three, maybe four versions and thought that I was doing a good job.

A few weeks ago I decided to post a lot of ads for a couple of my existing campaigns, so I logged into the account and started composing ads. I composed nine different ads for some of the campaigns. Some were my standard language but I decided to try a few totally different approaches.

The results are in and Google rotated all of those ads and now I have a couple of ads that are far more effective than what I thought would be hard to beat.

So, if you are running just a few ads, go in there and try something totally different. You might be surprised at how well it works!

Take a few chances with something really creative. If you try something a little different you will probably get a small change in results.... but if you try something extremely different you have a much greater chance at a huge improvement (or a total bomb).

This post has been edited by EGOL: Mar 26 2008, 08:58 PM

[A.N.Onym]

Are we talking CTR or conversions here?

Other than that, yes. Such simple things as First-Letter Capitalization work wonders (both in ad text and URL).

This post has been edited by A.N.Onym: Mar 26 2008, 11:45 PM

[EGOL]

We are in this case talking about CTR. There is no conversion associated with this traffic.

[saschaeh]

Nice one - Thanks for the tip.

I read a couple posts about how the smallest change can make up to 40% difference on CTR.

Like adding '...' at the end of you add improve CTR. There is more on the other side of the add that they dont know.

Simple grama changes like if you say you will do something for someone as a opposed to them having to do something....


"Come make money here."


"We will Make You Money." better


Quite a lot of those textual tricks.

This post has been edited by saschaeh: Mar 27 2008, 12:02 AM

[EGOL]

Free Beer

[saschaeh]

lol - that will work a charm.

[rynert]

We are in this case talking about CTR. There is no conversion associated with this traffic.

Did Google give you a boost in your quality score and subsequently reduced your CPC whilst maintaining the same positions?

[EGOL]

Did Google give you a boost in your quality score and subsequently reduced your CPC whilst maintaining the same positions?

It is difficult to give an exact answer. This is a placement targeted ad that appears on many pages of this domain - and there are many competitors. However, my ad appears to be ranking higher and getting a higher CTR, yet my total cost per day seems to be staying about the same.

This post has been edited by EGOL: Mar 27 2008, 10:20 AM

[MulaG]

Thanks for the tip!

Do you have to pay extra for different versions of the same ad?

[EGOL]

Google does not charge for using multiple ads. You can instruct them to show all ads with equal frequency or show the best performing ads with the highest frequency rate. Some of my ads that do not perform are rarely show.

Your only cost is the time to compose and enter the multiple versions.

[_jc]

I've split tested up to 6 ads at the same time, but never tried more. Thanks for the idea. I guess it partly depends on how long you're willing to wait for statistically meaningful differences to show up.

My own method is to alternate 2 approaches. At first, I try very different copy in each ad - like you say, get very loose and creative.

Later, I take the best one (however you define that) and make an exact clone of it, with just one difference - only showing the 2 ads, for a true A/B split test. Then I keep testing these minor differences, one after another, for continuous improvement.

Usually starting a completely new ad copy comes as a result of discovering new key phrases and using them to create new ad groups. But sometimes I just get a bright idea for a new approach - especially if the CTR or ROI is not great.

I've found it useful to study most or all of the competing ads in a market niche, analyze what their intent is and decide whether to adopt their market niche jargon and "language" (i.e. to "fit in") or how to look very different. I usually go for the very different

My choice is always to rotate ads as evenly as possible - else how can you possibly do unbiased split testing?