AdWords: The Strangest Test

[abbottsys]

Readers of this forum will know that I enjoy testing AdWords, and pushing it to the limits. But, this is without doubt the strangest AdWords test I've ever done. I'm not sure what it proves (if anything) but I'm posting it here to see it anyone has creative ideas.

Here it is..

If you Google ivf (a top keyword for one of my clients) you'll see an ad that has www.google.com as the display URL. Of course, AdWords requires that the display URL domain match the landing page domain. And my ad does, as you'll see if you click it. The landing page is a Google SRP that contains only my client in the organic listings, and not even any ads!

So, my ad is 100% Google AdWords compliant.

Does anybody have any creative ides for using/expanding this technique?

[jimbeetle]

Neat idea, though I see another ad from your client in the sweet spot on that page, so they might wind up paying twice for that one visitor.

[abbottsys]

Neat idea, though I see another ad from your client in the sweet spot on that page, so they might wind up paying twice for that one visitor.
Yes, good point, I'm working on that. But, they sure dominate the organics.

[jimbeetle]

they sure dominate the organics
He, he, he. Yeah, and helped along by an AdWords nerd at that. Guess there's no future left for us SEO types.

[Baylow]

I can't see this having any real use outside of misleading a user and leading to a low conversion rate. I don't know, but I seem to remember that it's better practice to give the user fewer clicks between the ad and the conversion, I'm not sure why you'd want to add another google SERP in there.

[abbottsys]

I can't see this having any real use outside of misleading a user and leading to a low conversion rate. I don't know, but I seem to remember that it's better practice to give the user fewer clicks between the ad and the conversion, I'm not sure why you'd want to add another google SERP in there.
I tend to agree. Still, I'm testing it in an A/B split with my regular ad. Both ads are identical except for the landing page. I'm interested in seeing exactly how conversions compare.

[abbottsys]

I tend to agree. Still, I'm testing it in an A/B split with my regular ad. Both ads are identical except for the landing page. I'm interested in seeing exactly how conversions compare.
I love quoting myself ;-)
Anyway, the new ad is blowing away my regular ad both in terms of CTR and conversions. Testing really does pay.

[abbottsys]

Update:
I just expanded the test to include Yahoo (same landing page technique I used with Google). So now I have 3 ads running. All ads are identical except for the display URLs - which are Google.com, Yahoo.com and my client's website.

Note: These ads are totally AdWords compliant, since the landing pages for the Google and Yahoo display URLs are Google and Yahoo site searches of my client's site i.e. AdWords display URL/destination URL policy is *not* violated.

[jimbeetle]

Do you think it's the URL that has the major impact?

[abbottsys]

Do you think it's the URL that has the major impact?
It must be. The only visual difference between the 3 ads is the display URL. One is Google.com, the next is Yahoo.com and the last is my client's domain.

I'm a bit surprised Google allows this, because an ad that uses Google.com as the display URL makes it appear that Google is promoting the product! Still, I want to stress that absolutely no AdWords rules are being broken here. These ads are 100% AdWords compliant.

Thanks Google ;)

[jimbeetle]

I want to stress that absolutely no AdWords rules are being broken here.
Not as yet, but methinks you might have one named after you. How 'bout the "abbottsys rider" to the AdWords T&C? ;-)

[j0nyDzine]

LOL... holy cow Abbott... Honestly I have never heard of anyone doing this before, which as I read this completely blows me away. Pretty brilliant idea to do this... How long you think before they close this hole now that you've posted this?
Now honestly I spend about 95% of all my time with actual SEO and very little on PPC stuff, but it would seem if they didn't, the actual potential spammy abuse could get out of hand.
Is this something guys on the dark side of adwords have been doing for a while and someone that doesn't delve much into PPC don't see?
Good post on this, thanks!

[abbottsys]

... holy cow Abbott... ..How long you think before they close this hole now that you've posted this?....
I guess we'll find out. Who knows, maybe Google will like this technique.. err, but then again, perhaps not ;)
It would be very easy for them to close it anytime they want. Until then I'm running my ads in production and enjoying the benefits that come with a Google.com display URL.

Thanks Google!

[BasicECommerce]

the new ad is blowing away my regular ad both in terms of CTR and conversions. Testing really does pay.

What sort of conversions are you seeing? Any changes here? (change in order size, item type, etc.) I'd be interested to understand how the consumer absorbs this information for their search - displayed by their purchasing decisions.

[eney]

I think maybe the loophole has been closed?? I can't find the results on Google or Yahoo. Well, hope it generated some traffic while it lasted. Nice one.

[jimbeetle]

Last I saw it was promoted to the top spot. Must be great CTR.

[abbottsys]

Last I saw it was promoted to the top spot. Must be great CTR.
Yup, the ad is going strong. I've been fiddling with it to try and fix some technical issues, so the CRT just dropped to 10% from 16% earlier today. However, it seems to be climbing back up again now.

[eney]

Yes, now I see it. Strange that I couldn't see it in either SERP previously. Maybe something with the data center associated with my IP? Do these change based on traffic volumes, or do I always get data from the same physical location based on geography?

[attak]

I hate to be the one that gives you a dose of reality, but honestly, this is boarderline amateurish. In fact, if it weren't for the creativity behind it, there is no other way to describe it.

The campaign's CTR is that high only because you are capitalizing on Google's reputation/credibility - not your client's. I would say that if you had an ad (including copy/url/landing page) that better targeted the user, I bet a 16% CTR could be considered low.
The target "customer" is desperate to be a parent. They will try anything to find a solution and will spend quite a bit more time online to find a company that they feel will be able to help them out. How hard could it really be to attract them to your ad?

Overall, this tactic is more suited to an affiliate marketing program...not an actual business.

I understand that this is only a "test", but seriously...

Someone who actually needs invitro shouldn't have to deal with a SPAM result.

-Attak

[miillss]

Maybe your click through rate shot up because everyone was reading this forum and trying it out?!

Let us know your conversion stats when you have them.

[abbottsys]

...Someone who actually needs invitro shouldn't have to deal with a SPAM result.-Attak
I hear your concerns, but don't agree. There is no spam involved. The landing page is very relevant, as is shown by the fact that the conversion rate is about 3X that of my regular ad!

If I was just getting great CTRs without conversions then I would agree with you. But the fact that I'm getting superb conversions shows that users are finding the ad/landing page highly relevant.

[abbottsys]

Maybe your click through rate shot up because everyone was reading this forum and trying it out?!...
I'm running the ad on 2,000 keywords, and only published one here ;)

[abbottsys]

SUMMARY:

Use Google.com as the display URL in your ad, and use a landing page with this type of construct:

google.com/search?q=keyword&sitesearch=yourwebsite.com

sitesearch= <your website>
q= <a keyword used to sitesearch your website>

So, your ad landing page is a Google sitesearch of your website.

Technical Notes:
1) Google serves Ads on sitesearch SERPs, and you clearly want to eliminate those. So it's important you choose a q= parameter to display the pages you want on the SERP but without ads. Adjusting this takes some trial and error. You might wind up using a q value that has absolutely nothing to do with your ad keywords.
2) This ad does not violate any Google AdWords ad policies. (At least, not that I know of.)

[abbottsys]

Test over.
After about a week of running the ad it was finally disabled by Google for trademark reasons. I'm surprised they did not catch this at ad creation time as they do for trademarks used in ad text. So, use of a Google landing page (and thus use of Google in the display URL) is a trademark violation. Makes sense.

[jimbeetle]

Kind of surprising that it took them a week to catch it, especially since you aren't shy about your tests.

Did you run just search or also run content?

Think beyond the domain name there might be a couple of other takeaways in there someplace. Fun test.

[abbottsys]

..Did you run just search or also run content? Think beyond the domain name there might be a couple of other takeaways in there someplace. Fun test.
Ran on both search and content. It was fun while it lasted.
Yes, there may be other takeaways from this test, let me ponder ;)

[abbottsys]

Wow!!! I just discovered a variation on this ad test technique that seems to satisfy all Google trademark concerns. It's now in production. I'll report results soon.