Let Google Pick the best Landing Page

Anyone tried this technique & what were the results?

[sailorjwd]

I've been struggling with the landing page quality issue since July. So far nothing I've tried has made any positive difference in min bids, in fact, min bids continue to rise.

I've gotten rid of many links on the landing page..
Added 1000's of words of content...
Even added some dynamic aspects to the page based on incoming query string...

Perhaps I haven't waited long enough for the results to show up.

Anyway, my latest exercise is to let Google pick the best landing page. I have 3-4 candidate pages for several adgroups. So, I created identical ad copy and only changed the landing page url for four ads in one adgroup.

After running for several days it seems the optimizer has selected what I assume is the page with the best landing page quality. Only problem is it is the last page i would have selected as an appropriate landing page.

Has anyone else tried this test?

Do you think it is a true measure of which page has the best quality?

Am i missing something here?

[chewy]

Interesting approach. Very clever.

Have you measured other metrics in light of this approach?

EG how much time the user spends on the site relative to which landing page they come in on or pg views / visitor per landing page?

I usually figure I need 100 clicks or so to get "siginficant" figures on which to make decisions.

How many clicks are you basing this on?

[poster_boy]

I tried running this test for a very important keyword for us... after a few weeks of introducing a new landing page (that increased conversions by 20%+) our keyword was disabled with a minimum CPC that was 17x the average CPC we'd been paying for 5+ years (and achieving position 1.0 - 1.9 consistently at that CPC).

And, despite eventually deleting the new test cell outright (among other attempts), I can't get the old CPC back.

[rbacal]

Do you think it is a true measure of which page has the best quality?

That's a really cool, creative idea.

The answer, however, is no, it's probably not going to work. The reason why simply changing the landing page doesn't necessarily work is that you are being assessed using a number of criteria, a number of which have NOTHING to do with the landing page, but things that have to do with "broader criteria".

That's the reason why you haven't succeeded by changing the landing page, and why there are so many reports of changes of this type not working.

[sailorjwd]

rbacal,

The only thing different is the ad group landing url. All else is the same. So I'mm not sure what you could mean by other factors or criteria.

Oddly, and perhaps a fluke - 60% of my keywords in one ad group just went 'active'. it happens to be the most active ad group of the bunch on which I'm trying this test.. about 1500 clicks a day.

[rbacal]

The only thing different is the ad group landing url. All else is the same.

Identical? Were they all created at the same time/date? Are any/all/some ranked in search engines?

What other ways do they differ, if any?

And, of course you don't know whether google as botted them all yet, or just one, or what?

Also, you can't gauge results/success except over a period of days or weeks.

[aleksl]

what rbacal is trying to pull off is to disguise a trivial price increase by Google behind some sort of magic “complex” algo. Nonsense. Sailorjwd, there is no spoon.

What I’ve been seeing is that keywords that don’t exist on the landing page get priced too high. That is a flawed logic, but it is as good as one gets. Also, they look at your domain, if it has penalties your site will be “low quality”. And at Smart Pricing.

What else they’ve done, is they used Google ANALytics data to figure out most converting KWs using statistical analysis or whatelse, and they increased prices for those. To prove that, take a brand-spanking-new domain, place a copy of your landing page, and point ads at it (to make experiment clean you may want to use different AdWords account). If ads prices stay the same for the same KW, that would STRONGLY support the theory of statistical price increases.

JMHO