Welcome to WebmasterWorld Guest from 18.206.48.142

Forum Moderators: buckworks & eWhisper & skibum

Message Too Old, No Replies

Changing the landing page affects quality score and price?

if that is the case, then G is doing itself a disservice

     
1:07 am on Jul 20, 2006 (gmt 0)

Senior Member

WebmasterWorld Senior Member 10+ Year Member

joined:Apr 6, 2005
posts:830
votes: 0


Joe Adwords likely doesn't have each and every keyword going to the most specific page for that keyword on his site.

He probably has all of them going to the homepage, or to one of 3 or 4 pages.

The people who DO probably have them going to specific pages more likely fit into a couple of molds:

Big Company paying someone to do it so perfectly.
Seo Guys that are that precise (many of you I guess)
People running ads sending them to "keyword related" generated pages for adsense.

So who loses out here?

The little guy, I guess?

Or is my assumption wrong?

Also, if quality score is already determined, can it be changed?

Is it the content of the ad? The landing page as a whole? The meta of the page? The page title? The landing url?

Can ANYONE explain any of this to me?

1:27 am on July 20, 2006 (gmt 0)

Senior Member

joined:Nov 29, 2003
posts:875
votes: 0


Can ANYONE explain any of this to me?

Yes, but not in a nutshell and I'm about done with spending time on this.

First, google has said it will run regular updates, as they do with other similar algos. So, yes, it will "change", as will the actual algo, as they tweak it.

As for the determining factors, they've said it's landing page, and ads (obviously tied to key words). But they have never said it's ONLY those things. It's clear there are others.

What are they? some factors connected with domains. Some things connected to CTR and CPC costs. Other stuff.

You (and since this might be my last note on the subject) have to understand that they don't need to use obvious variables -- they can use variables that are CORRELATED with certain kinds of sites they don't want. So, if they identify that domain name registration date is correlated with certain types of sites, they can use that, even though there may not be any logical link between the two -- only a statistical link.

To be completely silly, if they found that living in Salt Lake City in correlated with advertising certain sites they don't want, they could use that, without making any assumptions about the people there (it's a statistical prediction issue. It's really hard to understand this without at least some background in either the logic of, or the math of predicting with statistics. Because it's not completely intuitive. And, the algos are somewhat dynamic regarding the actual values and weights assigned to the variables in the formulas.

On the later point, that's partly why the results seem to be random, and inconsistent between sites and variable combinations. The other reason they appear so is that we don't know what variables to look at in what combinations.

So, what's the upshot? If they say to change the landing page, it's worth a try, but keep in mind that things may not change until the recalculations of the scores which could be tomorrow, next week next month or months from now.

Since the algo no doubt uses a number of interacting factors and variables it's unlikely that you could figure out a way to game the system that will be effective over time. If you do, it'd be a lucky guess.

If I was hit, one thing I would try, believe it or not, is put the prices of the products I am selling IN the ads.

I'll let folks figure out why.

Finally, I briefly explained how you can build such algos based on discriminant analysis, although I can't say they did it that way (they should have). I'll leave y'all with a thought.

You can use the exact same methods to understand the algoritim in general terms (which would be pretty valuable), in effect reverse conceptually engineering it. Thing is it would require a lot of data, a fair amount of number crunching (which isn't a huge problem), and a lot of time, so by the time you got to the end, it would have changed too much to be of value.

1:35 am on July 20, 2006 (gmt 0)

Senior Member

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month

joined:Feb 28, 2002
posts:1328
votes: 0


If I was hit, one thing I would try, believe it or not, is put the prices of the products I am selling IN the ads.

I was doing that before I got hit, was amusing to watch an item that retails for $9.99 go to $10 a click.

2:44 am on July 20, 2006 (gmt 0)

Senior Member

WebmasterWorld Senior Member 10+ Year Member

joined:Apr 6, 2005
posts:830
votes: 0


thanks for the reply.
5:38 am on July 20, 2006 (gmt 0)

Junior Member

10+ Year Member Top Contributors Of The Month

joined:May 20, 2005
posts:175
votes: 2


If I was hit, one thing I would try, believe it or not, is put the prices of the products I am selling IN the ads.

Yea, no luck here, except putting the prices in the ads kept the CTR lower and CPC higher on some ads, which of course were left untouched. The ads that I had a high CTR and low CPC were disabled, and some had price in the ads, some didn't.

 

Join The Conversation

Moderators and Top Contributors

Hot Threads This Week

Featured Threads

Free SEO Tools

Hire Expert Members