Is The 30DC AdWords Strategy Wrong?

[junkywunky]

Hello,

I've read the 30DC Adwords strategy and I've also started reading around on other strategies so if you could help me out I'd be very grateful:

30DC suggests having one keyword per ad group with exact, broad and phrase matches.

However, I've read different theories which seem to go like this:

Method 1

* Choose a subtopic or theme for the ad group
* Create 20-30 keywords for this ad group
* Do this for several subtopics of your main theme
* Continue testing and improving

Method 2

* Put into several hundred keywords of a subtopic into an ad group
* Test for a few days
* From that ad group select the top performers and place them into a new ad group.
* Continue testing and improving

So why such a difference in theories? Which one is correct? Is there a good way to combine all 3 theories?

Thanks for the feedback in advance.

[alfaprima]

Personally, I use method 1.
Method 2 sends the wrong message to google, and it will start charging more for the keywords.
The key is to be laser focused / targeted.

[Isha]

Ed tweeted an interesting item related to this on 22nd December:

Pay Per Click Update from Perry Marshall - VIP | The Underachiever Life

[lisahartwell]

The 30DC adwords strategy is very specifically as an extra testing method for researching a market, which is why you only worry about your main keyword phrase.

However, it is still a good method. The more keywords in each adgroup, the less targeted they are likely to be (you want your keywords to be highly targeted to the page you are directing them to). So, it makes sense to have many adgroups within each campaign, and each tightly focussed.

Certainly the days of piling hundreds of keywords into each group is long gone - Google will either charge you a fortune or turn your ads off altogether.

Lisa

[alfaprima]

Thanks for the link to the article.
I'm more confused now.. Although it seems a possibility that google is doing whatever they can to keep their botom line growing, my personal guess is that at the end, it may not be what it look like..
anyway, I'm testing separate groups, with broad and phrase to see it this shows some improvement..