Curious what your thoughts are about when to test? during the initial launch? after Google has optimized based on ctr? Is your strategy affected by potential search volume?

Great content on your site! Art

Ah Yes…the noggin..I use it myself.

Ad machine? The only machine I rely on to calculate statistical significance is my noggin’.

Normally, if you have more than 150-200 clicks per ad running concurrently and all else is equal, a statistically significant winner should be relatively obvious.

I should have said when to start testing adverts. Yes..I usually start with exact and phrase match with 3+ word phrases to start the account off on the right foot. Let Google optimize based on ctr and then start to integrate ad rotation. Landing Page testing can start in the beginning.

Does the Ad machine calculate the advert statistics…when the data is statistically significant?

Thanks for your kind words.

To your question, I normally go on the hunt for cheaper and more precise keywords first, winning a relatively high CTR on long tail exact and phrase match. More generic terms are tougher to win, obviously, so the long tail can give you a lot of information about search user intent and search query popularity.

But when to test, that’s based on a lot of factors. Always test, if you can. Testing is not just a matter of ad text, but also landing page relevance. So creative labor can be the limiting factor there.

Curious what your thoughts are about when to test? during the initial launch? after Google has optimized based on ctr? Is your strategy affected by potential search volume?

Great content on your site! Art

Thanks for the comments and for stopping by. How could I forget subdomains? Another easy way to differentiate your ads and increase ad quality.

Some have also said that alternate tokens work, such as “complimentary” versus “free” shipping.

To add a couple other:

-Split testing subdomains

red.shoes.com
blue.shoes.com

-Testing forms of verbs

“Washes” your windows
vs.
Window “washer”