Dynamic Keyword Insertion: Friend or Foe?

Last week, I described best practices for building large keyword lists, and the “research-shop-buy” process that should guide the way keyword lists are divided and combined with great, targeted ads.

After last week’s column, a reader asked “Why separate the 540 keywords into separate ad groups? Why not put all 540 into one ad group and use Dynamic Keyword Insertion (DKI) to make sure each keyword appears in the ad?”

Excellent question. DKI is a powerful feature, but it is frequently misused, to the detriment of many PPC campaigns. In a nutshell, DKI lets an advertiser designate a spot in an ad where a keyword will be inserted when a search term matches that keyword in the ad group. So, for example, if an ad group contains the keywords:

chicken, pulled pork, beef tips, marinated eggplant

And the ad text is:

Best BBQ {KeyWord:Sandwiches}
Mouth-watering {KeyWord:Barbecue} Smothered
in Smoky Sauce. Best BBQ in Town!
www.OwensboroBBQ.com

Google inserts the keyword from the ad group in the spots where it sees the {KeyWord} token. So the resulting ad would show:

Best BBQ Beef Tips
Mouth-watering Beef Tips Smothered
in Smoky Sauce. Best BBQ in Town!
www.OwensboroBBQ.com

I’ll go into best practices in the upcoming DKI column; but for now, understand that indiscriminate use of DKI results in ad copy that is worded for the “common denominator” of the keyword set, with diluted messages that fail to exactly match the searcher’s intent/need.

Also, if the keyword set includes many keywords that are too long to fit in place of the default text, the less targeted default ad is frequently shown.

For example, if the search term includes the word “Pulled Pork,” the ad will appear like this:

Best BBQ Pulled Pork
Mouth-watering Barbecue Smothered
in Smoky Sauce. Best BBQ in Town!
www.OwensboroBBQ.com

In this case, Google inserts the keyword in the headline, but inserting the keyword in the next line would make the line too long, so instead it uses the default word “Barbecue.”

And if the search term includes the word “marinated eggplant,” the ad will appear like this:

Best BBQ Sandwiches
Mouth-watering Barbecue Smothered
in Smoky Sauce. Best BBQ in Town!
www.OwensboroBBQ.com

In this case, inserting the keyword in either line would make the line too long, so Google uses both default values. This is hardly ideal, since someone searching for the specific term “marinated eggplant” is unlikely to click on an ad describing generic BBQ sandwiches. This is one of many examples where results obtained can be worse than those obtainable without it.

Another reason DKI should be used carefully: it’s frequently worthwhile to include misspelled words in a keyword list — incorrect pluralizations, for example. When these misspellings are made to appear in the ad text, it can negatively affect CTR – even if the correct spelling is relevant.

(I plan to cover DKI in depth in an upcoming column. In the mean time, you’ll find some good resources by one of the top authorities on the subject, Matt Van Wagner, here and here.)

So if DKI shouldn’t be used in most cases, what’s the best way to deal with long keyword lists like the 540-word one I described last week? It’s very straightforward, when you keep in mind the fact that every searcher is likely to be at a particular point in the research-shop-buy process. Here are some examples from our 540-keyword list — as you can see, one can often infer where the searcher is in the process by the words they use for their search:

Research: 12-string guitars, acoustic guitar, twelve string guitars

Shop: high-end acoustic guitars, compare collectable 12-string guitar, acoustic guitars online

Buy: buy acoustic guitar online, buy high-end 12-string guitar

Next week I’ll show you exactly how to split up the 540 keywords into ad groups, and the associated ads that take into consideration the searcher’s position in the research-shop-buy process. As always, let me know your comments and questions via the feedback form below.