To explaining keyword selection to my clients, I find it helpful to use a real-life example. My high tech step-father is particularly hard to shop for. Last year I decided to get him a new and exciting gadget in lieu of the overpriced tropical shirt that I usually get him -- but finding this unnamed and unheard-of gadget was complicated. I began my Internet search with the term "unique gift," believing that I would find an interesting selection of unordinary gifts. Instead, I found ads for cologne and shaving products, which are not particularly unique gifts, in my opinion. I tried another search for "unique gadget," which offered slightly more intriguing results, such as digital music devices, but still didn't fit my description of a unique gift. It wasn't until I had attempted a number of searches that the term "unique high tech gadget" brought up an ad for a new invention-type item that suited my needs.
Most engines rank keywords to determine position and cost in an attempt to help searchers easily find what they are looking for. A large part of the actual rankings are determined by click-through rate (clicks/impressions). With this model, it is possible for a word with a lower click-through rate to be in a top position, but the position would cost that company significantly more.
The term "unique gift" was not a good keyword for aftershave or electric shaving products, which filled the top ten positions. It is never beneficial to pay top dollar to keep an ad in a high position for a keyword that isn't highly relevant. Keyword selection isn't about being broad or specific, it's about relevance. It is possible that someone searching for a unique gift may have clicked a shaver ad to see what made those shavers so unique, but someone searching for a gift that is truly unique is not searching for shaving products. I'm sure that those companies saw some conversions from the keyword, but bidding for top position was probably not as effective as applying additional spend to more targeted keywords such as shaving products or electric razors, where they could benefit most from first position.
Of course, I wouldn't suggest that the shaving companies drop the word completely either. The beauty of search engine marketing is that a company pays only for clicks, not for impressions. The secret to making these relevant, yet less targeted, keywords work is to offer a lower bid amount, and make sure the ad copy describes how the keyword can be associated with the product.
"Unique Acme Technology Shaves Close & Makes a Great Gift" would be compelling ad copy for the term "unique gift." Remember, searchers are looking for a keyword, but choosing to click on an ad. As long as the ad captures the need of the searcher with that keyword, there is a good possibility of getting a conversion from a click.
As I mentioned above, I finally found the unique gift I was searching for by using the term "unique high tech gadget." Though this term is both very inexpensive and targeted, I wonder how much traffic finds its way to that site with such an obscure phrase. Since the company offered unique gift products, the term "unique gift" would probably have been valuable for them as well.
This is a great example of how using a broader keyword can be beneficial to a company. I imagine that its click-through-rate would have been good too, which would have lowered their cost to stay in first position for that keyword, by being more relevant to searchers.
Thoughtful keyword selection is the tool that puts your product in front of the most targeted searchers. Writing enticing ad copy that sticks out on a page can bring a great ROI, but not if it finds its way in front of an audience that isn't ready to buy, or worse yet, no audience at all. Inch per inch, the place where an ad appears on a search engine's results page is some of the most expensive real estate in the world. As with any large investment, research, investigation and testing will ensure that your advertising dollar is well spent.