How Keywords Are Researched

I use Wordtracker and/or KeywordDiscovery. These are the only 2 databases that keep track of actual search queries — up to a whole year's worth of searches. It really helps when you know what the search patterns are for your particular industry: how people phrase their search queries, whether they tend to use plural or singular terms, how they spell or misspell their searches, etc.

And once we know what search terms to target, we can test them out in a pay-per-click advertising (PPC) campaign.

(The following is a Q&A article from Jill Whalen's High Rankings Advisor newsletter, issue 186, Jan. 17, 2007. Republished with permission.)

Why Use Keyword Research Tools?

Jill,

I recently started subscribing to your newsletter, and since you offered to answer questions, I have one for ya!

Keywords. I have been designing web sites for 5 +/- yrs, so I am aware of what they are and how they are being used / can be used. I'm not saying I am an expert, just that I'm not a novice. Oh! And I read the LED Digest! ;-} .

I realize keywords can be used in Meta tags, although their importance there is now debatable. Keywords should also be used in the content. Got all that.

So here's the question: How does one use something like Wordtracker?

I am obviously missing something because I have heard some folks rave about these services.

If I have a site about dogs, a few keywords would be canine, dog, k-9, puppies, etc. and I would use those in my text when and where I can. I can understand using the service for finding which keywords are searched for the most, and I can even accept one might find words one never thought of before.

But other than that, what good does knowing "Britney Spears" (the poor pathetic thing) is searched on 8 million times in a particular week? What? I name a puppy picture after her?? ;-}

I think you may see my problem better than I can describe it, at least I hope so! Can you enlighten me?

Thanking you in advance,

Nancy C.

++Jill's Response++

Hi Nancy,

Great question!

From the sounds of it, you may not have seen all the features of the subscription keyword research programs such as Wordtracker and/or KeywordDiscovery.

You are 100% correct that knowing the top words searched is of no value at all. Some may think they want to optimize for the top words searched, but that would be silly and futile. You can bet that the top 100 words searched each week (heck, probably the top 10,000) are going to be way too competitive to realistically expect to achieve high search engine rankings for them.

You might be wondering what "competitive" means in terms of keywords. In basic terms, the competitiveness of any given keyword or phrase is based on how many pages online there are in the search engines' database which are optimized for the keywords in question.

Many people do a search at Google using their keyword phrase and erroneously judge their competitiveness based on the number of results that are returned. For example, if your phrase is "search engine spam" you might type that into Google (without quotation marks) and see that there are about 40,400,000 results for that phrase.

That's not much help in determining how competitive that phrase is, as it's simply telling you that there are over 40 million pages that Google knows about using the words "search," "engine," and "spam" somewhere on the page, in the indexable code, or in links pointing to that page.

Now, you could get a bit closer to determining the competitiveness of a phrase by putting it in quotes when you search for it in Google. Doing that with our phrase "search engine spam" returns about 150,000 results. See how that's a huge difference from 40 million? The search with quotes ends up showing you how many pages Google knows about that have that exact phrase somewhere in their indexable code or in links pointing to the page.

Still, just because a page is using a given phrase doesn't mean it's actually optimized for that phrase. Many of those 150,000 pages may not be optimized at all, and might be easily beatable in the results with just a bit of work.

To narrow down the field a bit more and have Google show you only pages that have been optimized (at least in some rudimentary fashion) you can type your phrase into Google using the "intitle" command. For our phrase, put this into the Google search box: intitle:"search engine spam" -- this command returns all the pages that have used the phrase "search engine spam" in their Title tag. As most of you know, keywords placed in Title tags are given a ton of weight by all search engines. If a page is using a phrase in its Title tag, it's safe to assume that it's been optimized, at least a little bit.

Google returns about 1,040 results for pages with that phrase in the Title tag. Big, huge difference from that original 40 million! At this point, it's clear to see that those 1,000+ pages are indeed your competition. Further study of them will be in order to determine just how optimized they actually are. Many of them are probably using the phrase *only* in the Title tag and might be easily beat in the rankings with just a little work on your part.

Now, compare the intitle results for "Britney Spears" and you'll find about 985,000 that have her name in their Title tag. So you'd have to have a page that was better and more relevant than the nearly million others that are about Britney Spears. Trying to be better and more relevant than 1,000 other pages is daunting enough, but 1 million? Just not worth the time and effort, in my opinion.

So now that I've illustrated why using keyword research tools to find the most-searched phrases is a useless exercise, let's get back to the original question.

What exactly are keyword research tools good for?

They are great for finding the keyword phrases that people are actually searching for when they would be looking for your products or services at the search engines.

So in your page about dogs, no, you don't want to use the keywords "canine," "dog," "k-9," "puppies," etc. Absolutely not! Optimizing for one-word, general keywords such as those is just as bad as optimizing for "Britney Spears." Our handy-dandy intitle search for "dogs" shows us about 5,650,000 results, and about 1,200,000 for "canine." Those are not optimizable keywords.

In SEO -- and more specifically when researching keywords -- the idea is to find the phrases people use in which decent search engine rankings are also attainable. Sure, you can guess at phrases like we used to have to do in the old days, but why guess when you don't have to?

What you need to do is go to the Research Keywords section of Wordtracker or KeywordDiscovery and start typing in your brainstormed phrases. The tools will let you know which are used by real people at real engines and which are not. They will also provide you with other ideas for phrases you may not have thought of. You can even put in a single word such as "dog" and get back every phrase that has "dog" contained within it. From there, you'd want to narrow it down to the phrases that most relate to what your site offers and which are also attainable in the rankings.

Be sure to research phrases for each page of your site. Certainly your site is not just about dogs in general, but about specific dog things. Maybe you offer all kinds of dog chew-toys. If so, you'll want to research all the phrases that have to do with this, including things like "squeaky rubber bone," etc.

Using our previous example, some quick research at KeywordDiscovery shows that the phrase "search engine spam" gets searched upon less than once per day, so it's not an ideal phrase to optimize for, in my opinion. Still, it does show up on their radar, and I imagine it might bring some traffic now and then if I had a page optimized for it. I see my buddy Alan Perkin's "Classification of Search Engine Spam" paper is in the top 10 for that keyphrase. I'll have to ask him if he gets much traffic on the phrase. And who knows, maybe this article will eventually show up in the results for it too once I upload it to the archives.

These types of phrases are known as the "long-tail" and really don't need to be specifically optimized for, as simply writing articles that utilize them is often enough to show up in the results. But that's an article for another day!

Hope this helps.

Jill