Google Keyword Tool Offers Insight for Search Marketers

The release of the free Google Analytics caused many people to think that paid analytics packages wouldn’t last long. Similarly, depending on who you asked, Google Trends spelled certain doom for businesses like comScore and Nielsen//NetRatings. I don’t think Google Insights for Search spells the end of the third-party keyword analytics tools. But in some cases, it highlights some paid tools’ shortcomings.

Google introduced Insights for Search two weeks ago on one of its blogs, which explains why it hasn’t been a particularly large story among organic SEO (define) practitioners.

This is an excellent tool that shouldn’t be confined to your PPC (define) toolbox. Google Insights offers plenty of relevant data for paid and natural search campaigns.

Touring Google Insights for Search

If you’re used to other Google tools, Insights for Search is easy to navigate. Start at the main page and enter a keyword, such as “swimming.” The resulting page shows the trend of demand over the last several years, news results (if applicable), most popular world regions (which, I assume, are per capita-based), and two very cool features, “Top searches” and “Rising searches,” which offer great insight into terms that may be logical targets in addition to the term you’re searching for.

Using Category Filters

Be warned: obtaining accurate keyword data depends on your ability to frame your phrases within the correct category. A general, non-categorized search on “cars” shows one search demand pattern. But “cars” is a term used in several different industries and social categories. Sure, it’s likely that most searches belong in the automotive category, but looking at “cars” within the music category skews the data due to the popularity of phrases like “Chasing Cars,” a recent hit song from The Killers.

This brings up two important points:

  1. If you don’t filter by a category, your results show the sum total of demand for that term across all categories, not just the assumed category.

  2. The graph results pertain not only to the exact term you input but to terms related to that term as well.

Here’s an example. We compare the popularity within the music category of “cars” and “cars -chasing.” The first term is all queries related to “cars,” while the second filters out any queries that also contain “chasing.” The result is interesting, because you can see that until Q2 2006, the lines overlapped, which means there was insignificant presence of “chasing” before that date in any “cars”-related searches. When “Chasing Cars” became popular, that term spiked and began to account for the majority of the aggregated popularity of “cars.”

This proves that when you analyze “cars” in Google Insights for Search, you’re analyzing not only “cars” but also “chasing cars” and any phrases based on the contributions of artists such as Gary Numan and Rick Ocasek. Otherwise, the minus sign operator would have no consequence.

Volume vs. Relative Popularity

Spend a few minutes in Google Insights, and you’ll quickly notice that keyword-demand graphs list data within a 1-100 scale, not with raw search numbers. Is this a problem? Not in my opinion. I’ve always believed that even the keyword tools that claim to have valid raw search predictions arrive at their data through wildly error-prone extrapolation methods due to their relatively small sample sizes.

The true value of a keyword research tool is its ability to accurately show the relative popularity of two or more phrases. If you know, for example, that “Chevy Trucks” is about two to three times as popular as “Chevrolet Trucks,” does it really matter whether there’s a specific raw prediction beside each term?

It’s easier to extrapolate traffic with your own rankings than with someone else’s traffic predictions. If you have top rankings for “roller blades” and you want to pursue “roller skates” as well, what’s the projected traffic and sales benefit? You’ll probably come up with a fairly accurate number by overlaying a relative popularity chart with your own traffic data for one of the terms.

Conclusion

It’s interesting that throughout my random use of Insights to research this column, I frequently encountered Google pages that told me I looked very similar to an automatic program, and it suspended my searches for a few minutes. I’m a quick typist, so I took it as a compliment. Still, I don’t recommend using the program to mine vast data fields — at least within a short period. Clearly, Google took precautions against that.

I’ll delve further into smart usage of tools like Google Insights for Search in future columns.

Related reading

site search hp
amp
Space Shuttle Launch
google-changes
<