Google Offers Insight for Search Marketers

My most recent column took a look at Google’s new keyword research tool, Insights for Search. Following that, an astute reader wrote to ask for clarification on some of the lesser-described components of the Insights for Search results page.

Her basic question was this: what is the exact relationship between the specific query term, its value, and the values of the top searches related to the specified term? She’s smart to ask, because it’s not necessarily what you might infer at first glance.

Without giving away the reader’s industry, suffice it to say that she came across several top searches terms that had a value of 100. Yet laid against her original query term on a demand graph, the terms had either significantly greater or lesser demand than the original query term. If that’s the case, what does the 100 value mean?

The Magic Number 100

Much of the understanding of Google Insights for Search revolves around its ranking mechanism for related search terms.

As you can see in the example of “boats,” the number 100 is associated with the very peak of search demand over the specified period. In the case of “boats,” it comes in mid-2004. Therefore, when you search for a single query term, it’s automatically assigned a relative rank of 100.

But don’t look at this 100 as the same scale as that in the related top searches search terms.

Google discusses these related terms in its Help documentation. There, Google says that its system:

    determines relativity by examining searches that have been conducted by a large group of users preceding the search term you’ve entered, as well as after. Keep in mind that top searches are also automatically categorized, so some errors in classification may exist.

Consider this top searches number as a value that measures how related the term is to the original query term, not necessarily how directly similar the actual query demand is. For example, the top related search term for “boats” is “boats for sale” (with a value of 100), yet when you lay the terms side by side and measure their relative search demand, you can see that demand for “boats” far outstrips demand for “boats for sale.” Yet what I think Google says by its 100 value for “boats for sale” is that it’s completely related to the term “boats” because the same people search for both terms when searching for the same bit of information. They might not search for them in the exact same quantities, but the same people typically search for them while attempting to satisfy a query.

The better way to measure difference in overall demand is the bar graph just below the “Search” button in the most recent link (“boats” vs. “boats for sale”). Here, you can see that Google places a relative value of 63 for “boats” compared to 8 for “boats for sale.” This roughly 8:1 ratio appears to be a close approximation of the contrasting demand shown in the line graph. So if you’re looking for an accurate representation of search demand, look at this image. If you’re looking for a set of terms that users routinely use alongside your core term (which may or may not be of a similar demand), look at the related search terms.

For example, “dogs” and “cats” may have similar search demand, yet they are unrelated. “Dogs” and “puppies” may be heavily related, but that doesn’t mean they have similar search demand. Use these two measures with these differences in mind, and you should be on the right track in making decisions on your keyword use.


I still have tremendous problems using this tool over long periods, drilling into various categories, without Google thinking I’m an automated program. I understand the wealth of information here and, consequently, I understand Google wanting to guard against people mining it and causing undue server load. But I’ve had dozens of cases of the site telling me that I may have a virus or malware installed and that I need to fill out a CAPTCHA form to complete my query. When I do, I’m told that I look like an automated program. This is after about five minutes of drilling through various categories manually. I’m hopeful this will improve over time.

Join us for SES Search Engine Marketing Training Day, September 26 at the Fairmont Dallas.

Related reading