Last week I had the privilege of visiting the nice folks at Google's New York City office for lunch. Lair of creativity and Web savvy that it is, I wanted to know which of Google's tools fascinate their own ranks most. One of the fascinating answers I got: Google Insights, a little-publicized tool Google launched last summer to analyze search terms over time and location. It lets you plot search terms against each other, figuring out who is searching what, when, and where.
It sounds like obtuse data-slogging, but it's an intuitive concept: Insights tells you which searches are most popular where, allowing you to cull the kind of accurate self-report data that psychologists only dream of. Of course, it's all anonymous; you can't tell who's doing the searching, except that they live in certain cities, states or nations. But as far as societal bellwethers, you can't beat what Insights has to tell you--and if you're at all interested in marketing or advertising, the tool is a must-see.
Here's how it works. You pick keywords you're curious about, and enter them into Insights. Since we've heard a lot about foreclosure, insolvency and other fiscal disaster since last fall, I used "bankruptcy" as my first keyword. Predictably, searches for the term in the U.S. began rising slowly in the summer of 2008, when knowledgeable canaries like Peter Schiff were telling us that the economy was about to have a coronary. If you search that term against predictable stuff--keywords relating to real estate, law, stocks, and so on--you get a correlation. But what Insights is useful for is finding the correlations that you wouldn't necessarily expect.
So what happens when you map searches for "bankruptcy" against searches for, say, "guns"? I'll tell you what happens: you get worried. Gun searches spike hugely in November of 2008, and correlate with bankruptcy searches in areas like Nevada and Arizona.
Of course, that's the chicanery of Google Insights: Its results are correlative, not necessarily causative. November was also election month, so there's a good chance that the spike is attributable to people researching the presidential candidates' firearms policies. Indeed, if you use Insights' regional interest menu to see where "guns" searches were highest, the winning states (Wyoming, Alaska, Utah, Idaho) also have some of the lowest incidences of "bankruptcy" searches.
If you buy online advertising for your business or you're curious about a product's endemic popularity (or lack thereof), Google Insights should be your next stop. It's free, easy to use, and presents results in a sensible, graph-aided format (and downloaded in CSV format).
Before you hang your ad budget on something Insights tells you, be sure to read Google's support page on the service, so you have a good handle on what your results actually mean. Statistics, as any pollster can tell you, are often saying the opposite of what they seem to be.