Google does “the wisdom of crowds.”
If you think about the millions of searches conducted daily using Google, (GOOG) there is reason to believe you ought to be able to divine patterns or trends from the activity. If everyone is searching for a particular song for example, you might expect that artist to climb the charts. If searches are fast and furious for a digital camera, sales ought to correspondingly increase.
But can examining a broader swath of Google searches related to specific industries inform where those sectors, and the broader economy is headed? The propeller-heads at Google think it can.
Tucked away inside Google Finance is the newly-launched, “Google Domestic Trends”. While it sounds like it might offer the latest in recipes for meatloaf or techniques for pressing a shirt collar, what it does is track search traffic across 23 specific sectors of the economy ranging from auto buyers, to jobs and the retail trade.
All indices use January 1, 2004 as a baseline from which a seven-day moving average is calculated. So if you look at auto buyers, you will see a more or less predictable seasonal car-buying pattern, peaking during the late spring and summer when people are looking at the new models and dealers cut prices on last year’s stock. The number of searches for cars falls off around the winter holidays before ticking up again at the end of the year, presumably for tax reasons.
The economy, reflected in search activitiy
As might be expected, the volume of searches for cars was generally lower this year compared to the last five years as the recession scared consumers into sticking with their jalopies. That is until the “cash-for-clunkers” program kicked in, producing an enormous spike in search volume Aug. 7 that dwarfs any point on the auto buyer index.
If you plot actual monthly retail clothing sales from the U.S. Census Bureau compared to the volume of retail sales searches on Google over the past two years, you see both decreasing over time, especially during the peak holiday buying season.
The idea that Google searches might be able to put a finer point on economic trends came from research around unemployment Google’s Chief Economist Hal Varian (that’s right, Google has its own economist) did using Google Trends. Google Trends constantly updates the most popular search queries.
What Varian found is that by building a model to forecast initial unemployment claims, and then adding in Google Trends data around unemployment the forecast results were more accurate. The idea is that you could better forecast car sales, travel, construction – you name it – using other categories of search volume data.
Google is not building forecast models for people with Domestic Trends; it’s just offering the general volume data. How you plug it into your own forecast model is your business. Is Google Domestic Trends a crystal ball into sectors and the fortunes of companies within them? No, but it is more data, and that can never hurt in making decisions nor in getting a clearer view on the economy.
As Varian writes on his blog: “It has been said that if you put a million monkeys in front of a million computers, you would eventually produce an accurate economic forecast. Let’s see how well that theory works…”