May
31
2011
On May 25, Google released Google Correlate – and statistic nerds everywhere spent the weekend playing with it.
Google Correlate is the math-nerd cousin of Google Trends. With Google Trends, you can input and compare any given websites or keywords and phrases. You can see spikes in search and web traffic, see where the most traffic is coming from geographically, and other details. But you’re not really working with any raw data and you certainly don’t need to know what r is.
Google Correlate, however, is a more statistical analysis of keywords and trends – telling you the correlation between your keyword and other keywords based on either time or geography. You still don’t have to be a stats junkie, but it might help.
Some examples:
When you input “jewelry” into Google Correlate, you wouldn’t really be surprised at the results using the “Compare time series” function (which find correlations based on time). At the same time people are searching online for jewelry, they’re searching for things like diamond necklaces, gold bracelets, and pendants. And, unsurprisingly, the biggest spike in traffic each year coincides with Christmas, and then a smaller spike for Valentine’s Day. Searching for “sandals” gives correlated terms like boat seats, lake fishing, and linen clothing – all summer items, searched most during the summer.
For an e-commerce site, this is a great tool; it can give you some additional ideas as to what to sell in certain seasons or what other items you may consider pairing with your current products.
When you switch to the “Compare US states” function, Google Correlate tries to find connections geographically. So the same states that are searching for “jewelry” are also searching for tunic tops, shoulder shirts, and makeup organizers; these are obviously fashion-focused people. The top-correlated term for “sandals” when searching geographically is actually “rash in groin,” with a big concentration in the Southern States. (So maybe you should sell some calamine lotion with your sandals!) Again, this data can help you understand what type of people are searching for specific terms – not just how many of them are searching.
Not every term is enlightening. If you input “Olympics,” there are huge spikes in traffic during the actual Olympic events (and not much in between), and the correlated search terms aren’t very diverse – all of them include something about the Olympics. However, if using the same term to switch to the “Compare US states” function, Google Correlate tries to find connections geographically. So the same states that are searching for “Olympics” are also searching for terms like “cats cradle string” and “banana oatmeal.”
What does it mean? Probably nothing; there likely isn’t a real connection between Olympics and oatmeal. But should that stop you from packing your bags and selling cat’s cradle string in North Dakota (the state with the highest correlation) during the 2012 Olympics?
Actually, that probably wouldn’t be that good of an idea. But for those of us who love to crunch numbers, these kinds of tools will make you giddy. And for those of you with a fantastic business sense, you may have what it takes to start a banana oatmeal business in North Dakota.
Who am I to judge?
Scott Spjut is a writer and editor who has been featured in various magazines, newspapers and websites - including Newsweek, the Washington Post, CBS News and the Las Vegas Review-Journal. Scott currently works with PMI Education helping people change their lives.