Co-Authored by Ekaagar Singh Hara
Google Correlate is a tool which is similar to Google Trends and Google Insights, but it takes the raw data and analyzes it. As the name suggests, it looks for correlations – that is, whether the demand for a given keyword matches the popularity of any other keywords.
Google Correlate is like Google Trends in reverse. With Google Trends, when we type in a query, we get back a data series of activity (over time or in each US state). With Google Correlate, when we enter a data series (the target), we get back a list of queries whose data series follows a similar pattern.
Initially this tool was a Google Labs project and the data set provided was based on US only, thus restricting users from other countries to use this tool. Lately, they have gone international and now, people from all across the world can use this tool to compare their data series.
After finding some useless correlations, I tried to find some parallels that really mean something. So I searched for ‘government grants’ on Google Correlate expecting to see a correlation of searches for other variations of that keyword.
The results that were shown for this keyword were very relevant to my search term.
The below graph shows the trend between the target keyword and the correlated keyword -
Keeping my term ‘government grants’, I now switch to "Compare US states" to see what it looks like geographically.
There are few changes with the correlated keywords when I click on 'Compare US States'. This is what the map looks like for this -
Another interesting feature is that, we can upload our own data (numerical data that shows a trend) in Google Correlate and Google will try to find search terms that had the same correlation/trend data as our own data. We can even draw our own line chart (if we don't have a data set) and Google will show us matches to the spikes and dips of the visual that we drew.
For example we drew a graph (x axis time and y axis searches) like the following -
Then Google will find keywords which had a similar time graph.
Hence, it can be inferred that -
We have found a few terms that might align nicely with our business goals (‘grant money’, ‘get a grant’). We could use these terms to expand our content offering, maybe with a page that will have variations of our keyword.
We observed how our main term trends over time, and we can compare it to correlated terms. Thus, we could use this to plan marketing campaigns (if any) around the time of highest search activity.
Using the GEO feature, we learnt about search activity within the US states, and how they correlate. We could use this to focus our search activity geographically (location based filtering).
- The only thing really lacking right now is the ability to enter two keywords and check the correlation between them.