What I Search, What I Say, and What I Do

Is anybody connecting these dots? Because I think that they are connected… we just need to be able to visualize those connections somehow for the purposes of predictive analysis. People search for things on Google, say things on Twitter, and do things everywhere. The first two are public, the last is proprietary to a specific organization.

If something is proprietary, then that creates competitive advantage and dare I say possibly, ROI. Yes, I’m implying that the future of money in social media lies somewhere in the data everyone is producing at breakneck speeds.

Keywords

Before we can explore the possibilities around the concept of connecting what people search for, with what they talk about, with what they actually do, we need to understand what we care about. If I sell a product, then I definitely care about keywords around that product. So I will need to create a list of those keywords before I start looking at Google and Twitter. Then I will need to use the same list in both platforms.

So build that list, and then we can begin.

google The Google Trends API is open to anybody, and although I don’t know the extent of what that means because I haven’t fully researched it yet… I think that it means anybody can pull specific search data into an application de jour. If you’re unfamiliar with Google Trends, a great example is what Google has done with the Flu. Typically, people use specific search terms when they’re interested in something. So connecting flu outbreaks to searches for flu-like keywords can be an early indicator of an outbreak in a specific region. Knowing when people are searching and where there searching from is something that we’ll be interested in for this experiment.

twitter The Twitter API is definitely open, and after working with @mathiask on myTPSreport.com, I am aware of the limitations of the word open in Twitter-speak. With a developer account, 20,000 calls can be made to Twitter every hour. Which is more than enough for something like this… Taking my master list of keywords and importing them into a Twitter search will reveal all of the conversations going on in real time. I will want to capture the data around those tweets, specifically time and geography information.

proprietary This is where things get a little bit awesome. We’ve been using open tools that anybody can use in Google and Twitter, but Google and Twitter can only tell us what people are searching for and saying… They can’t tell us what people are actually doing. Only we can tell what people are actually doing with regards to the products we sell. We know when they buy and where they buy.

So now we know what people are searching for, what they’re saying, and we’ve always known our own stuff… so what?

Social Media Predictive Analytics

If you’re familiar with the concept of Trend Trading at all, and let’s face it every blue blooded wannabe get rich quick capitalist has a book on Trend Trading in their personal library, then you understand the idea of finding indicators in data of future events. Fortune telling if you will.

When we have volume, time, and geography information around search, conversation, and sales… then overlaying the data could produce some interesting results. Systems theory tells us that systems are complex networks of relationships that can be analyzed. From a social media stand point, it could be very exciting to understand the connections between what people search for, with what they talk about, with what they actually do. Something equally interesting would be learning what could and could not be done to affect the system around a given product.

Smelling what I’m stepping in?

Let’s Talk

I’m very interested to find anybody else out there thinking about connecting what people search for, with what they talk about, with what they actually do… Is this the lamest idea you’ve ever read? Do you want the minute and a half back that it took you to read down to this point? Or are you interested in this stuff too?

Let me know.