Seeing the Future: Sentiment Analysis and Predicting Trends
As marketers we spend considerable time looking at analytics.
Much of the analytics has to do with prior performance. We look at what has happened, determined why we think it happened that way then we adjust accordingly.
In a real-time sense we look at analytics and we hope to learn what is happening “on-the-fly” and adjust accordingly.
So what else is there? Well, there’s always predicting the future. Yeah, that’s right. Predictive analytics are getting attention. I suspect that we are in the heavy hype phase where it sounds cool but will the delivery match the anticipation? We have enough of an Internet industry track record to know that hype often outpaces reality by a sizable margin (how many “This is REALLY the year of mobile!” claims were made starting in about 2000?).
One of the companies leading the charge into this brave new world is IBM which has released a study predicting a new fashion trend that should be hitting sometime in the future. If you are into that kind of thing it is called “steampunk” fashion. The infographic below shows some of the data that has been used to make this ‘prediction’ of what will possibly have some retail momentum in the near future.
IBM is calling this their “Birth of a Trend” project. Here’s a little more background. Note that the rubber really meets the road in the last sentence.
Through its sentiment analysis, IBM has found that steampunk is evolving into a cultural ‘meme’ via a series of leaps across cultural domains (such as fiction, visual arts, etc). A combination of science fiction and fantasy, steampunk is a sub-genre based around gothic machinery and the industrialized civilization of the 19th century. Rooted in the designs of the industrialized civilization of the 19th century, steampunk is a retro-futuristic style of fashion that is influenced by the works of Jules Verne, Nikola Tesla, HG Wells and more.
Using advanced analytics, IBM has been able to track the spread of trends geographically, chronologically and now, culturally. From 2009 to 2012, the amount of steampunk chatter has increased eleven-fold. Since 2010, more than two dozen US department stores and specialty retailers have become steampunk savvy. During the next two years, IBM predicts that steampunk will shift from low production, high cost “craft” manufacturing to mass production.
That’s bold stuff. Will companies be willing to ‘gamble’ that these predictions are accurate based on the data that produced them? Will they trust this kind of analytic prognostication? Based on what it costs to do these things right and to be as profitable as possible, having more data certainly beats having less. But what happens when the prediction is off? IBM is banking on this being a wave of the future.
“Smart retailers are using social analytics to better understand, predict and shape consumer demand for “must-have” products before a particular trend gets saturated in the marketplace,” said Trevor Davis, Consumer Products Expert with IBM’s Global Business Services. “By staying ahead of a trend as it develops, a retailer can more effectively control critical merchandizing, inventory and planning decisions. Technology can provide tremendous foresight to help businesses differentiate what is a fleeting fad, versus what is an enduring trend.”
Do you see this kind of predictive analysis as something that is ‘of the future’ or something to be looking into seriously right now? How much would you bet on this kind of planning?