Retail Analytics Gain Steam(punk)
Victorian clothing used to lie solely in museums or in historical depictions on film and television. IBM announced this week that an altered version of the style, called ‘steampunk’ is on the verge of being mainstream and, according to them, they saw it coming three years ago.
The steampunk statement harkens back to the days of Victorian England, a time where mercantilism was the prevailing economic/foreign policy theory and working industrials joined small farmers (not small people who farm but people who farm a small amount of land) as a leading occupation.
Steampunk, according to an IBM infographic, gained cult revival in 2006 after retro-futuristic depictions in novels and film. The trend really took off in a fashion sense after the New York version of ComicCon in October of 2010, after which conversation through social media spiked 300% and never slowed. “We’ve been tracking steampunk since 2009 when really there were just a few hundred, a few thousand posts to look at through to today where we’ve seen a 12-fold increase in chatter,” said IBM Consumer Products Expert Trevor Davis in an interview with Datanami.
“No retail trend is born in isolation,” claims the infographic. This announcement of the mainstream arrival of steampunk is part of a larger IBM initiative called “Birth of a Trend.” In essence, they are trying to leverage big data to not only pick out upcoming retail mainstays but to also look deeper as to why these trends show up in the first place.
To an extent, according to Davis, it has a lot to do with similar economic and social outlooks to that of the 18th and 19th centuries. “We’re trying to look inside short fragments and say ‘what is the context that is at work here?’ There are parallels with what happened in Victorian England,” said Davis.
Even though the technologies and economic situations were clearly vastly different to how they exist today, the core undercurrents of explosive growth, new wealth, and resulting uncertainty remain the same. As Davis explained, “You had the telegraph, a kind of technology explosion. In the mid-1800s there was a banking crisis so there was a recession and uncertainty. Looking at the social issues people were worried about back then, it was things like job security, poverty, national conflict. So there are a lot of parallels on a socio-demographic level. That makes steampunk work in the psyche.”
Of course, steampunk is not a simple rehash of the styles of Victorian England. It represents a callback to how the science fiction writers of the time imagined the future. Corsets, a staple of the noble leady in the Victorian age, are modified to reflect darker, more sheik elements imagined by Jules Verne.
To identify these underlying factors, IBM used a combination of advanced text analytics and algorithms to determine the root, the “birth” as IBM calls it, of the trend. “What we do is a technique called octagonal filtering, which is really three levels of keyword and regular expression searches,” said Davis on IBM’s process. “That allows us to break the text up into increasingly finer grained data to look at. When we’ve done that, what we’ve got is something very precise in terms of what it’s representing.”
Davis’s argues that any retailer can use simple keyword searches, track them over time and arrive at some helpful pure statistics. According to the infographic, chatter regarding steampunk has increased 12-fold since 2009. Twitter houses six times as much of that chatter as Facebook. A retailer would hypothetically be able to compile statistics similar to those and react accordingly.
However, without a more advanced analytics operation, it would be difficult to find these trends as they are happening or select potential related fads. “To really do the work in realtime on this kind of data, you do need to have access to some form of high performance computing cluster or at the very least a very big server,” said Davis.
IBM has the resources to pull off big analytics, but so to do the retailers according to Davis. The disconnect lies in the perceived best use for their high performance computing. “All of this probably sounds like science fiction to the retailer at the moment but I think you will see retailers as they grapple more and more with big data challenges thinking, ‘well maybe some of these technologies that we developed for research science are the kinds of technologies that we need.’”
Retailers have indeed been using big data analytics to track customer sentiment in an attempt to better advertise to them. However, this approach of finding popular statements before they happen can better prepare corporations for impending production and marketing plays.
Currently, IBM consults retailers on these matters but Davis predicts that it will be a matter of time before organizations turn their own analytics power toward trendspotting.