In the final weeks of a passing year and the first few weeks of the New Year, journalists, analysts and pundits alike tout the “big trends” for the upcoming 12 months. More often than not each article that passes through my RSS feed is touting a new trend discovery technique, proprietary to the author of course, that will bring amorphous human behavior into striking trend clarity hundreds-fold faster than previous techniques. Favored trend techniques of years past have been tools like Google Trends, Twitter Trending Topics, even Facebook’s “year in review” function was a trend discovery tool that appeared as a latecomer to 2012.
As we enter 2013, it is no surprise that IBM is sharing their “Birth of a Trend” tool that by using natural language processing, a targeted research set and, one can hypothesize, their Watson technology is able to analyze the trajectory of a trend and its adoption by digital denizens. As someone who works in advertising and marketing, this technology is pretty freaking cool. As a trend forecaster, I am philosophically calling their bluff.
Spotting cultural trends is inherently a human talent and has been for eons if we consider the practice abstractly. We are evolutionarily rewarded for our ability to spot patterns that are favorable to us, whether through avoiding danger or cultivating prosperity. Identifying cultural trends, changing behaviors in groups of people bound by similar attitudes and beliefs, is large scale pattern recognition with accolade going to the forecaster that can reconcile the most divergent pieces of evidence into a view of how culture is changing and will change over time. This practice differs from quantitative methods, like Google Trends or IBM, in that cultural trend forecasters focus on what is different about a few individuals rather than what is similar about many. By taking this stance, cultural trend forecasters are lauded for their intuition and ability to identify trends far before they hit the radar of qualitative analysts.
But trend discovery shouldn’t be a battle of human and computer – both are needed for successful forecasts. Humans are great at quickly spotting many potentially meaningful trends and computers can tell us when the chatter amounts to something meaningful. Our expertise in both areas is the key to reconciling both approaches: culture-led and data-fed. We identify trends in counter culture and track their development over time so that we can develop executions that are based on emerging trends – not emerged trends.
Going forward, expect to see more tech companies and data hoarders release algorithms that claim to predict human behavior—Minority Report style. However expect humans, which are not entirely predictable, to continue to be algorithmic outliers, skewing results and causing models to be revised. Methods that can reconcile quantitative and embedded cultural insights, graphs and guts, are best positioned moving forward. Whether to derive insight from anomalous human behaviors, sift meaningful information from a data set or use predictive modeling that accounts for and can react to multiple variables. No matter the trend-tracking fad of the moment, combining both quantitative analytics and embedded cultural insights will prove to uncover more impactful trends earlier and with stronger implications for society and brands alike.