I recently read an article in The New York Times about the ways in which companies use social media sites like Facebook, Twitter, and Foursquare as extensions of market research departments. The article described how one year ago, Walmart purchased the social media company Kosmix for $300 million primarily because Kosmix has a knack for extracting trends from social media conversations. According to the NYT:
The unit, now called @WalmartLabs, looks at Twitter posts, public Facebook posts and search terms on Walmart.com, among other cues, to help Wal-Mart refine what it sells.
In one of its first analyses, performed last summer, @WalmartLabs found that cake pops — small bites of cake on lollipop sticks — were becoming popular. “Starbucks had just started getting them in their cafes, and people were talking a lot about it,” Mr. Raj [VP for products] said.
His team alerted merchants at Wal-Mart headquarters. The merchants had also heard about the product, and decided to carry cake-pop makers in Walmart stores. They were popular enough that the company plans to bring them back this holiday season.
Now, while the Walmart purchase of Kosmix is not a deal of Instagram-like proportions, a $300 million investment in social media is significant. I’ve long been a proponent of monitoring tweets and Facebook posts to gauge public sentiment, and the aforementioned cake-pop anecdote certainly makes it seem that following the rises and falls of social media chatter is one way to unearth the latest and greatest new thing. I think it’s indisputable that people weigh in on new fads and converse with one another about their favorite products through tweets or status updates. And Walmart isn’t the only business employing social media listening strategies. WiseWindow is a company that uses real-time information (such as that found on social media sites) to predict and measure consumer purchasing and behavioral intent around different products. Derwent Capital Markets offers a “sentiment analysis app,” which analyzes social media to derive buy and sell signals for investors and traders.
Wait a Minute, Walmart
Yet, rather than nodding my head in agreement as I read about these social-media-analysis companies, I found myself doubting my preexisting convictions, and here’s why: extricating trends from social media conversations is grounded on the idea that Twitter and Facebook can portend the future. Walmart Labs listened in on digital discourse and discovered the popularity of cake pops. They then decided to carry cake-pop makers in stores, based on the assumption that people would flock to the pop makers the way girls flock to Ryan Lochte’s Twitter page (#jeah) , based on the assumption that social media can predict future buying habits of consumers. I see a problem with this. People may tweet about the charming pastel color and irresistible deliciousness of Starbucks birthday cake pops, but can tweets of this variety really be viewed as signs of a larger trend? Just because someone talks about cake pops doesn’t necessarily mean they will rush out and buy a cake-pop maker.
I think designer clothes aptly illustrate this disconnect between prattle and purchasing. Many people gush over high-end fashion labels, but the percentage of the population that can afford to routinely buy designer apparel is extremely small. Thus, the people talking about how much they love Louis Vuitton or Chanel most likely own only one or two pieces by these brands. If any of these brands decided to carry a new line of products based on tweets or Facebook statuses, they might end up sorely disappointed because people talk about these products more than they purchase them.
I think this phenomenon extends to various genres of consumer products. It throws a wrench into the idea that social media can forecast future trends. But then again, why would Walmart shell out hundreds of billions of dollars to buy a company whose existence is predicated on a shaky concept? Although, investing money and marketing-research resources into ultimately unsuccessful products and endeavors is not uncommon. Does anyone remember Pepsi AM or the Cocaine energy drink?
So, I wanted to determine if social media truly can serve as a crystal ball when it comes to product trends and consumer buying habits. Can social media foretell the future?
Social Media Is the New Nostradamus
There are actually a number of studies that indicate that social media is a kind of modern-day oracle at Delphi:
A study entitled “Predicting the Future with Social Media” found that Twitter chatter can be used to predict box-office revenue for movies. Researchers examined three million tweets and discovered that they could use the amount of buzz and attention on Twitter surrounding a specific movie (such as Twilight: New Moon and The Blind Side) to predict its ranking at the box office before its release. In fact, their predictions on which movies would generate the most box-office revenue were consistently better than those generated by the Hollywood Stock Exchange, the world’s largest prediction market for movie box-office forecasts.
Another study found that Twitter can predict the stock market. This study tracked collective mood on Twitter to see whether it correlated with the value of the Dow Jones Industrial Average (DJIA) over time. Researchers looked at six dimensions of mood (calm, alert, sure, vital, kind, and happy) and found that the degree of “calmness” on Twitter can predict the daily up and down changes in the DJIA closing three to four days in advance. One of the interesting things about this study is the fact that the calm mood dimension on Twitter can predict DJIA values with 87.6% accuracy. A similar study from Pace University found that social media could be used to predict the rises and falls of stock prices for Starbucks, Coca Cola, and Nike.
An article on Bloomberg talked about how Twitter could have been used to tip investors off on the plummeting price of Netflix stock. When Netflix decided to split its DVD and streaming business, Topsy analyzed public reaction on Twitter and found signs of an impending fall in stock price. In fact, Topsy is so confident that their tracking of negative and positive sentiment on Twitter can anticipate stock-market moves that it is building a “Twitter sentiment analysis” service to help investors stay ahead of the stock-market curve.
A Cloud in the Social Media Crystal Ball
Now, all of this evidence is very compelling and strongly suggests that social media can in fact anticipate trends, whether it’s the DJIA, stock prices, or cake-pop crazes. However, the above studies are not the whole story. I recently read an article on TechCrunch that portrayed social media as a kind of funhouse mirror: it distorts our view of the world. Gregory Ferenstein talks about the recent buzz surrounding Chick-fil-A. The media has been all up in the chicken-chain-restaurant’s business because of the anti-gay remarks made by the CEO. Ferenstein points out that blogs, Twitter users, and city mayors have used social media to voice their Chick-fil-A boycott plans.
People even planned a “Same Sex Kissing Day” at the chain. Yet, despite all of this hype, the restaurant recently enjoyed record-breaking sales! Ferenstein points out that social media makes it seem like the entire world is against Chick-fil-A and that consequently Chick-fil-A will suffer financially, but that perception could not be further from reality. Thus, social media doesn’t give us an accurate view of what’s going on in the world. One of the explanations Ferenstein gives to this social-media distortion is the fact that young people, who dominate social media, “have a bigger bark than bite.” They talk about things a lot, but that doesn’t necessarily mean they act on their words 100% of the time. With Chick-fil-A, Tweeters and Facebookers may have enthusiastically broadcast their support for “Same Sex Kissing Day,” but that’s not a guarantee that they will show up and take part.
This is what I was trying to get at with my spiel on designer clothing: I may talk about how much I love Trina Turk or Michael Kors, but that doesn’t mean I’m going to physically go out and buy something from these designers. There is oftentimes a disconnect between words and actions, and I think this is why attempting to predict the way people will act and the merchandise they will buy from the tweets/Facebook statuses they post is risky. It worked for Walmart with the cake pops, sure, but I don’t think it will always work.
I also think this idea that social media misrepresents reality is spot on. Consider this example: if I spend a few minutes on Facebook and Twitter, it begins to appear to me that everyone and their mother’s sister’s Lhasa Apso has a Facebook profile and/or Twitter account. But actually, only around 50% of the U.S. population is on Facebook. It’s the same thing with Pinterest. When I’m on Pinterest, the numerous comments on pins and all of the heightened repin activity make it seem as though Pinterest is used by everyone to get recipes and DIY ideas. Many of my friends even sing the praises of Pinterest on Facebook. But, in actuality, less than 4% of the U.S. population uses Pinterest. If I didn’t do my own research and relied only on the impression I received from social media, I would assume that upwards of 50% of the population uses the site. I would have fallen victim to the social-media-funhouse-mirror phenomenon.
Twitter > Miss Cleo
I think social media certainly can be an accurate predictor of the future and can be used as a trend-discovery tool (as the research shows), but, there are some risks to deriving brand intention from social media. If stores like Walmart can anticipate the latest fad and stock their stores with items that are all the rage, then social-media-sentiment analysis is certainly a wise investment.
And I almost neglected to mention something I think is important to remember: the social media oracle is a far better way to envisage trends than say, consulting a psychic.
With Twitter and Facebook, marketing research departments won’t have to worry about working with someone who was slapped with a $5 million fine by the Federal Trade Commission.
They also won’t have to deal with fake Jamaican accents.
Cool runnings, mon. Bobsled.