There’s been no shortage of finger-pointing in the eight months since Donald Trump stunned the world by beating Hillary Clinton to become the 45th president of the United States.
For political junkies, the culmination of that debate over how presumed front-runner Clinton lost the election to a political novice was the publication of Shattered, a campaign tell-all chronicling the infighting that purportedly paralyzed the candidate’s team. Since the book hit shelves in April, the conversation has shifted toward data: Clinton herself cited the Democratic National Committee’s “mediocre to poor, non-existent, wrong” data operations when discussing her loss with Recode’s Kara Swisher.
The DNC quickly moved to defend itself, but Clinton’s statements hinted at something that has haunted marketers since Election Day: In an era of increasing reliance on analytics, why did the numbers predicting a solid Democratic victory prove so wrong?
“The ways [Big Data] got used and abused are strikingly familiar to some in the marketing world,” said SS+K partner Rebecca Matovic, who compared the Clinton team to a junior high school class, noting that “sometimes people rely on data when they’re unable to have conversations with each other.
“Data doesn’t say anything interesting about innovations, how to combat a populist … or new products that people aren’t familiar with,” she added.
According to this argument, the Clinton team focused so much on running the numbers and playing to its base that it largely avoided the hard work of converting undecided voters. Some pointed to collaborations with Jay-Z and Lena Dunham, both of which were reportedly planned after the campaign’s custom “Ada” algorithm suggested the celebrities as good ambassadors for the Clinton brand, as examples of this shortcoming.
Political strategy and marketing are distinct disciplines, but they share common goals, chief among them connecting to targeted audiences on a personal level. In both industries, some argue that analytics have overpowered emotions.
“Maybe [the election] was a good wake-up call to marketers who pray solely to data as their god,” said Liza Landsman, president of Walmart-owned ecommerce company Jet.com. “A pure focus on science as opposed to a blend of art and science leads you to false conclusions, which can be very dangerous.”
Jet.com avoids what Landsman calls “very wonky decisions” by relying on its own in-house call centers and customer research labs, because no amount of well-curated data can substitute for real-life interactions with consumers.
Some in the agency world remain deeply skeptical of any implied link between advertising and politics, however. “Most people talking about their views of data have no idea what they’re talking about,” said Slavi Samardzija, global CEO of Omnicom’s data unit Annalect, “and this connection to the election is made by the same people.”
Samardzija calls opinion polling—which is the bread and butter of party politics—“an art of its own,” adding, “I have never used a survey asking for feelings. Should [the Clinton campaign] have been looking at behavioral data? I don’t know, because I’m in the business of using that data to help clients sell products.”
Yale School of Management professor Barry Nalebuff disagreed, arguing, “Neither politicians nor marketers really understand the data.” As an example, he mentioned a study in which Coca-Cola tried to figure out whether kids wanted a sweeter version of its Honest Tea line before determining that parents’ preference for a tea with less sugar was actually far more important than their children’s opinions. In the end, the company had lost time and money gathering data by asking the wrong question.
Likewise, in the case of Clinton’s loss, “I don’t think the message is that data is bad or that it failed,” explained Matovic; instead, Clinton’s team “clearly” used the wrong messages to pitch the public.
For Marc Simons, co-founder of media agency Giant Spoon, the election’s biggest take-away was that “intentions and decisions aren’t always in sync—and this applies to our business in a big way too. The crisis of confidence in Big Data should be more of a realization that the data can’t tell you everything and is not a crystal ball.”
While Big Data might provide a convenient scapegoat for Clinton’s loss, Yale’s Nalebuff still doesn’t believe it’s going away anytime soon. “In the meantime,” he opined, “a huge amount of money will be spent and wasted.”
Don’t miss the Adweek Livecast webinar on Aug. 22 with IBM chief digital officer Bob Lord on how marketers can succeed in an AI-driven market. Register now.