stevanovicigorIn 2016, the National Environmental Research Council (NERC), a quasi-governmental agency in the United Kingdom, decided it would be fun to let the public vote online to name the country’s newest research vessel. The agency was less pleased when it saw the winning entry: Boaty McBoatface.
Overruling the public’s wishes, NERC named the craft after British naturalist Sir David Attenborough. The public was outraged; newspaper editorials decried the lack of democracy, and citizens protested the unfairness of it all on social media. So much for having a little marketing fun with the public.
Boaty blowback highlights the potential danger of giving consumers the power to vote, even though customer engagement is a primary goal of almost every social media strategy.
The problem: Even though NERC never explicitly promised it would name the boat after the winner in the online poll, the agency implied that it would respect the public’s wishes, say Michael Norton and Leslie John, both professors at Harvard Business School.
“When firms conduct online polls, people frequently submit ridiculous entries; and with social media, those entries will go viral,” says Norton, Harold M. Brierley Professor of Business Administration. “But even when firms never guarantee that consumers will choose the winner, consumers infer an implicit contract and are upset when that contract is violated.”
Norton and John examine the pitfalls of those implicit contracts in a new working paper, Procedural Justice and the Risks of Consumer Voting, written with Darden School of Business Assistant Professor Tami Kim and Harvard Kennedy School Professor Todd Rogers.
Name this space
These days, firms encourage customers to vote on everything. You can thank online polling for such products as blue M&M’s, Spicy Street Taco Dorito chips, and even a set list for the band Metallica. The practice is appealing to companies for several reasons, Norton says.
For one, internet-based polls generate near-instant results. “Before, firms had to conduct these campaigns via the mail,” Norton says. “With social media, responses are immediate.”
Also, votes allow companies to perform consumer research cheaply, generating new ideas for products or flavors, or testing consumer preference for options it might be considering. “Firms can essentially outsource product development,” says Norton.
Allowing consumers to vote also engenders brand loyalty by giving users more of a sense of ownership over products.
Once a company opens those floodgates, however, they are essentially making a series of promises with consumers, whether they know it or not.
Norton and colleagues divide those promises into three categories: representation, consistency, and non-suppression.
Representation refers to the fact that if customers are given input into a product choice, they expect it to count as much as—if not more than—the company’s own input. In one experiment, participants were given the chance to vote on the Super Bowl Most Valuable Player award; they were told their votes would count anywhere from 0 percent to 100 percent of the final outcome, with the rest of the vote decided by a panel of experts. In another condition, participants were told their feedback wouldn’t influence the vote at all.
The researchers then asked participants to rank from 1 to 10 how satisfied they were with their votes. Those who had been told their votes wouldn’t count at all ranked their satisfaction at 5.47, much higher than the 4.04 satisfaction score recorded by participants who believed their votes would count 10 percent. In fact, participants’ satisfaction didn’t match the “feedback” condition until after they were given 50 percent or more input.
“As soon as firms introduce voting, they are implying a different relationship: ‘We are working together on this,’” says John, Marvin Bower Associate Professor. “Therefore, consumers believe their votes should be counted equally.”
A similar phenomenon occurs with consistency. Once a company asks for feedback on one choice, consumers assume that they will be able to offer input again in the future. To test this condition, Norton and his colleagues created an experiment featuring a fake Kickstarter campaign for a company called Ozzie’s Organics and allowed some consumers to vote on the company’s product lines. Afterward, those participants were told either that the company would continue to allow consumers to vote or it would be taking the vote away. Those participants who lost the vote subsequently rated their approval for the company lower than both those who retained the vote and those who never had it.
The last category, non-suppression, is the principle that was violated in the McBoatface incident, Norton says. For this experiment, the researchers read the NERC case study to participants, in some cases giving them the real vote totals of Boaty McBoatface at 124,109 votes, and David Attenborough at 11,000 votes, and in some cases reversing them. In a third condition, they were told that NERC named the boat without a vote. In all three cases, they were told NERC named the boat after Attenborough.
Mirroring reality, participants were significantly less satisfied with the outcome when they felt NERC overrode the vote—3.63 compared to 4.80 when told they abided the vote, and 4.58 when told they held no vote. In addition, they were less satisfied with the firm itself, 3.26 compared to 4.80 and 4.38 respectively.
“It’s offensive because consumers feel as though the firm broke the contract,” John says.
The right way to run a poll
For companies looking to let consumers vote on choices, Norton and John recommend that they be very explicit about how they will consider the voice of consumers before the vote begins. For example, if the vote is a one-time event, the firm should be sure to let consumers know that beforehand, to save on disappointment if they aren’t asked for input the next time.
In addition, companies can limit the fallout from off-the-wall choices by pre-selecting acceptable outcomes on which consumers can vote—or culling options from consumer suggestions without publicizing the actual number of votes for each. “If you love one that only one person suggests, you can include it without offending anyone,” Norton says.
Lastly, companies can set up some kind of screening process, by which only actual customers can vote, but giving out a voting code attached to products. That way, those voting are more likely to be invested in the actual outcome and may be less likely to deliver inappropriate responses.
If worse comes to worst, you can always just give people what they want.
That’s what the Australian government thought it was doing when it named a new ferry in Sydney Ferry McFerryface, telling the public that the people had spoken. It was later revealed that the government had chosen that name over the actual winner, an environmental activist.
Despite the fact that the government thought the humorous name would be more popular, once again, an implicit contract had been broken, and outrage ensued.