chef reading recipe book surrounded by ingredients

Behavioral Leaders on Starting Behavioral Units

bob_sullivan
Bob Sullivan
Writer

Editor’s Note: Ohmygosh, there’s so much to say about starting a behavioral unit within an organization, but I’ve decided it’s best to let leaders who’ve done it speak for themselves.

Mea Culpa: I’ve had a version of this piece in the hopper for a long while. ‘Twas given a kick by McKinsey’s report earlier this year.  We liked that, but needed more, wanted to get more granular and specific. Got another, ahem, nudge when we heard about the coming survey by Action Design Network and Steve Wendel.  So …

“How to get behavioral science into our work” is obviously on the mind of a lot of industry leaders – that’s why PeopleScience exists. I’ve just been hesitant to pull the trigger on this article because we keep learning more, and I want to solve everything, perfectly, now.  But, as someone may have once said, “Perfect is the enemy of progress.”  So here’s our almost perfect piece. I hope these leaders – you may recognize some! – move the conversation forward and that others will chime in. I’d never claim to be the definitive voice on anything; we’re honored just to be part of the choir. 

 

Uber had a problem. When the ride-sharing firm launched its new Express Pool service, riders kept ordering cars, then abandoning their purchase before a driver could be found. Express Pool is a bit more complex than Uber's traditional service, as it matches up riders going similar places. That saves riders money, but there's a catch: They have to walk a little to meet the driver, and they have to wait while Uber's systems search for compatible companions. That delay drove consumers to cancel.

But Uber also had a "secret weapon" - a behavioral economics team that was empowered to run experiments on new products. The group quickly identified three concepts that might be at play with the Express Pool problem: idleness aversion, operational transparency and the goal gradient effect. Working with the programming team, they ran A/B tests on simple fixes, such as "showing your work." By pulling the cover back on granular steps, like "How we calculate your arrival time," the BE team was able to produce an 11 percent reduction in post-request cancelation.

It was a quick win.

Stories like these are making the rounds among management teams across industries, as behavioral economics ideas are working their way from academia, through popular business books and into workplaces. Tales of successful behavioral “nudges” seem to be everywhere – like default settings that encourage employees to join retirement plans. So many executives are wondering if these techniques can be translated into bigger wins, both among their employees and their consumers. Many firms are hiring behavioral economics experts and forming teams designed to look for such wins. They're often called "nudge units," but if they are formed well, there's a much bigger idea behind them: BE teams are trying to spread the scientific method across every part of the workplace, in an effort help companies maximize sales and help workers be their best.

 

 

BE teams are trying to spread the scientific method across every part of the workplace, in an effort help companies maximize sales and help workers be their best.

Earlier this year, a McKinsey study of nudge units offered a helpful list of questions that management should ask when forming such teams: Where do they sit, who should be on them, how do they handle ethical concerns? PeopleScience.com reached out to a wide set of practitioners to see how they answered these questions and learn what other concerns they had along the way.

What? Where? Why?

As with most business propositions, everything begins by being clear about what problem you are trying to solve. Answering that question about your nudge unit can be tricky, however.

“No one has a behavioral science problem in the same way they may think they have a marketing or an IT problem,” says Matt Battersby, vice president, chief behavioral scientist at the Reinsurance Group of America. “They have a problem that often only we know behavioral science might solve.”

 

Everything begins by being clear about what problem you are trying to solve.

There are plenty of existential questions. There's the obvious, but overlooked – are we trying to nudge consumers or employees? (Ideally both, but you have to start somewhere.) Do we hire a bunch of PhDs? (maybe, maybe not). Is the team a central resource for all, like a consultancy, or is it embedded inside product teams, to understand their needs better?

Or, as author and independent consultant Sonia Friedrich suggests, is it sometimes easier to start with an outside firm?

“Clients use me … to see what it’s all about and get success. I find this a faster and lower cost way to get results and prove the value of BE to the organization,” she says.      

As for why, it’s critical to understand where the desire to start a nudge unit is really coming from – are you proactively innovating or rescuing a sinking ship? The distinction matters: Struggling firms often don’t have the patience required for the scientific method.

 

 

It’s critical to understand where the desire to start a nudge unit is really coming from.

“Some do not have the luxury of time and money to do a full baseline … Others want to try too many nudges at once, which means you get zero proof of what worked,” Friedrich says. “You have to hold them back. That is your job.”

As for the “who” of nudge units – Academics and PhDs might have the deepest research background, but they have to be balanced by behavioral economics workers who have entrepreneurial spirit and those who can be persuasive. As Charlotte Blank, chief behavioral officer at Maritz says, much of the job is “selling” science.

Battersby talks about using two distinct approaches: sometimes his team serves as mechanic, other times, a mechanical engineer. Some processes just need to be fixed or tweaked, like a tune-up. Perhaps a few words need to be changed to reframe an incentive. But sometimes, a process needs to be torn down and completely reinvented.

 

 

Sometimes the team serves as mechanic, other times, a mechanical engineer.

“That means working with the business to ensure that a better understanding of what drives human behavior is baked into the first stages of designing new products,” he says.

That’s why it makes sense to understand desired outcomes and potential language conflicts that can lead to institutional blockages, such as “we're already doing that” or the “not invented here” syndrome, warns Florent Buisson, behavioral scientist at Allstate.

Where nudge units “sit” can help with some of these difficult conversations. An independent unit that pushes in to solve problems has its own benefits – that setup might encourage outside-the-box thinking, the mechanical engineer in Battersby’s lingo. On the other hand, a BE team that’s within a unit might do a better job of understanding its goals and seeing beneath the surface.

 

 

Where nudge units “sit” can help with some of these difficult conversations.

Enrique Saborit, a behavioral economist at BBVA, created his relatively new team inside the chief marketing officer’s team, for good reason.

“Luckily, we have a chairman that is really into behavioral economics and gave us a clear mandate: help customers make better decisions," he said. Still, BBVA isn’t very siloed, he said, so he doesn’t find that limiting.

Buisson thinks the “where should it sit?” question might be a bit malformed.

“In many cases, the question ‘where to house the behavioral science team?’ is moot, because the initiative originates from a leader in a specific department, such as data science or customer experience,” he said. “The question for a behavioral science leader is not ‘where would I want my team to be?’ but ‘how can I make the best of where it is?’ Answering this question is a complex balancing act.”

What are you really up to?

The power of behavioral economics opens up a lot of ethical questions – using science to manipulate consumers or workers, however well-intentioned, is a power that should be managed with great care. One false step can make teams or consumers skeptical of what you are up to. Trust is essential for buy-in and to keep your experiments valid, to keep workers devoted to the “new way.” Gaining trust begins with early agreement on goals and expectations.

 

Trust is essential for buy-in and to keep your experiments valid, to keep workers devoted to the “new way.”

Quick wins are a big aid when building trust, Friedrich says.

“There are plenty of BE nudges you can use internally to show real impact to the bottom line quickly, from debt collection to reducing employee expenses to reducing stock piling of annual and long service leave, to name a few,” she said. “Use one. Show the measured benefit. Then ask to do the next one.”

Framing the experiments also helps. Sometimes it’s best to avoid arriving “guns a-blazing” as a disruptor. Explaining a nudge as a variant of “something we've done all along” can ease the anxiety, says Sam Tatum, head of Behavioral science practice at Ogilvy Consulting.

“Don't walk in and say ‘stop and change.’ (Say) ‘let's refine and add,’” he said.

It’s also important to manage expectations. Science does take time; experiments have to run their course. (Editor’s note: See “Testing at the Speed of Business.”) Corporations, and workers, can grow impatient. To deal with this, author and consultant Steve Shu has developed a “GRIT” model – Goals, Research, Innovation, Testing. It’s “essentially organizational fortitude to implement behavioral science,” Shu says.

 

 

It’s also important to manage expectations. Science does take time; experiments have to run their course.

“I typically have found that nudge units work best when they try to maintain a level of being project-focused,” he said, “and avoid lumping of bunch of activities together.”

But on a more basic level, it’s important to convince both workers and consumers that you are running experiments for them, not on them.

“I still worry that behavioral science efforts could easily be misconstrued if they are not properly presented … especially in the financial industry,” frets Justin Lindemann of Navy Federal Credit Union.

Behavioral scientist Garrett Meccariello also warned about the problem of behavioral snake oil salesmen who might hurt the credibility of BE ideas.

“What the industry needs is some form of credentialing or governing body to ensure that applications of behavioral science are not perceived as hocus-pocus after exposure to less qualified pitch/proposal presentations,” he said.

Blank, of Maritz, isn’t really a big fan of the term “nudge unit” because she thinks it’s limiting. Nudges suggest small changes, she said.

“Let’s make this one change, and improve this one metric. Let’s use social proof to get signups for a clinic, or let’s use a swat team to come in and make quick optimizations,” she said. Blank has grander ideas for what behavioral economics can do for a company.

“This is more foundational for us … What I am really selling is science,” Blank says. Many firms employ the scientific method when creating new products or spending money on marketing but fail to use those refined techniques internally, to help get the best out of their own employees. She’s on a mission to bring science to every part of the organization, so they are “optimized to unleash potential of people.” (Editor’s note: See also “Charlotte Blank Wants You to Lead Like a Scientist.”)

 

“This is more foundational for us … What I am really selling is science.”

Science, but not too much.

One proof that science, and not sales, is at work: failure. Good scientists are willing to admit failure and learn from it. (Editor’s note: See “Not All Failures Are Created Equal.”)

That’s what happened to Maritz when it tried to utilize the concept of endowed progress in sales incentives. Rather than provide auto dealerships with sales bonuses at the end of the month, Maritz experimented with providing funds earlier, with a “clawback” threat that the bonuses would be reversed if sales goals were not met. The theory of endowed progress suggested sales teams would work even harder to keep what they already had. But reality had other ideas.

“We found it backfired. It caused sales to go down. That was surprising to all of us,” Blank said. “But we learned a lot about incentives from it, and we continue to learn.”

Blank talks often about testing as paramount in any new program.

“Good thing we limited our exposure and tested,” she said. “In the real world, things get complicated and messy. You need to test in your own environment.”

Of course, the scientific method isn’t easy or cheap, so it often runs into reality. Randomized Controlled Trials, RCTs, are the gold standard, but many firms won’t have the time and the money. Is there a middle ground? Tatum, from Ogilivy, says yes.

“(Maybe) 90% confidence is good for them,” he said. “Good for business, but maybe not good for cancer research.”

Finding the balance between the ideals of science and the just-in-time needs of industry is a high-wire act, but it’s also exciting for both sides. Blank’s Maritz team maintains a list of academics ready and willing to work with companies in exchange for access to research results. Corporations are often surprised at how easy it is to find academic research that’s already asking the same questions they’ve encountered in their business. Balance comes from testing theoretical notions in a real-world environment.

“A behavioral science unit needs to be a muscle as well as a brain,” Battersby said. “The focus should be on applying knowledge.”

 

Finding the balance between the ideals of science and the just-in-time needs of industry is a high-wire act, but it’s also exciting for both sides.
bob_sullivan
Bob Sullivan
Writer

Subscribe

Get the latest behavioral
science insights.