The toolbox of an behavioral manager
For over fifteen years, Matt Wallaert has been applying behavioral science to practical problems, from startup exits to the Fortune 500 to an array of pro-social side projects. As one of the first behavioral scientists to leave academia and work in industry, he’s given hundreds of talks on the science of behavior change at the UN, SXSW, and beyond. He was most recently the healthcare industry’s first Chief Behavioral Officer at Clover Health, a multi-billion dollar insurance company, where he directed one of the world’s largest behavioral science teams, combining qualitative researchers, quantitative researchers, and project managers. His side projects consistently focus on the unrepresented, like GetRaised.com, which has helped underpaid women ask for and earn over $3.1B in salary increases.
In his new book, Start at the End: How to Build Products that Create Change, Wallaert details a science-based process to create behavior change that can be implemented in organizations of any size and industry. Filled with colorful examples from his own work and plenty of cursing, the book is an approachable guide to putting behavior back at the center of everything we build.
In this interview, we asked Matt 13 important questions about behavioral mangement, and the tools these utilize and the challenges they face! Here are Matt’s answers.
Why do organizations only succeed with 50% of their change projects?
The answer is in the question. I’m not sure we have any idea of how many change projects succeed, because we have no idea what success even means. If we don’t identify the behavior we want to change and what the world will look like when it has, how can we say anything has succeeded or failed? If we start articulating what outcome we want, we will start to get better at reliably producing those outcomes.
What are the top five frictions to change in modern everyday organizations?
The friction is mostly that people don’t pay enough attention to friction. They focus too much on promoting pressure (“let’s motivate people!”) instead of simply making the changed way easier than the unchanged way. I can’t count the number of times I’ve heard a leader who is baffled why people won’t change describe the new process as “so easy.” Then you get in there and look at it and it isn’t easy at all. You have to create a login, change your habits, figure out connectivity, and a thousand other tiny details.
Think about professional athletes. They practice, day in and day out, to make something automatic precisely so that it can’t be changed, so that it resists outside pressure. Professionals of other stripes are the same. Indeed, a company that wants to execute well tries to run nearly automatically, in a delicate ballet so that everyone knows their part. Is it any small wonder that change is hard? Leaders need to start learning when what they’re asking is a nudge and when it is a rebuild. And when it is a rebuild, they need to commit the resources to doing that, whether that is training staff or accepting a loss in productivity while half the company shuts down and retools. When you step back objectively from most change plans, they seem doomed to failure.
So if we want to talk about change, we should be talking about what we want. Certainly, autonomy has to be a big part of it — people thrive and grow best when they are allowed to make decisions and to act on them. But in order for autonomy to work within an organization, it must be bounded, which then takes us back to clear outcomes. If we are clear about the outcomes we want people to produce, then we can hold them accountable for producing them and give them the autonomy to do so. And when you couple autonomy with accountability, you get the transparency that people love to go on about. We don’t need transparency of process, we need transparency of outcome; we need to know what people are expected to achieve, we need to know if they are achieving it, and we need to know how we handle situations where people do not produce the outcome. Don’t want to spend endless time documenting what you intend to do? Then you have to agree to be measured on the results of those actions. Outcome-based leadership is the only way we don’t all end up in process documentation hell.
What’s the difference between a project management concept like ADKAR and behavior management?
First, ADKAR gets the behavioral part right (in focusing on the behavioral outcomes) but misses out on the science. It presumes that there is a standard template to rolling out change when in reality, we don’t know what works unless we test it and every situation is different. I can understand that — we want generalization because it is easy!
Second, it is far too conscious for me. It has these individual actors, who say things like “I want…” and “I need…” But the majority of behavior change (and the interventions that bring it about) are non-cognitive! Sure, sometimes we have to make people think about things but that should be a model of last resort. Starting from there is a no-go for me.
What is the job of a behavioral manager?
I think this is largely the previous answer but let’s break it down. I think there are two parts to behavioral management. One is managing for behavior change and that’s what I described above: having clearly articulated behavioral goals and then helping people design and deploy the interventions that create them. The other half is managing with behavior change and that is applying the same process to management itself. Take 1:1s. You can’t really have effective 1:1s if you don’t even know why you are having them. You have to say “what behavior do I want from my team?” and then science your way to those outcomes. So you’re actually following the same process as a manager that you would ask them to use to do their jobs.
What is the most important lesson from behavioral science which the behavioral manager should practice every day leading change to lasting impact?
Remember what behavioral science is: behavior as an outcome, science as a process. You should be constantly forcing yourself to examine whether you are truly leading with a clearly articulated behavior in mind and that you’re systematically measuring whether what you’re trying to do is actually creating that outcome.
Secondary to that, it is probably to be wary of the actor-observer bias. We consistently underweight the environment as a factor in understanding the behavior of others. If you’ve done even a halfway decent job in your hiring process, you should have enough high-quality clay to work with; it is up to you to create the environment that shapes it. Yet managers constantly attribute behavior to internal factors like personality. And it isn’t as if the person doesn’t matter; of course they do. But you can’t change the person and you can change the environment, so why not lead with the thing you can do something about?
What characterizes a tool in a behavioral toolbox?
There are two sets of tools that I think are ultimately helpful to creating behavioral outcomes. One of them is capabilities, which are organizational resources that are devoted to behavior change. For example, if I want to change behavior, I need to be able to do qualitative research with the population. So having a program that allows systematic access to those people makes it easier to do behavior change. Same thing on the quantitative side: if you aren’t collecting data or it isn’t well-organized, people are less likely to actually go through the science to make behavior change happen.
The second set of tools are frameworks. Whether that is the Intervention Design Process or the four squares we use for looking at exemplars (those who always do the behavior, those who never do the behavior, those who just started, those who just stopped), having frameworks for how you approach things in a reliable way can help align everyone and help you pick up the pace on the virtuous cycle of pilots.
What’s new in the toolbox, of a behavioral manager compared to the toolbox of a traditional change manager?
I’m not sure traditional change management really has a strong process to begin with. Most the time, when I talk to change management teams, it sounds like throwing spaghetti at the wall or defaulting to things like training (which we know doesn’t really to change behavior a lot of the time, yet we insist on defaulting to it because we have learning platforms that we paid money for). Really, the biggest difference is a scientific approach — it doesn’t matter what strategy you use, what matters is that you evaluate it for efficacy before you scale it.
What is an intervention in behavioral terms?
An intervention is anything you do to change a pressure that changes behavior. So if you want people to eat fewer M&Ms (like Google did), the pressure is “physical attractiveness” and the intervention is “putting the M&Ms in opaque containers”. Putting the M&Ms in containers reduces the promoting pressure that physical attractiveness.
What does a day in a behavioral manager’s day look like, on an hour to hour basis?
It depends where in the cycle of behavior change you are, but assuming that you’re high enough that you’re running multiple projects, there are a couple of key duties. In the early stages, it is stakeholder alignment around behavioral statements. Do you understand what outcome behaviors those above you are asking you to create? And how are you sitting down with your teams to make the sub behavioral statements that will get you there? A behavioral statement truly is a contract (because you are agreeing to be reviewed based on the outcome), so you are making contracts both up and down. Once you’ve got that done, then you’re in portfolio management mode. How are all the various projects going? Are they at the right phase of the process and how are tracking to outcome?
Ultimately, this is advisory — since you agreed on an outcome, you want to give them autonomy to get there, while being as available as possible for advice. It isn’t “do this”, it is “here is what I’d do”. If it is a junior team, you can be more prescriptive (you owe me a behavior statement on Friday, you owe me insights the next Friday, etc.), whereas you can leave a senior team alone with fewer check-ins. And then you’re also resourcing and unblocking. In most large orgs, your team is going to have to get other teams to do things in order to run pilots, etc. and so you are leaning in at a senior level to make sure they get done.
What is at stake when we talk about impact in behavioral change management?
Ultimately, if we accept that everything exists to change behavior, then behavior change IS impact. It is the only actual way to measure impact.
Why should a corporate organization hire a behavioral manager for their change projects?
Once you accept that allow projects are about behavior change then it becomes self-evident why you would hire an expert in behavioral change to guide the projects. In an ideal world, you actually get to a place where you actually pivot the org so that everyone is working this way, since you don’t need a special degree, just adherence to some basic principles and then continued experience down that path. So it may not always need to be some full-time outside expert.
In the effort to reaching impact in terms of behavioral, how could the behavioral manager apply the idea of decision-points as a tool to create the change?
At the crux of where we tend to go wrong today is that we view things as binary: something is either or untrue, works or doesn’t work. It is understandable, because the decisions we make right now are generally binary, like choosing to launch something or not. What we need to do is make that into a gradient. It is like multi-round betting in poker.
You get some cards, you bet based on what might happen, then you get some more cards and bet on the changed probabilities. Same thing in behavioral design: you bet some resources to get more info, then bet more resources base on that info. A pilot is just another way of getting more info. And even once you launch it, you continually measure so that you know when to change the bet. So anytime we see binaries, we want to make scales. The more finely gradiated the scale, the better the eventual outcome. We can become too finally gradiated (such that we lose momentum), but that’s usually not the error I see most these days.
What is the best way to scale behavioral change — and create lasting impact in terms of new habits?
I actually don’t think scale is as hard as people tend to think. In my experience, you can get an excited org to uptake a change to behavioral science as a method in six months or so, and then success just depends on commitment to building out the capabilities that let people turn the crank on that flywheel better and faster.
What would Kotter have done if he had a behavioral change tool?
Agree on behavioral statement (including population, motivation, limitations, behavior, and measurement). Form insights on why anyone would ever want the changed behavior and why they aren’t already doing it. Map pressures against those insights and design interventions. Pilot those interventions. Kill what doesn’t work, fill what does.
