Hacking inequality and behaviour
Matt Wallaert is a behavioral scientist at Microsoft, success lead at Microsoft Ventures, and a founder of startups including @GetRaised, which helped women earn $2.3B in raises. Matt is also working on #IAmListening, an upcoming Twitter-oriented web app that encourages people to think critically about the demographics of the voices they are listening to.
What’s your story?
“…morning glory! What’s the tale, nightingale?” I never really liked Bye Bye Birdie but that line has been stuck in my head since high school.
What’s my story? Work is probably the central theme, and movement. I grew up in rural Oregon to amazingly awesome parents. First generation college kid. Was on an academic track in social psychology, primarily studying decision making and choice environments, but got tired of writing papers that only academics read. I wanted to help people actually make better decisions. Got an opportunity to do a startup, found success in that, sold it, did it again. Ended up at Microsoft because I realized the size of the problems you get to solve is generally proportional to the size of the organization you get to work with, so to level up to the bigger problems, you almost always have to level up to a bigger place.
Just moved back to NYC with my wife (also a psychologist), have a six month old, loving spending time with him and figuring out how to be a parent. I think I’ll get the hang of it right around the time he moves out.
What does the role of a behavioral scientist involve?
It depends heavily on where you work and what you work on, as it is a bit of a catch-all term. Properly said, I’m a social psychologist but if you tell people that, they tend to think ‘therapist’, which I’m not.
But when I think about the broader behavioral scientist category, I think it helps to break down the pieces. The ‘behavioral’ is really about what we study — human behavior. As opposed to cognition, perception, emotion, etc. Thinking can be a behavior, and so can seeing and feeling, but generally speaking, behavioral science doesn’t care about thinking, seeing, or feeling until they actually modify an observable behavior. We don’t care that you feel sad as much as we care that you act sad.
The ‘scientist’ is the method. There are lots of way to observe human behavior; artists do it all the time in wonderful, successful ways. What differentiates what we do is the application of experimentation and theory to those behaviors. There are some who argue against the use of the word science because unlike a chemical reaction, it is difficult to precisely control humans, but I think that’s semantics: compared to most other explorations of the human condition, what we do is firmly in the science camp.
So if we think of the scientific study of human behavior, it crops up all over. Some people end up in marketing departments and they look at purchasing and usage as behaviors. Some people end up in government, doing public policy work. It really depends on what the behavior of interest is — the methods are similar throughout.
How does behavioral science influence the products and services people use?
I think Malcolm Gladwell does a much better job than I do at telling stories about behavioral science, so the first thing I’d say is to go and watch his TED talks, read his books, etc. I’m a great scientist and a lousy storyteller.
Though there are some great scientist storytellers, so I also think you can go to the source. I keep a list of books worth reading on my website, but Barry Schwartz, Adam Grant, and Dan Ariely are all worth reading and watching.
Personally, if there was one thing I could teach people about the influence of psych on products, it is that our notion of human behavior doesn’t have to be limited to the very obvious and conscious. Often when I talk about studying how people make choices, people think about the explicit choices: this versus that, what to buy, picking off a menu. But what about all the very unconscious choices that we make all the time, that are heavily affected by our environments? Things like word choice, how we engage with other people. Do we say please and thank you? How is the influenced by the other person? The world? Our internal states, like hunger?
If you start to understand that everything influences everything else, that our behaviors are rarely “he did x because y” but rather “he did x because xyz to infinity”, we can start to see where almost every product or service can start to have a behavioral design element. They recently put up signs in some NYC subways that tell you how long it will be before the next train arrives. And it made people feel like the trains run faster, even when they don’t, because unknown time tends to just crawl along. Everything can be hacked on. Everything can be improved. The much harder question is figuring out what is worth spending time on.
Should more entrepreneurs study the science related to their startup idea?
Absolutely yes. But in that there is limited time and you can’t study everything, really learn and take to hear the lesson of leaning on experts. One of the things I say in talks all the time is go find a grad student to talk to — they know the science, they are passionate about the topic, and they’ll work for free lunch. And the graduation rate for PhD students is atrocious, so if you can get them interested in your problem, you may one day be able to draft them into larger roles. That’s how I became the head of product at Thrive — I was on the board of advisors merely to talk about the science, and then I realized I’d rather go actually do something about financial inequity rather than just study it.
Tell us about the tool you are launching soon that encourages men to be more aware of the gender balance of who they follow on Twitter.
Sure thing. I almost always have a side project going and recently, I got together with some troublemakers to build #IAmListening, a Twitter-oriented web app aimed at helping increase equity by encouraging people to think critically about the demographics of the voices they are listening to.
For me, it is really just another step in the advocacy journey. With Thrive, we only looked at finances after people got paid and since women are underpaid by an average of 20%, that was a huge miss. So we built GetRaised to help correct that imbalance. But then while we’re proud of the $2.3B it has helped women earn in raises so far, I realized that I didn’t like the way it put the focus on people who are already disadvantaged. I wanted to build something that put the onus for action squarely on those who benefit for the privilege. And that’s where #IAmListening came in.
What do you hope to achieve with this?
Ideally, I’d like people to think more directly about the echo chambers they live in and, in particular, how those echo chambers are partially defined by their privilege. People tend towards homophily: they like people who are like them. And so they naturally gravitate towards a homogenous in-group.
There are also social forces at work that perpetuate the in-group biases. For example, if most CEOs are white males, they are going to group with other CEOs and therefore get an overall white male perspective, even if it wasn’t the whiteness or the maleness that necessarily they fixated on. So if we can get them to widen that network, hear new voices, I think we can disrupt that.
How does who we follow alter our perceptions?
Who we follow defines what ends up in our Twitter feed and thus what we read, think, and talk about, and what we encourage others to read, think, and talk about through retweets, etc. Which I think is problematic at a pretty obvious level — if we’re only reading things from people like us, we’re not getting the full story.
But it goes much deeper than that. Our brains are lazy, so they often use heuristics, and while it serves us well most the time, it can be quite destructive. Take the availability heuristic: the easier it is to recall something, the more important we believe it to be, because our brains use ease of recall as a signal of importance. So if your Twitter feed says over and over that something is a problem, you’re going to think it is the most important problem, simply through a trick of your brain. And that’s incredibly dangerous, if your feed isn’t diverse enough to show you problems in a wider lens.
This gets even worse because of the confirmation bias. Basically, once we believe something, we start to differentially find evidence that it is true in our environment. If we believe that people with red hair have tempers, for example, we’ll start noticing every time a redhead yells, confirming our own bias. So again, in a homogenous feed, we get a bunch of information that makes it easy to recall through repetition, which causes us to believe it is important, which causes us to find evidence that confirms that belief.
That’s why representation is so, so key. Having a diverse feed allows us to combat our own tendency towards a lazy brain. And the great part is that it does it passively: once we setup a diverse feed, it automatically broadens what we’re exposed to. That’s why environmental hacks like this are so important and so effective.
How else can we encourage people to widen their lens online?
I actually think Twitter has done this in some ways, by decoupling us from regularly reading a publication. That is, instead of getting all my news from the New York Times, I now see a wider range of publications because they pop up in my Twitter or Facebook.
But those feeds depend on who I am friends with, so one key is getting people to allow curation from people who they don’t know. I actually think sites like Upworthy are useful in that context: it is certainly curation along a theme and likely leans somewhat liberal, but it still does the essential act of curating content across sources with expertise and integrity. I’d love to see more content online with the source and the curation decoupled, so that deciding what is is separate from deciding which things are important.
Do you have a particular interest in behavior by gender? If so, which are the most important issues?
I do. I sort of fell into feminism. I’m one of those men who for the longest time refused to call myself a feminist and insisted that humanism was a better term. It is only recently that a really wonderful, patient women who are friends sat me down and said “Matt, you built GetRaised, you’re a feminist and you need to start being able to say that.”
I think maybe why gender is so top of mind for me is that I can see so clearly my own mistakes. I say things all the time that I retrospectively realize could have been much better.
There are a whole group of feminists so angry at me that they actively tell other women not to engage with me. And that is really, really sad to me and just goes to show how far I have to go.
I’m a man. So perhaps naturally, I think the most important issue in feminism is helping men get involved in changing their own behaviors and the systems they benefit from. Which isn’t to say I don’t want to empower women and I reject the notion that encouraging women to lean in is telling them to be more like men — there is nothing in the world that says that ambition shall be the sole domain of men. We didn’t invent it, we don’t own it. So I want to encourage women to speak up and out.
At the same time, I think it is a special kind of ignorance to turn to people who are already at a disadvantage and say “you need to do more.” To me, it just makes sense to turn to those who are advantaged and ask them to step up, recognize where they aren’t doing the right thing, and do better.
Tell us about Bing in the Classroom.
Sure. When I came to Microsoft, I was really interested in how kids were using technology in classrooms. And one of the things that surprised me was how little search behavior there was. I can admit it: I got a little obsessed. Sat in a bunch of classrooms and just watched and learned. I have a degree in education in addition to my psych degree, so it was actually nice to be back.
Quick digression. When I build product, I use something called competing pressures design. Basically, you start from the end behavior you want to see (more successful searching in classrooms) and then you look at all the forces that make that more likely, which we call promoting pressures, and all the forces that make that less likely, which we call inhibiting pressures. And then we just try to maximize the promoting and minimize the inhibiting.
So after sitting in classrooms, there were a couple of things that stood out. First, it wasn’t generally a promoting pressure problem: kids already had a natural desire to search. They had plenty of questions, they were curious, and they had a general idea of how to use a search engine. Most of the problem seemed to be in the inhibiting pressure area, with three main standouts.
First, privacy. Teachers had a lot of concern about what was being done with students’ data; they had a vague notion that Google was using it somehow to make billions of dollars, but they weren’t really sure how and that lack of clarity was scary for them.
Second, advertising. Search engines are primarily monetized through ads and yet schools are ad-free zones. We don’t put up billboards in the hallways, so why would we let the computer become a place where that happens.
Third, adult content. It wasn’t primarily a problem of kids going into search and directly searching for inappropriate things. Instead, the concern was mainly around what we refer to as “leakage”, adult imagery and results that tend to pop up in some innocuous searches.
That was the big step for Bing in the Classroom: reducing those three big inhibiting pressures. So we made a version of Bing that schools could sign up for, completely free, that minimized the data we collected from students, eliminated advertising in the search engine itself, and forced on safe search features to minimize adult results.
And just doing those three things turned out to be huge. Over 10m kids are in schools that have signed up and we see very different behavior among those students, with over 40% increases in searches not only at school but also at home.
Then we added two other features to address two special cases, both primarily around equitable access to the tech. First, in lower income schools, you don’t tend to get the same level of native tech skills coming in to the classroom, nor do you see as much formal education around tech skills. So we created free daily search lesson plans that teach kids to shift their behavior from just copying homework questions into Bing to a more authentic, curiosity-driven style of search. The lessons ask questions, aligned with Bing’s daily homepage image, that a search can’t just easily answer and then teach students to use multiple queries to build up to the larger question.
Second, we tied Microsoft Rewards to Bing in the Classroom, so that parents at home could use the points they earned searching to earn tablets for schools. Again, in lower income schools, you simply don’t see many computers available and the ones you do see are often outdated. So we wanted to create a pathway where communities can help update their technology, simply by using Bing.
How can digital literacy improve opportunities for the next generation?Honestly, digital literacy isn’t voodoo. Just because it comes with shiny tablets and bright screens, it is just another skill that allows you to accomplish something that you want, like carpentry helps you build a house.
I think we need to stop giving digital literacy this magical techno aura and instead acknowledge it is a learned skill that provides access to meaningful work and meaningful outcomes.
And in doing that, we then make it more accessible to people who often see it as out of reach. If anyone can learn to be carpenter, and carpentry is like digital literacy, then anyone can learn to be digitally literate.
Let me actually say it a different way. My mom is a nurse and for most of my life, she was a shift nurse at a hospital. Every time I tried to teach her how to use the computer as a teen, I got frustrated and she ended up crying. And now, at age 60, she’s in nursing informatics. She coordinates the rollout of huge tech systems at the hospital. And she does it because she finds nursing to be meaningful work, say how tech could help make nursing better, and she figure out how to make it work for her. My mother is awesome but she isn’t a tech savant; if she can do it, we can get everyone else there too.
How can we help people find meaningful work?
I think the first step has to be reorienting them to look for it. Right now, the dominant notion of work is trading your time and effort for money, which is a very economics driven viewpoint of the entire labor system. But just as behavioral economics has sprung up out of a desire to take psychology into account when considering economic theories, I think we need to put the basic psychology back into the labor system.
As people make more money and more of their basic needs are met, they become less and less motivated by money and more and more motivated by meaning. If you ask college seniors what matters in a job, their number one answer is salary. Why? Because we teach them, day in and day out, that the reason to go to college is to make more money. We run advertisement after advertisement that reinforces that.
Yet after just two years in the workplace, those same seniors increasingly say that meaning is a dominant factor in choosing work. But now they are out of school, outside most of our ability to influence them, and they don’t even know what that means or how to go about finding it. So I think helping people find meaningful work is really about introducing the concepts early on and then being explicit about how to look for work that you love.
And that has to start at home and at school. Too often we try to shield kids from work because it isn’t fun. Parents don’t talk to kids about their jobs and why they do them. Teachers don’t encourage kids to see their efforts as work and if they do, it is all about trading effort for grades, instead of effort for meaning. It is a radical shift, to reorient to meaning, but also a simple one. Sit down and talk to your kids about what you do and why you do it. Tell them about work you’ve loved and work you’ve hated. Be honest. Be passionate.
You mention the M&Ms experiment in your TechStars talk, why was this one of your favourite experiments? Which other experiments have been important to you and why?
Ha. That’s really an unfair question, there have been so many that I often go on and on during talks about a litany of experiments I love. But there is probably one that sticks out, which I always think of as The Good Samaritan study. The setup is pretty straightforward: divinity students go to the psych building to fill out a battery of personality surveys and then they walk across campus to deliver a sermon on the parable of the Good Samaritan. And in between the two buildings, they find someone in need of help. The thing we’re look at is whether they stop to help.
But there is a wrinkle. Half the students are lead to believe that after filling out the surveys, they have plenty of time to make it over to give the talk. The other half are told they are running late and must hurry.
You can see where this is going. The half with plenty of time tend to stop and help, the half who are in a hurry tend not to. But think about that. Put yourself on a park bench, watching to see who stops. We are so judgmental: the ones who stop are good, the ones who don’t are bad. We think it is about their personality, some deep inner characteristic of who they are. But it isn’t: it is just being busy or not.
How many millions of times do we look at people and say they are bad or good at some deep core level, without appreciating the circumstances in which they are operating? When in reality, we are all the Good Samaritan, and whether that comes out depends mostly on what else is going on for us.
Thinking about that study, I try to be more patient. To try to step back and look beyond the person and see what else is going on. I’m not always successful but I’m trying.
What would be your biggest learning about people’s behavior online?
That it is very much like people’s offline behavior. As I said earlier, I try not to fetishize tech. Yes, people do act differently online but generally not because of the virtualness but because one of the things that happens more often in virtualness, like anonymity. Yes, people can be hateful assholes online when they can hide behind a screen. But it turns out that people are also hateful assholes offline when they are given a mask and the opportunity to enact aggression.
I’m sure some much better academic out there will tell me ways I’m wrong about this and I’m sure there are special circumstances. I just don’t think that online is nearly as special as we make it out to be. Those are still people, their brains still operate with the same set of guidelines, they are still doing what they do.
How do we motivate people to take action?
I don’t. Or at least, I try not to. Most of the projects I work on are about reducing inhibiting pressures: the barriers that keep us from doing things we already want to do. And there are a couple of reasons for that.
First, it is incredibly hard to get people to want to do something they don’t already want to do. You can guide them to do specific things, like watching Star Wars instead of The Real Housewives of wherever, but generally that is just acting on preexisting motivations towards entertainment, affinity for scifi, etc.
Second, inhibiting pressures tend to be homogenous, while motivations tend to be heterogeneous. There are lots and lots of reasons people want to see Star Wars, but only a few reasons they don’t: time, money, availability. So you get much more upside from working on inhibiting pressures.
Finally, inhibiting pressures aren’t generally in competition. When you try to get people to want to do something, you have to fight with all the other things they want to do. There is a limited pool of cognitive attention, and when you ask for a piece, there is an implicit tradeoff with everything else they could be paying attention to. Making people want to watch Star Wars means beating out their desire to see every other movie. Whereas concentrating on making it easier to see Star Wars means that you just ask for a smaller piece of the pie — everybody wins.
Which companies are good examples of how you can embrace and harness behavioral science?
Alright, I’m going to use my canonical example here, but I want to give a big disclaimer first. Uber is killing it, behaviorally speaking. But I don’t love their business model, founder, etc. So recognize that good examples don’t always mean ethnically great companies.
So why do I love Uber? First, I think they did a really good job of focusing on inhibiting pressures at a time when others in their space were focusing on promoting pressures. Everybody else was trying to upgrade the black car experience through promoting pressures: better cars, in-car wifi, distinctive experiences.
But Uber recognized people already have sufficient motivation: they want to go from Point A to Point B. So it was just about minimizing the inhibiting pressures to getting that done. They made it easier to pay (don’t underestimate this; if you’ve ever been in NYC on a one way street, feebly trying to pay for a cab with a credit card while a block of traffic lines up behind you and lays on the horns, you know just how important simply hopping out can be). They made it easier to book. They made it easier to find your car and to reduce trip times. They made it cheaper.
And that relentless focus on inhibiting pressures has carried over into their marketing. If you read Uber emails, every once in a while they do a promoting pressure event: you get a puppy in your car or get picked up in a Tesla or whatever. But the vast majority are about three things. One, it is now cheaper than it was before. Two, there are now more Uber drivers on the road so you’re more likely to get one quickly. Three, we can go somewhere we haven’t been able to go before.
That’s it. Price, availability, removal of restrictions. Welcome to a billion dollar company.
I also think they did a good job of recognizing that experience is about emotion as much as practicality. It gets really easy to get over focused on the hyperlogical, practical aspects of our products. Yet as the Uber founder frequently points out, he just wanted to make a way for people to feel badass. And even though it has become ubiquitous for many people, there is something sort of awesome about just hitting a button on a phone and having a car and a driver magically appear. If you look at the deals and integrations they are doing, it is based on inhibiting pressures and that magical feeling. Like with Microsoft; you can now schedule a meeting, check a box that says you’ll need an Uber, and it will just magically show up to get you in the right amount of time to get you there. That’s sort of amazing when you think about it.
And the big mistakes you have seen in tech? How could these have been avoided?
I mean, tech is ruled by startups and the landscape is littered with dead ones and ones that didn’t reach their full potential. Look at Gilt. They got a very basic psychological principle, that adding urgency and then giving a clear signal (this is the cheapest you’ll ever get this) is a brain short circuit for spending. But when I went and gave a lunchtime talk there years and years ago, it was so clear that they got there by accident. They didn’t really understand people and the rules they were playing by, so when it came time to launch other products, they couldn’t get there.
Who knows, that could still happen to Uber. It isn’t enough to just blindly stumble on one innovation: you need to have a theory of humanity, and you need to work that theory to continually produce new products based on it.
What are the success factors for creating a successful startup?
There are a million possible answers to this and others have written more and better than I possibly could. People is certainly key. Having a viewpoint is key. Maybe I’ll go with this one: you have to fall in love with your problem, not your solution. Because we are constantly finding better solutions and if you stick to yours, you can get disrupted easily. But if you’re in love with the problem you’re trying to solve, as long as it is a worthwhile problem that someone will pay to have solve, you can iterate endlessly on better solutions to continue to expand and improve.
What are the most interesting future trends?
Augmented reality is certainly a candidate but it isn’t the one I’m most interested in. For me, it is still about access. There are huge parts of the world that still don’t have indoor plumping, power, efficient agriculture, etc. The future trend we ought to care about is solving those, because it will make AR look pale in comparison. You think the internet is amazing for making innovation happen? Try lighting up the world with power and then see what happens.
What would you recommend for others to read?
I keep a list of books on my website but honestly, I’m a bad source for this because I only read fiction (and scholarly psych articles). Here’s why: no matter what the latest book is, someone will paraphrase it for me. Or someone will do a TED talk. Or I know an expert who will tell me about it over dinner.
But you can’t do that with fiction. Art is that which is intrinsically reduced when summarized. So that’s where I spend my reading time — other people teach me the rest.
You have a new baby (congrats!) — how do you hope tech will change the world for the next generation?
I hope tech continues to solve the basic sustenance problems, so that my son’s generation can build on top of a solid base. Selfishly, I hope we unlock exploration again, which probably means space or the sea most immediately. I’d love for him to be able to go boldly into the unknown…but only if he wants to. He can also be a nurse. That’s cool too.
How can people find out more about what you do and the tools you are launching?
I honestly have no idea. I’d normally say my website but at this point, I’ve become horribly bad at updating it. The hope is that you won’t ever have to fixate on Matt Wallaert the person, that you’ll find out what I’m doing because you’ll be affected by it, because it will enter your daily life.
I’d actually love it if you never knew my name. I’m fascinated by the everyday inventions around us and how great it would be if you could go your whole life never knowing who I am and yet being enriched by something I built. I love products that build themselves out of existence. What if Bing in the Classroom didn’t have to exist because all search was safe and private and ad-free? What if we didn’t need GetRaised because everyone was paid fairly? I’m really just trying to work myself out of a job so I can go find something else to solve.