(This is a post primarily about the results of an experiment, but I did build a tool as part of it that you can use to gather behavioral feedback from former coworkers; you can go directly to WorkWithMeAgain.com to use it for free without reading about the journey or my data.)

It started, as so many of my projects do, with a conversation on Twitter. I proposed a public Glassdoor-like system with a single rating: would you be willing to work with this person again? Slice that by demographics and we might have a tool worth building to increase inclusion in the workplace.

After much healthy debate about the risks (it could be used as a harassment tool, for example), much of the value seemed to stem from the collection of feedback itself. Shifting away from the goal of a public system for discovering bad actors, it is worth asking: does the average person even know whether former coworkers would want to work with them again?  

I certainly don’t and especially not in any systematic way; despite being much-touted by thought leaders in the management space, organized feedback collection is relatively rare.  For years, I’ve had a link at the bottom of my email that allowed people to leave me anonymous feedback via Google Form and I’ve received plenty of responses, some helpful and some less so (apparently, someone doesn’t like my cowboy boots).  But I never collected demographic data or any concrete behavioral metrics, like willingness to work together again.

So I started building.  I iterated through several versions of both the survey questions and the analysis before arriving at the results I’ll talk about today.  The dominant behavioral measure is a simple 1-5 scale: “Given the opportunity, would you actively avoid or seek out working with me again?” With that, I asked gender and ethnic identity, age when we worked together, reporting relationship, and recency of the relationship.  

To avoid as much bias as possible, the final survey was sent out indiscriminately to ~800 people I’ve worked with over my ~15 year career.  The sampling wasn’t perfect, because people have moved between companies, I’m not connected to everyone I ever worked with, and many people who have worked with me don’t have LinkedIn profiles (due to socioeconomic skews of who is on LinkedIn).  But this whole project is a lesson in not letting the perfect be the enemy of the good; it is easy to find reasons that feedback will be incomplete, but that isn’t a reason to not get feedback at all.

In all, I received ~70 completed surveys, which seems extraordinarily high given what we know about average survey response rates.  Since I phrased the survey request as an experiment, the results of which I was committed to publishing, there may have been some increased promoting pressure but I’d like to believe that we might actually be on to something here; people want to give feedback, if we choose to hear it.

Before we get into the result that drove the title of this post, let’s start with resolving one important concern that came up in the Twitter discussion: that many people would choose not to add demographics for fear of being identified.  This didn’t appear to be the case, as only ~10% chose not to volunteer full demographics.  Assuming that response rates were high enough, it is thus likely that meaningful demographic filters can be applied.

Now the results at a high level, with the boring stuff first.  If you want the nitty gritty, I’ve linked to a copy of the report so you can see what you’ll actually get if you do this yourself using WorkWithMeAgain.com, although I have redacted the free entry portion to preserve the anonymity of respondents.  I also hadn’t yet added the recency question (which was the great suggestion of Leigh Honeywell) in the version of the survey I sent out, so there is no data for that section.

The basic finding is that statistically speaking, there were no significant differences in the averages within demographic groups.  Men/women/non-binary people and white/non-white people were equally likely to want to work with me.  And the averages were all above three on the scale, meaning generally all of those groups would seek out working with me again.

Does that mean I’m free of bias?  Of course not; as a white male, I absolutely have racist and sexist associations that express themselves as behaviors, despite my intentions.  What it does mean is that at least in this sample, these biases were not large enough to meaningfully affect any particular demographic subsegment, which is a good thing.

There are some interesting correlations that are neither good nor bad.  For example, the older someone was when they worked with me, the more likely they are to want to work with me again, which means I may need to change how I relate to younger people in the workplace.  I can come across as condescending, so this is a known problem that I actively work on, and that data backs it up.

Where the data starts to get interesting is when we look at standard deviations.  So rather than how much would people seek me out generally, how much disagreement was there within the group about that behavior.

I am, unsurprisingly, a fairly polarizing coworker.  I’m outspoken on social issues, have a distinctive personality and work style, and frequently clash with others who I perceive as acting against the group interest.  I’m not unaware of this, although I do vary at times in how much I see it as a strength, weakness, or simple fact.

This polarization, however, is not evenly distributed.  Even though men and women/non-binary people don’t differ in their average desire to work together again, they do differ in their polarization: women have much stronger feelings about avoiding or seeking me out, as do people I managed and people who worked with me across teams.  In my case, because I have enough data to cross gender and ethnicity, the finding is really around white women: men and non-white women consistently would seek out working with me, while white women are basically either 1s or 5s.

This is in someways rather puzzling. For example, I once lead a customer service team of ~100 people who were almost all non-white women and subsequently outsourced the function, resulting in a significant layoff. While I remain convinced that it was the right decision that was made in the best interest of our customers, all of those people would be entirely justified in avoiding working with me again. So I would have expected a potential main effect of ethnic identity or at least an interaction with gender.

I’ll confess, I don’t know what the varying feedback from white women means yet.  And that’s a fine thing: survey data like this rarely gives us a conclusive answer.  Instead, good quantitative analysis clues us in to where to do qualitative analysis; data tells us what, not why.  For me, the survey is a jumping off point to conversation that I can use to improve and I intend to seek out further conversation, particularly with white women who I’ve worked with before.

Feedback isn’t a silver bullet for changing our behavior.  But it can be a start.  Personally, I’ve committed to publishing this data and to following up with the 800 people in the initial sample, as well as periodically adding new co-workers with each new job.  As part of that, I’ll also resurface the link to the anonymous feedback form I already have setup; for any of the 10 people who would prefer not to work with me again, if you are willing, I’d deeply appreciate you using that to give me feedback on what I could have done differently to change your experience with me.  Finally, readers of the writeup can use the same form to give suggestions on actions I should take or email me at matt@mattwallaert.com if you don’t need anonymity.  I want to do better, I’m committed to doing better, and I hope when I one day rerun this experiment, the results will be different.

If you’re committed to doing the same kind of work, the tools I used are now publicly available at WorkWithMeAgain.com.  It will allow you to copy the Google Sheet that contains all the calculations, as well as instructions on how to launch the survey to your former coworkers.  I am absolutely convinced that putting in the effort to gather meaningful behavioral feedback can be a key component in how we change our individual behaviors, as the collective culture is simply made up of those individual behaviors; if enough of us do this, we can live in a substantially better world.

Side Note: As usual, this project required collaborators: frontend coding from Grayson Null and design by Patricio Hunterkhozner. Plus feedback from a WhatsApp group full of folks working toward inclusion and 70 people who were willing to share their thoughts about working with me again. It has been a good year for WorkWorthDoing projects (Mediocre White Men and Project 4255, with another on the way soon) but they don’t happen without a bunch of people willing to work at reduced rates and in suboptimal circumstances.

Posting my survey results, online and unfiltered, is making me nervous. And that’s a novel feeling, since I usually don’t really feel nervous in circumstances that don’t involve heights (I really don’t like heights). That makes this a rare brave moment, because if I don’t normally feel the feelings, I’m not actually being brave: it is easy to do something with low inhibiting pressure. There is no concrete outcome I am afraid of…but I’m certainly afraid.

In a tweet that spawned a million tote bags, Sarah Hagi said “God give me the confidence of a mediocre white dude”.  And I love the meme in that special way I love anything that is both funny and scientifically valid.  A plethora of studies about the confidence gap between white men and both women and people of color, coupled with the brain’s tendency to inappropriately perceive confidence as a proxy for competence, explains a world full of mediocre white dudes with disproportionate power.  The term “failing upwards” comes to mind.

But why are white men so overconfident in the workplace? 

In Start At The End, my book on how to create behavior change, I talk about behaviors as the result of a competition between two sets of pressures: promoting pressures (reasons to do something) and inhibiting pressures (reasons not to do something).  

One of the two must be at play for these white men.  Either there is a strong promoting pressure (like believing they are simply incredibly competent) or a weak inhibiting pressure (like believing that failure isn’t such a big deal).  And because nobody seems to know exactly which one it is, I grabbed former colleague and equity-minded data scientist white dude Tyler Burleigh and WhyMenAttend.com co-author Rhapsodi Douglas and we went off to gather data.

The survey design was fairly simple.  We asked two sets of questions to 500 people over the age of 18.  All respondents were in the United States and as demographically representative as possible.

The first set of questions assessed occupational self-efficacy, a construct that is essentially a measure of how competent we think we are in the workplace.  Typical items are things like “No matter what comes my way at work, I’m usually able to handle it.” and “When I am confronted with a problem at work, I can usually find several solutions.”  This tested the strong promoting pressure explanation: white men are overconfident because they believe in their own absolute competence.

The second set of questions assessed psychological safety: the belief that a work culture is gentle, human, and forgiving.  It is measured by items like “People at work are able to bring up problems and tough issues.” and “It is safe to take a risk at work.”  This tested the weak inhibiting pressure explanation: white men are overconfident because there is really no reason not to be, since work is a place where it is acceptable to fail.

Two sets of explanations, 144 white men.  Next up: compare their answers to those of our 356 women and people of color.

Before we get to the big reveal, there are a few trends that are worth noting. First, the average person feels relatively competent at work, scoring 5.44 on a 7-point scale.  This isn’t all that surprising, given the very strong motivation to both take work that you can actually do and to believe that you can do the work you have. In contrast, people generally don’t feel as psychologically safe at work: the average score was 4.32, so higher than the midpoint but not as high as self-efficacy.

Second, having high workplace self-efficacy and feeling psychologically safe are moderately correlated, around r = 0.38.  For comparison, that’s about the same correlation as thinking something is a good idea and actually doing it across a variety of domains (which explains why we all know we should go to the gym but tend not to go).

And now, the secret to the peculiar psychology of #MediocreWhiteMen: confidence or psychological safety?

The simple answer is both.  White men had higher self-efficacy than women/people of color (5.65 versus 5.36, p = 0.02) and felt greater psychological safety (4.43 versus 4.27, p = 0.04).  Note that these aren’t huge differences in absolute numbers but on a 7-point scale, it is a combined ~10% difference. And since confidence is seen as a proxy for competence, which translates to compensation, suddenly 10% starts looking quite large.

So now we’ve got data: both promoting and inhibiting pressures are acting in favor of white men.  We’ve quantified white male privilege. But there is reason to believe that one is actually much more important than the other if we want to disrupt that privilege and create a more equitable workplace.

Underrepresented people consistently underestimate their competency in almost every domain you can measure, and the workplace is no exception.  And thus interventions that help foster self-efficacy will likely be effective at increasing promoting pressures for risk taking and confidence in the workplace and may lower the compensation and promotion gaps (so long as we manage to get underrepresented people credit for the work they do, which is its own struggle).

But it is increasing psychological safety that potentially holds greater promise.  Remember, for women and people of color, the mean for workplace self-efficacy is already 5.5 on a 7-point scale; there is only room for a ~20% improvement.  The mean for psychological safety, by contrast, is 4.2, implying room for a ~40% improvement. We need to make workplaces feel safer for women and people of color by constantly reinforcing collaboration over competition, finding both personal and professional common ground, and moving from the Golden Rule (treat others as you would want to be treated) to the Platinum Rule (treat others as they want to be treated).

We also need to acknowledge that psychological safety may be a misnomer.  We measured perception, not reality, but given that there is abundant research that suggests that workplaces actually are less safe for women and people of color, it may very well be that these groups are simply reflecting an accurate understanding of their environment.  Underrepresented people are more likely to be judged harshly for the same failures, more likely to be sabotaged, etc. And white men are more often judged on their potential than their actual demonstrated experience.  

The overconfidence of #MediocreWhiteMen isn’t irrational, but rather the product of an environment that has been designed to reward them.  And it is only by disrupting the design of that environment that we can create change.

Side Note: For a rabbit hole on psychological safety, check out Amy Edmondson’s work on the topic, particularly this review piece.  Google has been very strident about their belief that it is the defining characteristic of high performing teams, so much so that they’ve implemented manager training on it across the country, using materials like this.  Personally, I feel very safe working with Tyler and Rhapsodi, which may be why I keep making so many mistakes in front of them.  For an excellent deeper dive into the statistics behind this article (including sample descriptives and other geekery), Tyler put together a longer stats-focused post.

Also, we decided to honor the original Tweet with the hashtag in the title but I did crowdsource alternatives and feel some deserve to be included: #historyoftheworld, #BornOn2ndBase, #himpocracy, and my personal favorite, #himposter.  If someone wants to make me a t-shirt that says “Don’t be a #himposter”, I will gladly wear it to GHC.

Recently, Maia Bittner tweeted about the top three things people want to know about you, using Google’s search suggestions. This inspired me to check what came up for “Matt Wallaert” and then be subsequently horrified that my marital status is apparently more important than my work on behavioral science. Bing, on the other hand, doesn’t care who the hell I married so I’m doing those as well. If I remember, I’ll update this periodically.

matt wallaert seattle
Yes, I lived there while working for Microsoft and am a native of Oregon (go PNW!). Why is this the most searched thing?

matt wallaert getraised
Yes, I (and others) built a tool that has helped women earn over $3.2B in raises called GetRaised.

matt wallaert book launch
Yes, I wrote a book called Start At The End: How To Build Products That Create Change, which attempts to distill my experiences leading Product and Behavioral Science at companies both large and small into a replicable process for behavior change. Blame Merry Sun, who bribed me with Silvana. My son Bear was a prominent feature at the book launch – even my parents flew in!

matt wallaert clover health
Yes, I currently work as the Chief Behavioral Officer at Clover Health. I joined in 2017 and wrote about why on the Clover Health blog.

matt wallaert a behavioral psychologist
Yes, I am an applied behavioral psychologist and have written extensively about what it means to be a Chief Behavioral Officer working in industry.

matt wallaert videos
Yes, there are many videos available of me on YouTube.

Recently, Maia Bittner tweeted about the top three things people want to know about you, using Google’s search suggestions. This inspired me to check what came up for “Matt Wallaert” and then be subsequently horrified that my marital status is apparently more important than my work on behavioral science. After some contemplation, however, I decided the right approach was to make a page that tries to address the popular searches in a straightforward manner. If I remember, I’ll update this periodically.

matt wallaert wife
Yes, I was married to the awesome Dr. Sugar. We are no longer married, although I don’t refer to her as my “ex-wife” (although technically true) because that emphasizes our past relationship; instead, we use the term “co-parent”, which focuses on our current relationship as we work hard to raise Bear Sugar to be as badass as his name.

matt wallaert clover health
Yes, I currently work as the Chief Behavioral Officer at Clover Health. I joined in 2017 and wrote about why on the Clover Health blog.

matt wallaert book
Yes, I wrote a book called Start At The End: How To Build Products That Create Change, which attempts to distill my experiences leading Product and Behavioral Science at companies both large and small into a replicable process for behavior change. Blame Merry Sun, who bribed me with Silvana.

matt wallaert ted talk
This one is hard. I think people mean my TEDx talk, which was done as a favor to Arjan Haring in his final year of organizing. Maybe TED will one day invite me to give a mainstage talk; you can suggest that here. For what it is worth, this isn’t my favorite talk (since I really, really dislike slides) – I prefer another Dutch talk I did for Nibud, in which I sound distinctly like someone with an electrolarynx.

matt wallaert start at the end
That is the name of the book I wrote.

matt wallaert twitter
Yes, I am on Twitter.

matt wallaert net worth
I don’t know exactly but let’s call it high single-digit millions? I put the maximum into my 401K pretty much every year of my employed life, got lucky with Microsoft stock and the house in Seattle, and generally live a fairly fiscally conservative lifestyle.

matt wallaert linkedin
Yes, I have a LinkedIn.

matt wallaert wikipedia
There is not a Matt Wallaert wikipedia entry (you can make one here), although I do get mentioned in the entry for Thrive.

A few weeks ago, I got into a discussion on Twitter about why more men did not attend gender-focused events. In the world we want to live in, men recognize that they benefit from privilege and actively address it. In the world we do live in, change has been slow and male involvement low, which leaves many women taking on the double burden of both sexism and the emotional labor of ending it.

One way to lessen that is to better understand why men become active feminists so that we can hasten the shifting of the work. But unfortunately, we don’t actually know that much about why men become active feminists. Studies have typically looked at fairly specific phenomena, like venture capitalists with daughters being having better performing funds, and even those are fairly rare on the ground.

So I decided put my money where my mouth was and fund a little research. Using Survata, I paid for two, 200 person surveys to be run. Both groups were all male, all in the United States, and all over age 18. One of the surveys was for men who said they had attended an event focused on gender (examples were “Celebration of Women in INDUSTRY” and “Gender Equity in INDUSTRY”), the other was for men who said they had not. Both groups were offered a variety of reasons for their attendance or non-attendance and were allowed to select as many as they wished and provide other factors, as well as filling out a variety of demographic questions.

I then called my friend and fellow Harlem dweller Rhapsodi Douglas, a consultant in Deloitte’s Diversity and Inclusion practice, and we got together to do some data analysis and talk through the findings. I’ll be switching to the plural now as she comes into the picture.

At the highest level, the most popular reason on both sides was simply about the importance of gender-focused events: 55% of non-attendees said they were not interested in gender-focused events, while 41% of attendees said that focusing on gender is important to supporting women. We should take that finding with a grain of salt, however, because of cognitive dissonance. Because our beliefs change to line up with our actions, non-attendees may simply say it isn’t important because they didn’t go (rather than not going because it isn’t important) and vice versa for attendees. So importance is…well, important! But it isn’t the end of the story.

Many of the other general findings are fairly obvious. Younger men are more likely to have attended a gender-focused event and more likely to acknowledge the existence of sexism. Having at least one daughter was associated both with acknowledgement of sexism and attendance, as was being employed.

But we can’t change age or employment or having daughters. So let’s look at the reasons for attendance or non-attendance that we do have more control over. And let’s start with some complexity. There was a significant difference in the number of reasons that attendees and non-attendees selected to explain their behavior: 40% of attendees selected more than one reason, while only 15% of non-attendees did.

Digging deeper, men who didn’t attend generally fell into one of two fairly distinct groups: those who would go in the right circumstances (36%) and those who wouldn’t (64%). Those are roughly equal to proportions found in Matt’s previous work on acknowledgement of sexism in the workplace: around 2 in 5 men acknowledged at least general sexism, while around 3 in 5 men denied it.

Men who did attend cited a much broader range of reasons, usually including either that they believed attending would personally benefit them or was important to supporting women (what Matt would call promoting pressures) plus at least one form of acknowledgement that men were welcomed (what Matt would call the removal of inhibiting pressure) like being specifically invited, having male speakers, or a session description that specifically mentioned men.

Given that 36% of non-attendees cited inhibiting, rather than promoting, pressures as the reason for not attending, one potential interpretation of these results is that we could substantially increase male attendance at gender-focused events by creating the right circumstances.

So how do we remove the inhibiting pressures? Well, there are two things that about 20% of attendees cited as being important: having at least one male speaker and having the session description mention being open to men. Interestingly, very few non-attendees cited the lack male speakers or direct fears like being disruptive to women or saying/doing the wrong thing. But some noted that it conflicted with other events they wanted to attend, so there is a third potential recommendation: make sure gender-focused events are uncontested on the main stage.

All of those interventions, however, are in the control of conference organizers. Fortunately, the most powerful intervention, cited by large number of both attendees and non-attendees, is something each and every one of us has the power to do. It costs nothing and means everything: the simple act of invitation.
40% of attendees cited an explicit invitation from a man or woman, while 16% of non-attendees specifically said they didn’t feel invited. In a survey of this nature, those are large numbers, especially for such a seemingly trivial intervention.

Maybe a tool is needed, a simple one page site like SalaryOrEquity.com that makes it easier to invite a man to a gender-focused event. Or perhaps an email template is enough. Maybe just this pure look at the data will do it. There is certainly more work to be done on what gets people to take that first step and start inviting.
What we do know is that helping men show up and be affected by the content of gender-focused events is critical to shifting the work of dismantling sexism to men. And that introductions are a powerful part of that. Matt would never have attended his first Grace Hopper Celebration of Women in Computing if it weren’t for Betsy Aoki telling him he should go. Rhapsodi clearly felt the turning point with a man in her own life when she started involving him more in her feminism. Personal experience and the data seem to converge: now is the time for invitation.

Side Note: These are just surveys. What we really need are experiments. We need to send half of the men attending a conference a personal invitation to attend the gender-focused session and see if they are more likely to attend than the half that don’t get an invite. We need to know the male attendance of talks with and without a man on stage, with and without men in the session description. And that all starts with tracking the gender of attendees. Without those gender attendance numbers, we cannot know how well we are doing and how far we have left to go.

I have discovered, over the years, that I’m a spectrum thinker. On white boards and bar tables and with wild air gestures, I always seem to be explaining how there are two opposing endpoints and why I’m only interested in this or that part of the area between them.

So it is perhaps unsurprisingly that over the past few months of evangelizing the idea of a Chief Behavioral Officer and talking to companies both large and small about how psychology fits in their business, I’ve started to see a spectrum in how application is happening.

On one end is Insight. The function typically lives in data science/analytics/whatever the heck we are calling it these days and reports to the head of that area. The primary inputs tend to be variables that are already instrumented, and the primary output is typically some sort of report that indicates a potential surface area for change, with some more ambitious companies also including a few recommendations for high-level potential intervention. This report is handed off to whoever controls the variable itself (marketing, product, ops, etc.), while the Behavioral Scientist returns to the data puddle to investigate something new.

This is where the bulk of the job openings are at the moment: Allstate, Amazon, Facebook, you name it. In some ways, the descriptions often sound like a hybrid of user research and data science, with the goal expressed as “We want to understand our users’ behavior, particularly where it is irrational”. Understand is the key word; this role is about the why of human behavior. Certainly there is an implicit belief that the understanding will lead to better behavior change, but the actual change lives elsewhere, with whoever owns the lever that may need pulling.

Contrast that with the other end of the spectrum, Intervention. This function is focused on the actual changing of human behavior and seems to be living in Strategy/Innovation/Global Services. While this role may touch data, it doesn’t seem to have analysis at the core of its function (think SPSS instead of R) and if paired with a solid data team, may not actually be doing much data work at all. The output isn’t a report but rather an intervention that has been experimented and iterated until it can be shown to reliably change a behavior and is ready for scale.

Similar to Insights, there is still a handoff at the scaling point, where the intervention is handed off to the relevant team for ongoing ownership, but relatively speaking, the Intervention function is picking up the ball later (after an insight) and carrying it farther (a scaleable intervention exists).

There have been comparatively fewer roles I’ve seen here, in part because Insight already fits into existing structures (Data Science reports on trend, someone else pulls the lever), whereas Intervention requires creating a new step in between. But I believe that this is a little like the recent pseudo-bifurcation of data science as analytics (BI, Insight teams, etc.) and data science as product (machine learning, AI, etc.). A conversation with an insurance company recruiter sticks in my mind: “We’d love to be doing intervention, we just don’t think we are there yet, so we’re starting with insights.”

Is this spectrum rigorous? Absolutely not, and every company is thinking about it differently. There is no science here, only an attempt to pattern match the signal out of the noise. But I think that in order for behavioral science to catch up to data science in terms of corporate understanding, it behooves us to start understanding how to use a common vernacular. Executives need terms they can buy in to and recruiters need roles they can recruit for.

One potential option is to recognize the commonality of the two roles by keeping a single title, Behavioral Scientist, but emphasizing differing job requirements and responsibilities. Speaking very broadly, I’ve seen more postings use “behavioral economics” as a requirement when they are looking for Insight and “behavioral design” when looking for Intervention, although both of those terms are about the modification of other fields to incorporate psychology rather putting psychology at the center.

In my conception of CBO, both insight and intervention are needed. I’ll admit that I’m biased toward the intervention side, since my expertise is mostly in the building of things, but look at Bing in the Classroom: there were initial insights (“School search volume is lower than expected”, “Curiosity is not the root cause”) that allowed for the intervention. Ditto GetRaised (“Women are significantly underpaid”, “Women are less likely to ask for raises and less likely to get them when they do”).

But as with data science, we must resist the urge to simply relegate behavioral scientists to insight functions. There is a natural tendency to look at the black box of human behavior and long for understanding.  But in reality, business is driven by the ability to change behavior, so to not apply science directly to the intervention design seems foolhardy.  Regardless of which is more needed, however, the predicting of behavior and the modification of behavior are related but not the same, and should not be painted with a single brush.

Side note: For years, I resisted calling myself a feminist. Typical arguments about humanism and striving for equality not being gendered and blah blah blah. And now it is in my damn Twitter bio. Similarly, for years I’ve resisted the term behavioral design. Science is so important to me, it is hard to leave it out. And yet as people increasingly use behavioral design to differentiate from behavioral economics, it may be something to consider. I’m not convinced enough to yet start using the term in a self-applied way…but I’m tempted. Particularly because I distinctly don’t want to spend the rest of my life predicting behavior; I want to create it.

Earlier this week, I bombed a talk at the Professional Convention Management Association’s Education Conference.  This is actually fairly rare for me: because I love the science, it is normally easy for me to talk fluently and authentically.  This week, though, I just couldn’t get it together.  So I’m going to do something I try not to – write what I should have said.

Behavior Change
Before I melted down, I did a pretty decent job of explaining at least the basics of competing pressures.  But I missed a few key points that are worth surfacing.

First, because of the natural tendency to focus on promoting pressures, there is a great deal of whitespace on the inhibiting pressure side.  But it is not just because of our focus that this remains true.  In my M&M example, you’ll notice that the promoting pressures tended to be heterogeneous: one person wants to eat M&Ms because their blood sugar is low, another because they are delicious, another because they’re a bit sad and need a delightful moment.  But the inhibiting pressures tend to be homogeneous: we are all affected by cost, availability, etc.  Thus, while strengthening a promoting pressure may help a select few, weakening an inhibiting pressures tends to help everyone.  Thus, on a dollar-for-dollar basis, the money you spend making things easier to do will have higher ROI than those you spend on making people want to do things in the first place.

Another benefit of focusing on inhibiting pressures is that it helps move away from a nuclear arms race for attention.  To a behavioral scientist, the limited resource in the world isn’t time or money but mental energy.  So when we go the traditional route of trying to maximize share of mind (“If you only attend one conference this year, it should be X”), we force all conferences to compete with each other for a fixed and limited resource.  On the other hand, if instead what we do is make things easier and more mentally efficient, we actually unlock greater potential.  If I can provide the same value, but at half the mental, then that creates new focus that can be used to attend another conference or take another action at the same conference.

In many ways, this is really a plea for focus.  Conferences have become choice overloaded, a sort of fear-of-missing-out hell where everyone feels like everyone else is getting more value than they are.  I’m not saying any of the following products should exist, but imagine if they did.

What if, instead of making me try to find friends and make dinner plans, your conference app automatically set me up for dinner with people you thought I might enjoy?  This might sound a little crazy but let me tell you about an experiment I did once.  We built an app that advertised itself as doing one thing: sucking in all your personal data, then recommending the absolute best place for you to have lunch.  You logged in with Facebook and got a recommendation, but on the backend, we didn’t actually personalize at all.  We simply pulled randomly from a list of restaurants with good ratings that were nearby.

People loved it.  They said it knew them so well, marveled at how much it could tell just from their Facebook data (this was a few years ago, so people might not be surprised now), and how much better it made choosing food.  We can so often get obsessed with the idea that if we can’t do things perfectly (deliver on those promoting pressures), they aren’t worth doing.  But sometimes, just holding value constant and making things easier is worth it.  Everyone has to eat.  Don’t make them choose where and who with unless they want to.

Another example is around increasing representativeness of speakers.  So often, I hear from organizers who say “Well, we just don’t get that many proposals from women” or “There just aren’t that many underrepresented folks that want to talk”.  Bullshit.  There is a huge difference between not wanting to talk and not feeling like you are the right person to do so.  What if you ask your existing speakers (particularly if you’re paying them) to run a quick speaker training for attendees who might want to try speaking next year?  Or create speaking slots with lower barriers to entry, like lightning talks of two minutes, talks that are responses to a prompt, etc.  There are a million ways to reduce inhibiting pressures here and you’re the real experts – it simply starts with not accepting the status quo.

Conference organizing is a tremendously difficult creative endeavor.  And as with many such things, those who are responsible for it frequently resist the notion that data can be helpful, in part because it feels like it may destroy the creative impulses that take a good conference and make it a great one.

I want to bring a different perspective.  As a behavioral scientist, data is one of my primary tools.  And yet my job, which can be summed up as the designing of interventions that change behavior, remains highly creative.  Rather than removing the need for ingenuity, data allows me to spend more time actually doing the creative work I like doing.
First, data tells me where to look, not just for behaviors that can be changed but also at what levers might be most effectively pulled to do so.  For example, let’s pretend I have a behavioral goal around connecting (a frequent topic at PCMAEC), something like “All attendees will leave the conference having met three new people that they talk to at least quarterly for the next eight quarters”.

Without data, I have no idea how close or far I am from that goal.  But more importantly, data isn’t just a scorecard.  A data-driven perspective on what is already happening allows me to spend more time on the why and thus on the how of change.  For example, by understanding who is already meeting the connection goal at my conference, I can investigate what is different between that group and those who aren’t connecting, and then design an intervention that bridges the gap.  But to do that, I need a data-driven perspective on what is already happening.

Data also unlocks new features and products that wouldn’t otherwise be possible.  Think about trying to address the frequent manager question of “Where should I send my employees to get professional development?”  In a world without data, the best we can do is persuasive marketing: employees get sent to whoever tells the best story.  But with data, the story can itself be driven by verifiable truths.  A manager can decide what variables matter, like where their competitors are sending people, the ratings, novelty, and diversity of speakers, average seniority and background of attendee, etc. and then make a decision based on their priorities.  That’s powerful and it will make for better conferences.
Even if none of that convinces you, data is here:  Sponsorshipped, Feathr, and others aren’t going anywhere.  Conference sponsorship and attendance is the last unmetriced frontier of marketing, and as we’ve seen across other industries, you can expect that sooner or later, everyone will be using data to evaluate this spend.

But if you’re confident that you’re producing a great conference, that should be empowering.  The only people who should fear data are those who are actually bad at what they do.  In every industry where data has been embraced, spend has gone up for those who do it well.  Think of it this way: Google would never have built Google Analytics if they thought it was going to drive down spending on Google Ads.  Putting clarity into the value of both sponsorship and attendance is an opportunity to show how importance conferences actually are.

So when the startups come knocking, give them the right feedback to control your own data destiny.  Because you can influence how this all plays out but only if you lean in rather than out.

Paying Speakers
People are often surprised to learn that I don’t belong to a speakers bureau or charge speaking fees (don’t judge me by my PCMAEC appearance; I’m normally providing more value).  But I want to argue that not only will this becoming increasingly common, it will also become the dominant norm, and it will change the entire conference industry.
Before trying to prove this, it is important to first clarify my policy.  I do allow conferences to pay for my travel, so that speaking doesn’t actually cost me money.  And if they already have a budget and are paying all speakers, I’ll ask them to donate the money to a domestic violence shelter in that market, as Sanders/Wingo recently did for me in El Paso.  So there is still some budget changing hands here, although I’ll pay my own travel if the audience is important or unique enough.

I also have to own some privilege here.  One of the reasons that I can afford not to charge for speaking is because I choose not to make speaking my job and have had financial success in other areas.   I build things for a living and that is literally the only thing I allow people to pay me for.  And as a white dude, I don’t have to make conferences pay me in order to have them take me seriously.

This is real and important.  Many of my underrepresented speaker friends have had horrendous experiences, from being denied a private space to nurse in to being asked to write extensive blog posts that white male speakers weren’t to extra “content vetting calls” before they were allowed on stage.  Nobody should have to ask to be paid simply so they can be taken seriously and the speaker community is small; when conference organizers engage in these kinds of activities, they lose both access to premium speakers and risk potential public exposure and the corresponding loss of attendance and sponsorship.

Issues of respect aside, there are several reasons I see speaker fees going away.  Let’s start with a logical axiom: all things being equal, the speaker who doesn’t charge can make more appearances than the speaker who does, simply because more people can afford to put them on stage.  And now more than ever, getting on stage has downstream monetizable effects.  As the working world moves increasingly away from execution-focused tactical work to knowledge-focused strategic work, demonstrating an ability to be strategic and thoughtful is what gets you a high paying job.  Even if I charged for every speech I did, it would pale in comparison to my actual salary, which is in part based on demonstrating the competencies that I show on stage.

So there is a simple economic motive to not charge for speaking, if getting on stage elevates the chance that you will take some other higher value action, like booking someone for consulting, hiring them into your company, etc.  This is different from a stage action in a non-monetizable field.  For example, if you want me to dance, you do have to pay me, because dancing doesn’t lead to Chief Behavioral Officer.

More than just speaker economics have changed, however.  In the traditional conference format, you were paying a significant amount of money specifically to gain access to a speaker.  But this was developed in a pre-internet era, where people were unable to get free, high-quality content with the same ease.  Fifty years ago, the only way to understand my view on a competing pressures model was to attend a lecture in which I spoke about it.  Now, you can just go watch a free YouTube recording of one of my talks or read my book due out next year, and you can get access to my knowledge easily and for a fraction of the cost.

But despite MOOCs and other methods of potentially accessing knowledge, people still go to college in droves.  Why?  Because self-motivating is hard.  You could go watch a YouTube video of me giving a talk but it won’t be affective in the same way watching it live will.  Because speakers respond to audiences, every live talk I give is different.  And there is still an important part of learning that requires human interaction, not only during the speech but after.

To put it differently, at PCMAEC, people identified the two dominant reasons people attend conferences: learning and connection.  On YouTube, you can’t ask me a question or grab a drink and introduce yourself.  You can’t walk out shaking your head and talking to someone you just met who was sitting next to you about how terrible I was, then exchange business cards.  Yes, YouTube has commenting and I could do a Q&A video, but there is a very real difference between computer mediated interactions and the ones we experience in person.  Indeed, as one person so eloquently put it on Twitter, because they are a remote worker, they now go to conferences just to be around people.
But what does this have to do with the death of speaking fees?  Well, if we accept that the movement is away from conferences that are simply about providing access to knowledge and toward a more interactive form of both learning and speaking, the monetary value of speakers will eventually decline as the conversations we have before, during, and after, and the corresponding connections we make around those conversations, become the primary value driver for attendees.  If you think of speakers as simply the fodder for that connection, then their individual attractive power lowers.

Now that doesn’t mean speakers are valueless; in the way that a star professor can attract students to a university, a star speaker can certainly drive ticket sales.  But if the value of each individual speaker to do so is going down, and the ancillary value that speakers harvest from simply speaking is going up, at some point those cross over.  Couple this with the fact that more people can actually be trained to become better speakers as equitable access to education continues, you’ve got a recipe for the end of speaker fees.
To look at it differently, think about sponsored speaking slots.  At the moment, companies pay big money to essentially buy mainstage speaking time, because they recognize the brand halo that it has: not only can your smart exec talk about your product, people also respect the company for employing said smart exec.  But companies could easily pursue an alternative strategy, like we did at Microsoft.  One of the smart folks on my team had the brilliant idea of simply removing inhibiting strategies to grabbing the speaking slots we didn’t have to pay for, by training smart people to be better speakers, helping them apply for slots, and then paying for their travel costs.

So if the many smart people get trained on better speaking and you end up with a plethora of amazing speakers who are willing to do it for free because they can find other ways to monetize, why would any conference organizer reasonably pay speaking fees?  Instead, they can use that cash to democratize access by paying for travel and double-down on actually respecting speakers’ time and effort.

A Final Note
It is important to end by recognizing the graciousness of the PCMAEC audience.  As is my habit, I was entirely authentic onstage and noted that I was having trouble – after asking for an extra round of applause to give myself a moment to recenter, the crowd graciously obliged and afterwards many of them said they thought it was just shtick because the talk itself wasn’t bad.

But trust me…it isn’t shtick.  The talk wasn’t good.  But hopefully at least some of what I would have said came through in this article and we can all go make better conferences.  Because that’s what really matters.  Now more than ever, people need to come together, to debate and learn and connect and just generally cause trouble while opposing the status quo.  As conveners, conference organizers are far more than functionaries – making sure the drinks are cold is table stakes, but the real work worth doing is the behaviors that remain changed days and weeks and months after attendees go home.  Done right, conferences can change the world.

Side note: As everyone is painfully aware (because I won’t shut up about it), the lack of representation on stage pisses me off. I’ve tried a variety of small experiments to tackle this, but I’m ready to step it up a notch. Using some of the advance money from my book, I’ve hired someone to manage a small project we’re calling Speakershipped.  The idea is very simple: if you are an underrepresented speaker, we will essentially act as your free speaker bureau. I will personally help you uplevel your speaking skills, we’ll construct a bio and several proposed talks, and then we’ll actively pitch you to conferences. This will come at no charge to you and you’ll be in complete control over where you want to speak, what your acceptable parameters are (they have to pay for travel, main stage only, etc.), and how you want to appear. It is my hope that by taking a more active role than the traditional “let’s make a list of underrepresented speakers” approach, we can see much faster change.  If you are interested in speaking or are a conference organizer willing to accept pitches, please email and we’ll get started.

Recently I’ve been working to evangelize the use of behavioral science in business, which often means working my network to get to the top levels of companies so I can do a bit of evangelism. And in doing so, I’ve learned something interesting: I don’t know corporate leaders. I don’t even know people who know corporate leaders.

You could argue that the first-degree gap isn’t surprising. After all, I came from academia and specialize in a field that is only newly being applied at scale in the corporate world. But it isn’t like I am generally unconnected: if you name a successful startup person, particularly in NYC, it is a fair bet that I either know them or am one degree away. I give about fifty talks a year, all across the world. I’ve got social media followers. Hell, you’re reading this!

And the gap isn’t unique to me. It isn’t only that I don’t know many people at the top of corporates, but due to the glory of LinkedIn, I know that I’m not even a hop or two away. There is something bigger going on here and I think it is at the root of a serious problem in increasing the efficiency of innovation.

If you look at who is running the top three levels (CEO, C-level direct reports, and their direct reports, the Baby C’s) of the Fortune 1000, a pattern emerges. Take my old boss, Satya Nadella of Microsoft, as a sort of archetype. Graduated from college in 1990, goes to work at Sun, joins Microsoft in 1992, finishes his MBA in 1997, appointed CEO in 2014.

Now admittedly, that’s cherry picking. But look at Nadella’s top level: minus the folks just acquired from LinkedIn, only one has ever worked at a startup (which was acquired by Microsoft in 1997). They all follow a fairly similar path: corporate, MBA, more corporate or consulting, then pick a company and spend 10+ years there. Along the way, you meet a bunch of other people doing the same and you all employ each other as you move around the companies.

Entrepreneurs have a template as well. You work at a startup, you found a startup, you fail or get acquired or get big…repeat. And just like the corporate folks, you meet a cohort of people who you value and they become the tribe that you recruit from and party with. They become your friends.

Neither of these two templates are bad on their own. But what this difference in path creates is a gap in social circle. And in a world where social circles create innovation because of creative collision, that’s a real problem.

Take recruiting as an example. I’m acutely aware of how limited my brain is. When a founder asks “Do you know a good X?”, I generally think back over the people I’ve talked to recently and who I have coming up, because that’s about all my puny memory can hold, a month’s worth of people.

I have to imagine that is probably how it works for corporate folks as well. When they see each other for drinks and ask those casual recruiting questions, I’m sure they suggest people who are top of mind for them. And because that social circle was formed over years in a corporate environment (and MBA programs; for entrepreneurs, there is probably an incubator cohort bias), they regurgitate other corporate people, just like I did for other startup people.

Again, this isn’t terrible on its own, except that we need crosspollination. Corporates need more innovation from entrepreneurs. Entrepreneurs need to do more business development deals with corporates. And as research points out over and over again (mostly because we continue to do nothing about it), diversity of viewpoint makes for stronger, more profitable companies.

I’m sure that at the very highest levels, this gap probably isn’t as real; Reid Hoffman probably knows plenty of CEOs and Satya likely knows plenty of founders. Although I do wonder how personal those connections are – at the backyard BBQ, is there really a mix? Does Indra Nooyi invited Matt Salzberg over for coffee?
And maybe that’s the challenge, the go forward action: we all take a second to find someone on the other side of the gap, in a similar role, and invite them to have a drink or a coffee and just talk about the areas of mutual overlap. Or the Yankees. I’m not sure it matters that we talk about business so much as we simply take the time to get to know each other.

And yes, I promise to do it as well: I just sent Mauro Porcini a note.

Side note: Because this was on my mind, I tried a quick Twitter question, asking my followers if anyone could think of an example where a Fortune 100 that wasn’t recently a startup hired a startup person into their C-suite, other than through acquisition. Grand answer? A massive blank. Nobody could think of one. And even if one or two trickle in after this post, it is telling that this isn’t on the tip of our tongue.
Shouldn’t it be common? Take a serial entrepreneur; if they have two or more exits, do we seriously think they can’t contribute meaningfully at the top levels of a company? If we want innovation, we better start hiring for it.

One of the most consistent questions I’ve been asked over the years is where to find good product people (a term I use because nobody can seem to decide if they are product managers, project managers, or some other exotic variant). Most companies recognize the value that a highly-skilled product person can bring, so they are in high demand, and yet definitions of the role itself vary: It sometimes emphasizes deciding and scoping what will be built, sometimes managing the process of building, and sometimes owning the iteration back and forth between the two.

So unsurprisingly, when you look at the best product people out there, they don’t come from any consistent background. Some are former engineers or designers, others are subject matter experts, and some just seem to take to it naturally. But with a role so vague, it is hard to predictably find the right people to fill it.

And education isn’t helping. There are very few full programs that train product people and not even many good courses that do more than teach potential tools. So I’ve always struggled to answer the question, because if there is no consistent training and no consistent definition of what they should be doing, how can I possibly tell you where to find them?

Which naturally means I’m going to now try to do that. Recently, I’ve become increasingly convinced that the answer to Product is Science, and the answer to finding good product people is to look to scientists and those with training running experiments in a lab setting. There are a couple of reasons for this.

As I wrote briefly about recently, scientists are trained interventionists. The whole process of science is looking at the world as it currently exists and all we know about it, and then theorizing variations and potential outcomes. Which, I’d argue, is exactly what good product people should be doing. When you build a product, you are fundamentally trying to change some sort of behavior in the world. You hopefully have some understanding of what currently happens and why, what you want to happen and why, and then you experiment with methods that bridge the gap.

That’s exactly what scientists do on a daily basis. And they are highly familiar with the iterative process that it takes to get there and generally not frustrated by it. To have succeeded in science, they must necessarily have become accustomed to and patient with the universal truth that most experiments fail to produce the theorized outcomes (or often any discernable outcome at all). Behind every paper with a few studies in it, there are often a dozen failed experiments from which the scientist learned something about what didn’t work and then got busy creating another attempt at one that did.

And scientists are already trained to evaluate the data that helps in that iteration. While most are certainly not full-fledged data scientists, they are accustomed to looking at the statistical output of an experiment, evaluating the results, and then iterating to the next version. They can run regressions, are familiar with at least one statistical package, and don’t need to run off to a separate team to understand what worked and didn’t.
They are also accustomed to cooperation and advice, so when they do need help, they are likely to go get it.

In academic science, it is virtually impossible to go it alone: The process as it currently stands requires an advisor. Rare is the entirely independent scientist, alone in a tower trying to resurrect their personal Frankenstein. Scientists are trained in a tradition of teams, with lab meetings, plenty of white boards, and discussion and debate. It isn’t a solo enterprise and that natural emphasis on collective action translates incredibly well into the modern development environment.

And you can actually hire whole teams because we already have lots of scientists. One of the benefits of discovering that we can repurpose scientists is that we don’t have to wait for the development, iteration, and subsequent years of operation of specific training programs for product people. While science certainly has challenges and we radically need to increase both the raw number and diversity of those that enter it, there is still a wide pool to draw from: According to the National Center for Educational Statistics, we get about 300K new grads a year with some form of science degree.

But perhaps more important than intervention, iteration, experimentation, data science, and abundance, we have orientation. Scientists understand the random. They won’t simply abandon a method because it doesn’t initially produce outstanding results. They’ll tinker with it, even run it again to make sure some outside variable didn’t interfere. Successful scientists, by definition, are those who don’t give up, who refuse to be discouraged.

Yet good scientists are in love with the problem, not the solution. They have only the lightest attachment to the experiments themselves and a strong passion for discovering the overall truth of the subject. I’ve never seen any group so ready to be flexible in method while holding steady on desired outcome, which is the literal meaning and whole point of a pivot.

And this is essential to good product. The number one reason companies fail is because they refuse to let go of product elements that are not driving the desired behavioral outcomes. Innovative products are by definition initially unknowable. The right solution must be teased out through experimentation and exploration. And to do that, we need trained tinkerers. To run product, we need scientists.

Side note: Yes, I fully recognize and embrace that one of the reasons I love the ideas of scientists as product people is because my background is science. But science is not without flaws. One of the reasons I left academia was because I felt (and feel) that academic science has lost sight of application. When I was shopping for PhD programs, the legendary social psychologist Tom Gilovich was the head of Cornell’s program and he called late one afternoon. “We’ve talked and we really want you,” he said, “but I just want to make sure that you’re clear on what we do here. You keep talking about applying psychology and we’re a research program – that’s what we do here, all day, every day.”

Tom saw something that I wasn’t yet mature enough to see, that despite my early successes as a researcher and love of expanding the knowledge of the field, I was never going to be happy with just research. My year as a Cornell PhD was one of the worst of my life and I was incredibly glad when it was over and I left to become Head of Product and Lead Scientist at Thrive. I probably needed that year to discover for myself the truth about a research-only focus but kudos to Tom, for calling it advance.

It is a strange time to be a behavioral scientist in business.  When I left my PhD program almost ten years ago to focus on real-world applications, I spent the majority of my first few years explaining over and over what behavioral scientists actually did.  Now, I regularly get inbound requests to speak at large companies and books by Adam Grant, Dan Ariely, Jonah Berger, Angela Duckworth, Amy Cuddy, Barry Schwartz, and others sell millions of copies.  Executive confidence that behavioral science is both valid and interesting has seemingly never been higher.

And yet hiring behavioral scientists to explicitly apply what they know has remained somewhat rare.  When I left Microsoft, I got far more recruiter inbound based on my background in startups and venture than I did in behavioral science, despite being considerably better at the latter.  And while there are a few corporate behavioral science groups, like Om Marwah’s burgeoning team over at Walmart, Steve Wendel’s at Morningstar, Charlotte Blank’s at Maritz, Prasad Setty’s at Google, and Jeff Helzner’s at AIG, by and large the Fortune 1000 seems to love behavioral science without actually applying it.

There is a fundamental gap here.  How can executives believe in behavioral science, espouse its virtues and recommend its writings, and yet not be investing in its application to their own companies?  Surely it can’t be all TED talks and no desire to see it work?

And we need it to work.  We live in a world where ~70% of people are fundamentally disengaged with the work they do (and that number hasn’t meaningfully changed in 15 years) and despite record highs in health and wealth, personal happiness is actually declining in the United States.  We are doing better but feeling worse and our current approaches to addressing that aren’t working.  If we want change, we need to employ people trained in and focused on behavioral outcomes.

So why isn’t that happening?  To me, the gap isn’t a failure of business but of behavioral scientists, myself included.  If we accept that executives across industries want to incorporate behavioral science into their companies, then we need to remove the barriers to implementation until it is applied repeatedly and at scale.  We need to demonstrate the kind of problems to which behavioral science can be applied and create clear processes for that application.

We’ve done a bad job so far.  When I was preparing to write this post, I asked some of my fellow scientists if they wanted to contribute examples of their work and methods.  The overwhelming response was “we wish we could”.  Because talking about human behavior and its explicit modification remains an area of corporate secrecy.  We trumpet the results but not the methods.

So I want to give an example from my last job at Microsoft and to call out the potential for a replicable process that could underlie the application of behavioral science at scale.  Because I believe every large company should have a Chief Behavioral Officer and team actively applying behavioral science to identify and improve human experiences, both internally and externally.  And I believe that showing how we do it makes it easier to do.

When I first came to Microsoft and looked at Bing, one of the things I heard anecdotally was that search volume was lower than expected in schools.  The working theory seemed to be that kids simply weren’t curious enough and didn’t know that search could be used to satisfy curiosity, and there were vague plans to run a marketing campaign encouraging curiosity-based search in schools.

So I got my hands on query volume logs and ran some regressions to confirm the deficit with data, creating a metric of “searches per student per day”, then went to go visit some classrooms.  Unsurprisingly, there didn’t appear to be a problem of inquisitiveness: anyone who has interacted with kids for more than a few minutes knows there isn’t a lack of curiosity.  Instead, three big inhibiting pressures stuck out, inherent in teachers’ concerns about search: a vague and unspecified fear about how student data was being used, advertising in the classroom, and potential exposure to adult content.  Plus teaching search itself was a grey area.  Was it a media skill, best left to librarians?  How did it really fit in the daily rhythm of a class?

After the classroom visit, I got a few people in a room and used a Competing Pressures model to design Bing in the Classroom: ad-free, safe, private search that was free to schools, coupled with daily mini-lessons that fit in existing classroom models.  We helped Engineering to build it, acquired a small budget to publish lesson plans, and convinced large districts to sign up in advance of the launch.

The net result?  A 40% increase in student searches at participating schools and an additional 15% at home, larger than any single previous product innovation at Bing.

Bing in the Classroom isn’t the only project I’ve built that changed behavior at scale.  One of my startups, personal finance tool Thrive, got bought by LendingTree because we proved that we could increase people’s credit scores by 20 points in six months through behavioral finance.  And GetRaised, a tool we built that helps women figure out if they are underpaid and what to do about it, has helped tens of thousands of women earn over $2.3B in raises by changing the approach from encouraging women to “lean in” (as if they don’t already want fair pay) to simply making it easier to act.  I can rattle off a laundry list of behavioral science interventions that have worked.

But of all the things I’ve built, Bing in the Classroom feels most like a process that can be repeated inside corporations to change behaviors of value.  And that’s the key.  I believe behavioral science can be used to affect a myriad of business problems that have proved resistant to other efforts: the hiring and retention of the underrepresented, the modification of products and services to work across cultures, entrenched user experience problems like payment and support.  But if we want the chance to address them, we need a replicable process for applying behavioral science to these problems.

It starts with theory-based observation.  Behavioral scientists, like all scientists, are trained to look for the systems that underlie outcomes.  And all undergo statistical training, often fairly advanced, in order to try to extract the signal of behavioral patterns from the noise of the human condition.  For Bing in the Classroom, it wasn’t enough to believe something generally about kids and search; we needed data to confirm it and observation to understand it.  Without those, we would be stuck running brand campaigns about curiosity.

Then there was the design process.  I use a Competing Pressures model and that is only one of many options, but what the methods share in common is a firm grounding in the science of intervention and behavior change.  Behavioral scientists are fundamentally interventionists, rather than operators, because the experimental process itself is about taking how people normally behave and then proving that a change in environment or interaction or situation results in a replicable change in behavior.  The goal from the outset is to alter the status quo.

That orientation is very distinct from the processes we teach to those whose primary job is operational.  Think of it as the difference between a test (atheoretical evaluation, which tells you that the green button is better than the red one) and an experiment (theoretical evaluation, which tells you why the green button is more likely to result in an outcome than a red one, and thus allows generalization to purple, blue, and orange).  Because behavioral scientists don’t just care about what but also why, they can better create interventions that change behavioral outcomes in more permanent and integral ways.

This focus on the why of human behavior also brings an openness to a wider range of potential interventions than most operators.  Unsurprisingly, when asked to generated solutions, people gravitate to those within their locus of control.  Marketers tends to find marketing solutions, like a curiosity campaign, and the engineers find technical ones; for Bing in the Classroom, they wanted to build out a special desktop client to be individually installed on each student machine (school district IT people, you can thank me for vetoing that outright).  Behavioral scientists bring something novel to table because their domain is behavior itself; all levers are fair game.

A removal from operational components also creates another uniqueness for behavioral scientists: The ability to let go.  There is a part of the Bing in the Classroom story that is rarely told.  Once it was launched and scaled to around 10M users, with the query volume increase proved, I stepped away and handed the program over to Marketing.  Indeed, the two marketers who took the handoff were named Marketers of the Year by the CMO; I was the one who nominated them.

While many at Microsoft thought this was odd, my scientist friends instinctively understood, because this is a unique facet of scientific research.  When a theory is proved, it becomes part of the corpus of knowledge that belongs not to the researcher but to the field as a whole, and researchers immediately transition to the next area of investigation.

This is fundamentally different from what we instill in operators, who are generally rewarded for retaining ownership when something is successful and thus change only when forced.  By constantly working toward solution and transition, behavioral scientists can be a potent force for change.  And politically, HR, marketing, and other disciplines have everything to gain and nothing to lose, because any positive lift created remains with the operators and the behavioral science team moves on.  This is why I like internal NPS as a potential measure of the efficacy of a behavioral science team: if you are doing your job correctly, other teams will recommend working with you.

Taken collectively, there is a replicable process here: theory-based observation coupled with data science leading to intervention-focused behavioral design, with a build-test-refine cycle and then a handoff of an empirically validated program that is ready for scaling and operation.  It can be applied to novel human experience problems that operational teams have had difficulty solving because it brings an entirely new perspective and is distinctively apolitical.  The team and expenses are small because your only costs are human capital and the potential upside can be directly linked to profit.

And that is an important word: profit.  In a world where businesses at scale have done everything in their power to maximize earnings, behavioral science is an untapped well.  Imagine the efficiency that comes from addressing the 70% of the workforce that is disengaged.  Or reversing the negative consumer happiness trend in the United States.  For the majority of large businesses, improving the human experience of a company is where the next massive increase in profit will come from.

There will be a first CEO willing to commit a few million dollars to hiring a CBO and team to drive change in their business.  But it is on behavioral scientists to make this work easier to embrace.  To show structure, process, and results.  Not a skunkworks of product ideas thrown against a wall, but a disciplined approach to behavior change.  Not tests but experiments.  Science, well applied.

Side note: One of the chief difficulties of enlarging the presence of behavioral science in business is the tendency toward secrecy.  Governments, including both the US and UK, have behavioral science teams that are transparent about their process and success.  Yet business lag behinds.  At Microsoft, we discussed my hiring but never noted that Bing in the Classroom was a direct result of a behavioral science experiment.  If we want to close the gap, we need to do so in public.  So if there are other behavioral scientists willing to contribute case studies to a followup post, send me a note.

Editors: This post would not be possible without the input of both other behavioral scientists and businesspeople, all of whom I am lucky to count as friends.  Thanks to Stefanie Sugar and Dominic Price, William Leach, Kevin Brilliant, Lauren Woodman, Kara Silverman, Jonah Berger, Adam Grant, Om Marwah, Steve Wendel, Jeff Helzner, Josh Wright, Michael Butera, Val Tsanev, Charlotte Blank, Dan Storms, Erik Johnson, Avi Karnani, Bill Cromie, Carson Miller, Andres Glusman, Michael Norton, and Kelly Peters.