Over the last few years, I’ve done thousands of office hours: first-come-first-served blocks that mentees set the goals for. And this is the dominant model for how mentorship is viewed in the workplace today: a 1:1 meeting, set on a calendar, with an explicit agenda. I’m also the dad of a wonderful, quirky nine-year-old. He does not schedule his growth around my calendar and follows no agenda. Guidance happens spontaneously, as part of the context of our time together. And this used to be the dominant model of work mentorship, against the backdrop of co-labor: a more experienced worker paired with a less experienced worker, doing a task together, with the knowledge transfer occurring contextually. As work became increasingly differentiated, the meeting model became the norm. But it doesn’t have to be that way. My son is nine and finally old enough to play video games designed for adults. Our first outing has been Valehim, a Viking-themed sandbox survival game, and it has created a context for conversations that I cannot imagine having outside of co-labor. This week, my partner decided she wants to play with us. But in order for her to not die immediately, she needs armor that matches our level. Which means we need troll hide. So my son had to form a plan to collaboratively hunt trolls and then act in a coordinated fashion, with one of us as bait while the other focuses on shooting the monster.We talk as we do this, sometimes about the tasks of the game but also about life and his experiences more generally. Sometimes the mentorship is explicit (I’m teaching about how the meta-game works), sometimes it happens implicitly (I’m listening and nudging as he rambles), but both happen organically. Contrast that with the modern work environment. Often we do our work alone, because increasing profitability means reducing labor costs means using technology as our co-laborer (this will get worse, not better, with AI). We pop out of this solo labor for meetings, of which mentorship is one. But what if mentorship reintroduced co-labor? In a perfect world, this would be role-relevant: we would pair code or write a communication together or design an intervention. But it doesn’t have to be. It can be as simple as introducing a quick game, a 60-minute hackathon on an orthogonal business problem, or anything that creates a context within which mentorship can happen. To borrow an educational phrase, “guide on the side, rather than sage on the stage.” This doesn’t mean returning to the office. But it does mean being more deliberate about planning for context, which is the third participant in any mentorship experience. Because if the goal of mentorship is to encourage growth, we need much better soil than we have today.
Last week, someone was murdered. In reality, many people were murdered last week but someone leapt to mind the moment you read it. On social media, people argue about the validity of who you think about and why you should have thought of someone else. I have never seen these arguments at a funeral.
I have been an executive at a health insurance company. I neither approved nor denied claims directly, but was responsible for the call center that handled both talking to customers and providers. Claims are denied on the presence and absence of evidence; my team was part of the evidence-gathering process.
Every claim is a research project. As with science, you start with the default that the claim should be denied and then gather evidence to reject that null hypothesis. Over time, heuristics form and then become laws, as immutable as gravity, because you cannot afford to revalidate them every time. So every flu shot claim, even if you’ve already had one this year, gets approved.
During my time as an executive, I am absolutely certain we denied claims that we should have approved. I am also certain that we approved claims that should have been denied. These mistakes happen in insurance for the same reasons they happen in science: our sources of information are imperfect and our judgement fallible.
Mistakes were made. I have made many mistakes. Many mistakes are mine.
(I pause here to make dinner for my son, who has a father that is still alive and has not yet been murdered.)
(Sometimes, my son drops his bowl, scattering soup across the kitchen. I am angry. He protests that it was an accident. I point out that it was an avoidable accident, that he made choices that made the accident more likely, that the accident was a matter of policy. We settle there, hug, clean up the food. I do not murder my son.)
(What if he dropped the bowl on purpose, because he is not a fan of soup? What if the dog gave him a big-eyed look in the certain knowledge that spilled soup increases Rover Oral Integration, also known as doggy dinner? What if, in spilling the soup, he burns my leg, does me harm, forces me to take out a loan to pay for more food and burn cream?)
(Should I kill him then?)
Science does best when free from outside influence. I have been a health insurance executive and I believe that people are best served by a single payer system run by the government. I believe this because evidence exists from the study of our own system and from the practical experience of other countries.
I enact this belief by voting. Sometimes my candidates win; sometimes they do not. Sometimes the candidate that wins says publicly that they will turn more of our healthcare system over to companies. I believe he will.
I also enact this belief by being visible. Sometimes, being visible gets you a nasty note. Sometimes, it gets you a death threat. Sometimes, you ask the police to do an extra pass by your house at night and they do it. Not to shoot you but to shoot those who would shoot you, because you are white and male and affluent and because someone cares if you die.
I would work in health insurance again. I would try to make it better. I would be glad when we got a single payer system. I would find new problems to solve. Because it is core to my beliefs that no one deserves to die. Not the murdered. Not the murderer. Not the murdered murderer.
(My son will spill the soup again. He will be a grown man some day. Aggravator and aggrieved, please do not kill him, through action or inaction. There are already enough ways to die.)
People often seem to forget the root of compensation is compensate: a force exerted in order to counterbalance an opposing force. Think of it like a balance scale; something is loaded on one side, so we compensate by loading the other.
But what exactly is it that we’re loading?
There are generally two frames that occur in a variety of contexts. The negative frame suggests that what is loaded is injury, compensating by recompense. For example, work creates harm by costing me effort and my paycheck is restorative justice for that harm. This justice view often creates a wage-style system with tightly-defined borders; I give you 8 hours, you give me $200.
The positive frame, in contrast, suggests that what is loaded is value, which is then compensated by shifting value. For example, work is a value-producing activity performed in sync with others and my paycheck is my share of that collective value. This value view often creates a variable-style system with aligned incentives.
Neither frame is right or wrong, they’re just different. But how you choose to engage with them will dramatically change the work choices you make.
For example, I have tended toward the value frame and while it is tempting to think of that as a reflection simply of white collar work, I’ve also had some very mundane jobs: IT repair, bouncer, retail, farm work. And I know plenty of white collar workers with a justice frame, who treat their job as something inflicted upon them.
Because I’ve chosen the value frame, I tend to only accept direct payment where I am certain I am creating value. I don’t charge for speaking or advising, for example, because both activities have potential, rather than realized, value; my advice may or may not be good and so the value it creates is highly variable. To capture the potential value, I often joke with founders that if they sell for billions, they can buy me a boat.
If I instead chose the justice frame, then I would likely charge for far more activities. If I fly around the world to speak at your event, you owe me for that effort; if it took me an hour to advise you, you owe me for that lost time.
Because I have chosen a way of viewing my work, it is difficult to write about both sides empathically; no doubt others may better express why they see their paycheck as a form of restorative justice.
But what is important is that like all frames, these can be shifted. Every workday is an opportunity to choose a different frame for your compensation, with both direct and indirect consequences for that choice. So be deliberate and do not let the frame be imposed upon you; you can decide how you want to view the relationship between the two sides of the scale.
Recently, I was talking to a company that helps employers gather feedback from their in-the-field workforce in order to boost employee retention. And while they have a ton of customers fielding surveys, the product folks were concerned that managers don’t actually seem to be looking at the feedback dashboard.
My suggestion was to use a Max Headroom: take a few customers and do away with the dashboard entirely, instead having one of the product managers meet with them weekly to deliver the feedback. That way, the SaaS company could eliminate the dashboard itself as a variable and focus on the core behavior of managers responding to employees.
The tactic is named after a character from the 80s played by an actor in prosthetics but billed as computer-generated; the goal is to have a human frontend with a computing backend. And it is similar to another testing methodology, the Mechanical Turk, where the frontend is computing but the backend is actually human.
When I suggest this to folks, the objection is usually that humans aren’t scalable. Which isn’t entirely true; for many SaaS companies, a bit of quick math shows that you can have a CS person take quite a lot of meetings and still be profitable.
But it also doesn’t matter. The point of a Max Headroom isn’t to find a scalable business model; it is to reduce the inhibiting pressures of a behavior to as near zero as possible. A meeting to go through the dashboard together doesn’t quite accomplish that; there are still plenty of inhibiting pressures like time, mental energy, etc. But it is meaningfully easier than trying to interpret and action a dashboard on your own.
The much larger, unspoken concern is: what if that doesn’t work? What if, even with human-delivered feedback, managers are still unwilling to do what it takes to retain employees?
And the sad answer might be that it is time to pivot. If you try a Max Headroom and your target audience still doesn’t do the behavior, even with inhibiting pressure reduced in unsustainable ways, then you may need to address your underlying behavioral statement.
There are exceptions where adding a human increases friction, like calling a taxi company instead of using the Uber app. But that is more typical in consumer products and less in SaaS, where value is usually produced by removing the cost of having a human do something.
Max Headroom tests are underutilized in product creation, because of the bias toward seemingly-scalable software as the solution to everything. But in an AI era, where we need to better decouple backends and frontends, they are more important than ever. Because if a human reading your AI’s script isn’t useful, your AI probably isn’t either.
Movies will have you believe that all fights happen in bars. But to me, few places feel quite so ready to break into a spontaneous brawl as the line to get on an airplane.
And sure, there is sometimes utility to boarding the plane early. When my son was young, it was helpful to have an extra few minutes to get him settled. And you’re guaranteed to have overhead bin space if you need it.
But most bags do make it onboard, the seats are assigned, and we’re all leaving and arriving at the same time. The value of being in Group C instead of Group D is marginal at best.
So then why do people crowd?
Some of it is about social standing. In a million ways, humans pay to demonstrate our privilege and that is unlikely to end anytime soon.
But there is another factor at play: people like to feel like they are getting their money’s worth. And airlines have clearly connected boarding early to value. Early boarding is associated with Executive Diamond 100K Premier customers, higher-priced ticket types, and even explicit charges if you fly Spirit or Frontier.
So if you’re sitting in the waiting area and they call your group but you don’t board, you feel like you’re giving up value that you’ve already paid for. And humans are extremely sensitive to that form of loss.
This is, of course, a deliberate strategy by the airlines – controlling who gets to go first costs them nothing, yet creates the illusion of value. But there are plenty of times when “use it or lose it” is an unintentional consequence of designs that creates unnecessary feelings of missed value.
All-you-can-eat subscriptions like Netflix are an easy example. By putting so many options on the front page, they maximize the chances you’ll find something to watch. But they also highlight all the things you might want to watch (and more importantly, are paying for) but will never actually have time to enjoy.
Ditto for multi-function devices. There are a plethora of tools in my garage that can absolutely do more than I am capable of (I’m looking at you, compound sliding miter saw) and every time I use them for my basic needs, I’m keenly aware that I’m somehow missing out.
In a world where products compete on value, it often feels like the optimal is simply stuffing in as many sources of value as possible. And it is great that Netflix has many options and my miter saw can do many things.
The key to balancing the gain of more features and the loss of not using them is to control how visible those features are. Netflix can make fewer recommendations. The miter saw can hide the more advanced controls. As applied behavioral scientists, we can sometimes create greater perceived value simply by being more direct with the highest value features and deemphasizing the visibility of those that most people will never get value from.
A few years ago, I was talking to an entrepreneur working on an easy way to share STD test results and he said something that stuck with me: HSV is a dumb reason to break up with someone, because about 50% of the adult population in the US has it, and so it is 50/50 that even if you dump this person, you’ll have to dump the next person too.
To me, this is a very clever reframe. It doesn’t try to debate anything about the morality of breaking up with someone or minimize HSV itself. Instead, it just focuses on the probability of an event occurring in the next instance.
This is on my mind this week because of a conversation with a mentee about her immigration status. She’s looking to change jobs, so she’s been applying and getting interviews, but doesn’t get selected to continue because she only has 3 years left on her OPT visa.
So I invited her to do one of my favorite activities: a reverse roleplay. She got to be the interviewer, I got to be her. And when she asked me the question about my immigration status, I used a similar reframe:
“The average tenure of someone at my level in a tech company is less than three years, so if you pass me over because of my OPT, it is just as likely that the person you choose to hire will actually leave in the same amount of time. So why not consider whether I’m the best person for the job instead of focusing on how long I’ll do it?”
Her job dropped.
Note that I’m not arguing the fairness of immigration status being a form of legal discrimination. And I’m not trying to convince the interviewer that I’m so much better than the next candidate that they should overlook the deficiency. Instead, I’m arguing it isn’t a deficiency in the first place because it is true for the majority of the population.
This works for all sorts of perceived shortcomings in an interview. For example, take a job description I read once that had both “Ten years of applied behavioral science experience” and “Ten years of insurance experience” as requirements – a combination that literally no one had at the time.
Rather than arguing that whichever you don’t have isn’t important or that you can make up for it in other ways, you can just rely on probability. By pointing out the non-overlap of those experience sets, you can level the playing field by arguing that everyone is going to have to learn something in order to do the job and so it might as well be you.
As the joke says, you don’t have to be faster than the bear, you just have to be faster than the other person they’re chasing. By pointing out equivalencies, you can reset the frame on yourself as a candidate to minimize shortcomings. You’ll still have to highlight your strengths, but it can sometimes eliminate at least some forms of bias.
Over the weekend, the former CFO of Oura Daniel Welch posted about the success of narrowing the product focus to a subset of women’s health. And one of the comments, from Diana Torgersen, caught my eye:
“Wonderful – but still wondering how Women’s health is a “narrow focus”? By focusing on women you’re focusing on 51-52% of the population so it seems it’s not that narrow, merely the other half of the population.”
The reason it perked me up is because a) that math is wrong when you’re talking about product narrowing and b) it is a mistake that tons of product people make, resulting in insufficient levels of focus.
Take Fitbit as a representative of the default mainstream wearable device (you could insert Apple, Samsung, or dozens of others in here if you want). Torgersen is right: they generally try to target 100% of the market, with an implicit focus on men (because the patriarchy sucks).
But they do more than that: they try to target 100% of the market across multiple features, including sleep, nutrition, exercise, stress, etc. That is 100% x 5 categories, or 500 product points.
If Oura had taken that approach, they would simply have halved the population but kept the same features; a Fitbit Fem that exclusively focused on women would still have 50% x 5, or 250 product points.
But Oura didn’t just narrow the population: they also narrowed the features by focusing on just fertility, pregnancy, and menopause. That’s 50% x 3, or 150 product points. And in reality, because fertility, pregnancy, and menopause are non-overlapping, it is actually just 50 product points: 50%/3 (each member of the population is in only one of the three states) x 3.
I suspect that Torgersen was really just trying to use a rhetorical question to make the point that focusing on women isn’t narrow and more people should see women as an important market; a true statement, for sure.
But it inadvertently also made a point about how people conflate population and features. When seeking product focus, many product folks will narrow population OR features, instead of population AND features. And the “and” is where all the magic happens.
Now to be fair, Welch’s post was really about expansion; Oura had already done 100 products points of everyone x sleep. But you could argue they started with 50 product points, by doing men x sleep; Welcome acknowledges that in the beginning, it was seen as a very masculine device. They could (and perhaps should) have started with women x sleep, but would still be 50 product points. And that was with a $2.3m seed.
Every product starts by changing one behavior for one person. Over time, you’ll remove Limitations and add Populations and Motivations as you expand. But start with one.
There is a routine that plays out in the workplace time and time again. I ask HOW you’re doing, you tell me WHAT is happening, and then we proceed as if those two things are exactly the same.
We don’t say “I’m sad because XYZ.” We just say XYZ, as if any reasonable person should know how we will feel in reaction to that specific combination of events. But just because I know what is happening around you doesn’t mean I know how you feel; humans are variation machines and what feels good to me doesn’t always feel good to you.
So a few years ago, I started explicitly trying to draw out the use of emotion words. First by moving away from coded wording; instead of asking how you’re DOING, I ask how you’re FEELING, even though we typically tend to think of those as the same thing. And reactions vary in a typical 2×2: not/noticing, not/emotional.
Some people don’t seem to notice and don’t change their wording. And that needs to be alright; inclusivity doesn’t mean that we demand emotionality from people who aren’t comfortable using that framing in the workplace. Ditto for noticing and not changing; if I offer the prompt and they choose to take it a different direction, it is on me to respect that.
Others don’t seem to notice but implicitly switch to an emotional frame. And that’s really lovely: I’ve learned a lot about colleagues that I wouldn’t normally have known. In particular, I hear more about racism, sexism, and classism in the workplace this way – opening the door to emotion also opens the door to events that are likely to have emotional consequences.
And then some people notice, pause, and then switch to an emotional framing. This is honestly my favorite, because I imagine that it is now slightly more likely that they’ll do this for someone else, making the whole workplace a bit more humane.
The other thing I’m trying (and largely failing, because my superpower is having a two hour long conversation entirely about you) is to use more emotion words myself. Now I could argue that rather like a noticing/not emotional person, I should have the right to keep my emotions private. And I do.
But creating change often isn’t about exerting our right to comfort; it is about being personally uncomfortable in order to make other people more comfortable. Think of it like standing up to give someone your seat on a bus, except you’re standing out emotionally so others have the chance to do so as well.
Maybe this isn’t the hill you want to struggle up; there are plenty of other ways to make more inclusive workplaces that don’t evolve emotionality. But consider trying it as an experiment: at least for today, use and solicit more emotional wording and then reflect on how it changed your interactions. Do you feel…happier?
TLDR: Almost all economic increases come from change and so profit maximalism demands a world where everything changes, all of the time. But change has a cost for individuals and communities; be cautious, in yourself and your designs.
Recently, I was introduced to the Benedictine ‘Vow of Stability’: a promise that, upon entering a monastic community, you will remain a part of it for life. Philosophically, it is rooted in a belief that instead of seeking out an environment of perfection, you have an ethical responsibility to improve where you are. And practically, it was historically good to have monks that weren’t monastery hopping, trying to find greener grass.
My introduction to the VOS was secular; it came from someone who agreed early on in their marriage to pursue stability as a family wherever they could. They recognized that voluntary changes often created additional stress and even as changes could create a step up in circumstance, there was a hidden cost of turbulence.
As they told me about the choices they made, it struck me just how much of modern capitalism requires paying a change fee in order to advance.
The easiest example is their home, which has quadrupled in value over the 20 years they’ve been in it. They have often contemplated selling in order to harvest that value but that would require destabilizing their family. So instead they bought an empty lot nearby, with the goal of eventually building a home they can comfortably retire in while remaining part of the community.
Slightly more nuanced is education. Their school district has consistently been in the bottom 25% of their already less-than-stellar state. But rather than move, they have sought out enrichment activities for their kids and volunteered to make things better.
I’m not suggesting that their VOS-inspired choices are right. I’ve moved for job opportunities, to give my son an education that fits him, and to be in places that support my individual and family happiness. And I’m content with the choices I’ve made.
But I was in NYC last week and it was just so easy to get business done, because my network there is expansive. When we moved to San Diego, I essentially hit the reset button, both personally and professionally, as did the rest of the family. As individuals, it is worth considering whether all change is worth it.
Perhaps more importantly, when we design interventions, we must be on guard against demanding too many change fees from those we intend to benefit. Turbulence upsets not just the individual balance but also the group; when my son changes schools, it is not only he who loses friends – each of his friends also loses him. What can feel small rapidly becomes larger.
VOS behaviors are all around. When a stock that pays a dividend rather than trying to constantly escalate in price. When a factory stays rather than offshoring. Many changes are choices and we could all stand to choose a little more carefully.
TLDR: Learning is not a smooth curve; we frequently grow in spurts and jumps. But those rarely align with external validation like graduation or licensing. Credentialism isn’t just inequitable, it is a business and cultural liability. And there is a market opportunity in refusing to accept the bias.
Recently, there has been a trend in mental health to switch to associates: psychologists who have not yet been fully licensed and cannot practice on their own, without supervision. This is driven largely by cost, as associates are 20-40% cheaper than their fully licensed peers, largely because they cannot create their own practices and must sacrifice some money to a mediating entity.
The requirements for an associate to become fully licensed vary by state but generally include an exam, some number of supervised hours practiced across various populations, and an annoying amount of paperwork that gets returned to you if you misspell something. Bureaucracy gonna bureauc.
Obviously, experience matters. In study after study, we can prove that people do actually get better at things over time; some version of “practice makes perfect” arose in pretty much every language for a reason.
But that improvement is lumpy. When we look at learning, people tend to proceed in epochs, with long periods of level performance between jumps in ability. And those jumps are hard to pin down; if you do a cohort analysis, when people make a leap up is largely unpredictable. People who are learning and working together don’t necessarily hit milestones at the same time.
And so it isn’t necessarily true that you’re a better therapist today than you were yesterday and it certainly isn’t true that you’re a better therapist the day after the state declares you fully licensed than you were the day before. No one magically produces better work simply because they walked across a stage and got handed a degree.
And yet the day after graduation, your chances of getting a job go up dramatically. Credentialism, or the tendency to rely on formal qualifications over demonstrated ability, is rampant in hiring simply because across a large number of applicants, it is very difficult to find a better proxy.
The problem, of course, is that not everyone has equal access to credentials. Licensing fees can run into the thousands of dollars in some professions, let alone the time investment of a bewildering number of forms. Credentialism tends to amplify sexism, racism, and classism, because the systems that bestow credentials are themselves sexist, racist, and classist.
It is easy to look at associates as inferior and lambast the companies that are increasingly relying on them. Because hiring isn’t the only place credentialism occurs; consumers can just as easily interpret a credential as a valid signal of quality, without actually determining whether there is a real difference.
But ultimately, that drives up market pricing. Consumers who put false faith in credentials end up spending more, without any additional return. And because consumers are buying based on the credential, people are forced to get credentials to increase their wages; the tail is wagging the dog.
As employers, we have structural tools at our disposal: temp-to-hire as a method of determining actual ability, removing degree requirements, etc. And as consumers, we can use reviews as an alternative to credentials. But all of these are choices: ultimately, it is on us to give people the chance to demonstrate their ability to perform beyond their credentials.
And there is plenty to be gained. Besides combating inequity, in a market economy that overvalues credentials, there is a price opportunity in not making the same mistake: you can get more and better, for cheaper, if you’re willing to fight the bias that others are falling victim to. Fighting credentialism is a moral imperative but if you need an economic justification, it is certainly there.