How Radical Candor Can Revolutionize Your Team’s Performance | Jason Rosoff | Glasp Talk #17
This is the seventeenth session of Glasp Talk!
Glasp Talk delves deep into intimate interviews with luminaries from various fields, unraveling their genuine emotions, experiences, and the stories behind them.
Today's guest is Jason Rosoff, the CEO of Radical Candor, where he helps leaders and their teams build better relationships for exceptional results. Jason was a founding team member of Khan Academy, a nonprofit education platform offering free online courses, lessons, and practice for students of all ages. At Khan Academy, he started as its first designer and product manager and later served as Chief People Officer and Chief Product Officer. During his seven years at Khan Academy, he helped scale the product from 100,000 users a month to over 10 million users.
In this interview, Jason discusses his journey, including his transition from Khan Academy to Radical Candor. He shares insights on leadership and team building, explaining the core concept of Radical Candor: caring about the people you work with while challenging them directly. Jason also explores the challenges and successes of building a company around this concept, emphasizing the importance of feedback in fostering growth and improvement. He provides valuable advice for future founders and product managers, highlighting the need for a strong support network and an experimental mindset. Join us as we dive into Jason Rosoff's fascinating career and impactful leadership and education work.
Read the summary:
Glasp
Transcripts
Glasp: Hi, welcome back to Glasp Talk, and tonight we are very excited to have Jason Rosoff with us. Jason is a CEO of Radical Candor, where he helps leaders and their teams build better relationships for exceptional results. He was a founding team member of Khan Academy, a nonprofit education platform offering free online courses, lessons, and practice for students of all ages. At Khan Academy, he started as its first designer and product manager and later served as Chief People Officer and Chief Product Officer. During his seven years at Khan Academy, he helped scale the product from 100,000 users a month to over 10 million users. Today, we'd like to ask him about his journey and insights on leadership and team building. Please join me in welcoming Jason Rosoff. Thank you for joining us today, Jason.
Jason: Thank you so much. I'm excited to have this conversation with you. I'm really looking forward to it.
Glasp: Thank you. So, first of all, I know Radical Candor is a book written by Kim Scott. Could you explain what the concept of Radical Candor is to us and our audience?
Jason: Sure. So, Kim is my business partner. She and I founded Radical Candor, the company, together about seven years ago, about a year after the book was published. One of the reasons why I decided to do this with her after Khan Academy – it's hard to follow Khan Academy; it's hard to figure out what your next act is after something like Khan Academy, which has had such a profound impact on so many people. One of the reasons why I decided to do this with her was because of how the book and the concept helped me better understand what I appreciated about the leaders who I admired and what I was doing when I was at my best as a leader and a people manager. So, Radical Candor at its core is a really simple idea. It just means caring about the people that you work with while being willing to challenge them directly to continue to do really well or to do better. That's it. And it's when we combine both of those things simultaneously that the magic really happens – when we combine care and challenge. The way that I like to think about this is that, as I look back at my career, the best relationships in my career were all defined by people who really demonstrated to me that they cared about me and my success but pushed me to be the best version of myself, right? Pushed me to be better at my work. And that doesn't always mean correcting mistakes. Sometimes it was really about encouraging me, pushing me by encouraging me to lean into things that, though they made me a little uncomfortable, actually produced exceptional results. So, I think that at its core describes what Radical Candor is and why it was so helpful to me because I just didn't have words before Radical Candor. I didn't have a way to describe succinctly what it was that people did that made me appreciate their leadership and what it was that I was trying to do as a leader.
Glasp: I see. And also, that's really impressive that the book is a New York Times and Wall Street Journal bestseller for multiple years and also translated into 20 languages with more than half a million copies worldwide. That's pretty impressive.
Jason: Yeah, we just actually passed one million this past year, so we hit a new milestone. Yeah, one million copies sold worldwide. Incredibly impressive. I mean, Kim has this amazing talent for reducing ideas down to their essence, the core, the nugget at the middle that is most valuable, and she's a fantastic storyteller. So, I think the combination of those two things is what made the book so popular.
Glasp: I see. And how did you meet Kim?
Jason: So, my transition out of Khan Academy – obviously, if you're Chief People and Chief Product Officer, it's not like you quit, right? There's a period of time where you're transitioning responsibility to other people and things like that. I made a commitment to myself to meet as many people who I thought were doing interesting things, especially in the space of leadership. That was kind of my focus at the time. I was considering becoming a coach, actually, like a full-time executive coach, and I read Radical Candor. I thought it was amazing, and I was like, "Oh, there's no way. Who do I know that would know Kim or how would I even get in touch with her?" So, it was sort of on my list but kind of fell off. Then a friend of mine just mentioned that, hey, he was raising money for a startup, and he's like, "Hey, one of my investors introduced me to this person, Kim Scott. It seems like she would know something about this." And I was like, "That's great, but before you make the introduction, let me think about what I would want to say. What do I want to have a conversation with Kim about so I can make good use of her time?" She had just written this blog post, because while she was writing Radical Candor, she actually started a software company. If you read the first edition of the book, it's very present in the first edition of the book. It slightly changed in the second edition because, you know, the time had changed, and she was able to update the story slightly. But in the first edition of the book, there's all these references to this company that she started. This is not the company that she and I have been working on. Really, what they were trying to do is build software tools to help people put Radical Candor into practice. They had built a handful of prototypes, all of which had failed, and it had become fairly clear that, in Kim's words, Radical Candor is about putting the phone down and looking each other in the eye or being really present and having a conversation with a human being. If that's what it is at its core, taking your phone out is a sort of value-subtracting round trip. It's actually a distraction, not a help. She wrote this blog post about shutting down that company and how painful it was and how challenging it was in that moment to realize that this was a really, really difficult thing to do – to teach people social skills, essentially, like the skills that are fundamentally relational in nature. That blog post just resonated with me so much because at Khan Academy, I think one of the biggest learnings that I had was that context matters. Content is incredibly important, and what Sal was amazing at – Sal, who's the founder of Khan Academy, was amazing at creating great content. But in order to learn, context matters. An example of this is when we were doing user research, and we were talking to teachers who were working in schools in Silicon Valley, where you wouldn't think things like food scarcity are a problem. You're like, "Richest area of the country, one of the richest areas in the world." But these teachers were telling us stories about how kids were coming to school hungry, and that was really hard to manage, right? You're not in a state to learn if you haven't eaten that day. It seems so trivial, but what it made me realize was that technology can't gather – it doesn't have perfect context. It can't gather all that information. So, as we were doing all these things like building recommendation engines to figure out what concept followed which other concept and how to help people plan their learning journeys, one of the things that I realized was that that was only a tiny fraction of the context that was needed. We needed to understand in order to really help that person learn. Kim's blog post basically said the same thing, which is like, "I didn't fully understand or appreciate the context in which this tool was going to be used and what people really needed. As a result, I think we failed to be able to build something." In her defense, I think this was 10 years ago – well, eight or nine years ago now. I actually think that the technology required to build what she wanted to build didn't really exist. They needed to be a real tech company that was building AI or machine learning or whatever you want to call it way back when. It might have been a different case if Radical Candor is being published today with generative AI being as prevalent as it is. She might have had a different kind of success. But anyway, that was the crux of our conversation. It was just a commiseration of how difficult it is to teach people things with technology as an intermediary because the technology on its own often fails to have enough context to actually be helpful to the person on the other side of it. It was really through that conversation and just talking about how challenging it was that we built a bond. We both were at a point where we're like, "Maybe we just want to do some stuff that doesn't scale. Maybe we want to try some things that are just smaller and simpler, where we're more in control of the kind of impact that we're going to have, even if it's on a smaller number of people." I think it was those two things, the alignment around those two things, that was sort of the foundation of our relationship and just a genuine appreciation for the other person's point of view.
Glasp: I see. But you serve as a CEO, right, at Radical Candor. It's a book, and I'm curious, what do you do as CEO at Radical Candor? Do you do in-person teaching, like private sessions with several companies, or advising companies? Can you elaborate on that part?
Jason: Yeah. When Kim and I met in the summer of 2017, we didn't start the company until the winter. So, our anniversary date is November 6th or something like that. In those intervening six months, Kim and I were trying to decide if we were going to build a company, what would that company be like, what would it do, and why was it meaningful to build a company around Radical Candor. From Kim's perspective, she said, "Look, long before I published Radical Candor, I was teaching Radical Candor to groups of managers and individuals at companies." She was using the ideas of Radical Candor when she was on the learning team at Apple. She was using it with her team at Google – the precursors to these ideas before she had written them down in a book. She said, "I've seen the impact of it. When people actually implement it, I've seen what kind of positive impact it can have, and I'm case study number one. It has transformed the way that I approach things, and it's been really helpful to me. I've seen the impact it can have on others." So, she said, "I think that this is the product that works – this sort of executive education, teaching leaders how to think about this stuff and then teaching them some new behaviors that will help them practice these ideas with their team." That's really what we founded the company on. The other thing that she told me was that she was not looking to run a business anymore. Kim, in her heart of hearts, just wants to write. That's what she's always dreamed of doing. In some ways, all of her career has been a way to fund her writing. In addition to Radical Candor, she wrote Radical Respect. She's also written four novels and is working on a fifth novel. Kim is quite prolific in her writing, although Radical Candor is by far the most successful of those books. She said that if we were going to get into business together, whatever the business was going to do, I was going to have to run it, and she was going to act more like a board member, advisor, type of coach role. I was like, "Great," because one of the reasons why I decided to leave Khan Academy was because as much as I enjoyed scaling somebody else's business, I wanted a chance to do it my way. I know that's very egotistical, but I did. I wanted to see what it would be like if I got to make all the decisions that I had to defer to Sal and other people as we were growing Khan Academy. So, as CEO, I have 99% of operational responsibility for the business. What does the business do? We're an executive education company. We do talks and workshops primarily, but we also sell content and asynchronous digital experiences – think courses and things like that – on the topic of how do I apply Radical Candor, how do I use Radical Candor in my work to build stronger relationships and get better results. Over the last – we bootstrapped the company, and we were profitable from year one. I did the teaching. Kim did a bunch of talks – she likes to do the keynote – but I did basically all of the workshops, all of these highly interactive teaching sessions. Over time, we started to scale that up. Kim had a couple of people that she had trained on how to teach Radical Candor, and we brought them into the fold. We started essentially acting a bit like a speakers bureau, connecting people to trainers who could facilitate sessions on Radical Candor and how to apply it on their teams. Year one, we probably made like a million dollars in revenue. We're looking at six million in revenue this year. We've quadrupled the business in the course of six years, which for a services business, I feel very proud of. We did this all very intentionally. We've never taken debt. We've never taken investment. We've grown only based on the free cash flow in the organization. At the start, it was me doing everything – accounting, sales, operations. I quickly realized that that was an untenable thing. I brought on, initially, a part-time director of operations. Kim brought in one of her friends, Amy Sandler, to help with marketing. Amy's our lead coach today. She runs all of our coaching work, helps define all the content, and teaches our facilitators how to teach Radical Candor. Bit by bit, we built up the organization. Today, we're nine full-time team members plus Kim, and we have 10 coaches who live all around the world who teach Radical Candor in addition to a couple of us internally who teach it. We grew from something quite modest to something fairly substantial, not in a Silicon Valley hyperscale way, but in a very methodical, thoughtful growth to achieve a particular outcome in the world kind of way.
Glasp: Could you tell me why you don't like that style – the "get investment and try to grow as fast as possible" style?
Jason: Yeah, I think two reasons. One, I didn't want people telling me what to do, and neither did Kim. As much as she appreciated the investment that she got as part of building Candor Inc., which came before Radical Candor, the company that exists today, she didn't like the pressure that that put on her to achieve some objectives, even if she thought they were unreasonable, which I strongly understood because even though Khan Academy was a not-for-profit, we had major donors like the Gates Foundation and others. They were lovely to work with, but they were very opinionated about what we should do. Even though it's not investment, Kim and I both had that experience and thought, "We don't necessarily want that again." We were unconvinced that hypergrowth, the ability to grow very fast, was actually going to be particularly helpful to us because the whole idea was to get closer to what worked and what didn't work and try to pick apart, stay very close to – in some ways, this is Paul Graham's "do things that don't scale." Teach the sessions yourself if you want to understand the kind of impact that a concept like Radical Candor can have. Don't hire somebody to teach it. Teach it yourself or with a small group of people first and figure out what it is about this that actually resonates or what it is that works. We didn't want people telling us what to do. We didn't want scale necessarily right away because we weren't sure what parts of the business we wanted to scale. We could have taken on debt if we wanted to grow faster. That's another common way, especially for services businesses where you have cash flow and a customer base. You can take debt that allows you to grow faster. Kim and I talk about this not infrequently, but we've decided over and over again we've come to this decision point where we say, "We could do this. Do we want to?" We've always answered no. We'd rather grow more slowly and grow within the bounds of what we can afford than take on either investment or debt. That might change in the future. I'm not closed off to the idea. Kim's not closed off to the idea, but it's been a good way for us to grow. It's allowed us to build the business in the way that we want without the external pressure of either large financial commitments or investment relationships where they have some level of control, like a seat on the board or something like that.
Glasp: I see. I'm curious, when you teach Radical Candor concepts and practices to executives, it's easy to understand the concept, but it's really hard to apply in daily work, in daily practice. How do you navigate them to apply the things they learned into their daily work? Do you have tips or advice, or do you follow up with them?
Jason: Great question. I think you're asking a couple of questions packed into one. How do we, in the moment, translate the concept into behaviors or actions – things that people can do in order to achieve better results? And then the second question is, how do we help sustain that learning over time? How do we help ensure that people are applying that learning after maybe an initial session? The answer to that is, I think we have good answers on the first and okay answers on the second. One of the things on the first thing – one of the things that Kim did really well and the place where we've probably spent the most time iterating over the last few years – is getting very practical or tactical about what Radical Candor looks like in the real world. In addition to this care personally and challenge directly frame, she also has a couple of other frameworks that we've co-developed over time. One of those frameworks is the order of operations, as she calls it in the book. This is the idea that if you want to build a team on which it is easier to give and receive feedback in order to help one another grow, there are some things you have to do in a particular order. The order of operations is get it, give it, and gauge it. The idea here is that somewhat counterintuitively, if you want to build an environment that is rich in feedback, the most important habit to build is a habit of soliciting or asking for feedback on a regular basis. There's a great quotation, and maybe I can share the specifics with you if you all do show notes, that basically says that the act of asking for feedback, one, makes it easier for us to receive. Even if the feedback is delivered the same way, if we asked for it versus it was delivered to us unsolicited in the exact same words, we react differently. At a biological level, we react differently. They've done fMRI studies of people's brains and things like that to see what's happening inside us when people give us feedback. Two, the act of asking for feedback is one of vulnerability and builds trust. I think this is one of the first major lessons for a lot of people who walk into a Radical Candor session. They assume that the most important thing they can do is tweak the way they deliver their feedback if they want to get better results. "I'm going to learn a little change that I need to make in how I say my feedback to other people, and that's what's going to help me succeed." We're like, "Nope, actually, that's not where to start. The place to start is by asking for feedback." Then we tell people you need to give feedback, and that means both praise and criticism. This is the other thing: people think the only valid form – many people live in a world where the only valid form of feedback is criticism. That's the only thing that's going to help you grow, so that's the only kind of feedback that matters. It turns out that that's wrong on every level. The most valuable developmental feedback that people get is praise. It is helping people identify – another way to say it is positive target identification. It's basically helping people see or understand or perceive what they do that actually works. The example I like to give is there was a study done – it has nothing to do with work – it was done with bowling, the American sport. I think your audience will be familiar with it even if it's global. It started with a bowling coach, and they did this study where they had the coach review tapes of people bowling. In one condition, the coach would criticize – "Oh, you put your foot in the wrong position. Your foot was turned in. You need to turn your foot out," for example, or, "Where you released the ball was too high or too low." They were giving tips – "Here's the correction you need to make." So, it wasn't just, "Oh, you did it wrong." It was, "You did that wrong, and here's the correct thing to do." In other situations, they had that same coach review the tape and say, "Oh, here's one thing that you did right. You should focus on doing that thing correctly again." The improvement was significantly more in the group that received praise than it was in the group that received criticism. It wasn't just about good feelings, because that's the thing that people assume. It's like, "Oh, it feels good to receive praise, and so that's why it worked." What was actually happening was if you're a novice and you're getting coached in something, if you're new to something and you're getting coached, it's very hard to know how to do something you've never done. Let's say the foot position thing – let's say you always make that mistake. You always put your foot in the wrong position, and then someone says, "Put your foot in a different position." Well, it's actually kind of hard to do that because you've never done that before. But if they're like, "Where you're releasing the ball is perfect. I really want you to stay focused on releasing the ball in that position." Well, that's something you've done before. It's easier to redo a thing that you've done before than it is to do something you've never done before. Once you focus on a couple of those good behaviors and those become rote, those become innate, you know how to do them, then it's easier to layer on top of that correction to the mistakes that you're making. Part is working really well, that feels really strong. Now let's layer on a correction to the mistake. That example seems really trivial, but it turns out that result has been shown in all kinds of feedback contexts. Coaching people on their strengths or the things that they do really well is a much more effective way to rapidly increase performance than focusing on the things that people do wrong. Now, obviously, we need to stop mistakes from happening, and criticism is very important, but the ratio of praise to criticism should be greater than one. In almost every organization that we walk into, when we ask people, "How much feedback do you give?" It's usually not a lot. That's one problem. The second question is, "How much criticism and how much praise?" It's almost always the case that they're like, "About equal or maybe a little bit more criticism. We don't really praise each other here." That's a really common thing that we hear. While there's no perfect ratio of praise to criticism, the research on that is that it's multiple – three times, five times, seven times the amount of positive reinforcement to criticism is actually ideal. Get it, give it – both praise and criticism. While we're not encouraging people, we're not saying it has to be two to one. We're saying be conscious of the fact that you want to focus on the good stuff and make sure that you're talking about what's working because that's actually going to help you and your team get better results faster. Gauge it is the last step. Gauge is really about measuring how your feedback is landing. This is one of the other misapprehensions that people have, which is if I give feedback and I do it well – well in air quotes – my job is done. I've done the thing. It turns out that also is not true. We need to be concerned about the impact or effect that our feedback is having on the people that we're giving it to. Whether it's praise or criticism, we need to pay attention to how the other person is responding and be willing to adjust our approach. There's lots of good reasons for this. First of all, we often give feedback the way we like to receive it. I like pretty direct feedback. It's helpful to me when people just say, "Hey, here's how I think about that thing. This is what was good about it. This is what could be better about it." People getting straight to the point is really helpful to me. But there are people on my team who do not feel the same way, who would rather have a more curious approach where I ask questions instead of sharing my perspective first. We need to be willing to make those kinds of adjustments because our goal as colleagues or as leaders is to make sure that we are actually communicating with the other person. It's not enough to say that. We want to make sure the other person is understanding what it is that we're trying to say to them, which means there's no one-size-fits-all approach to Radical Candor. In the book, Kim says Radical Candor is measured not at the speaker's mouth but at the listener's ear. I think that's a really simple way of thinking about it. We don't get to claim we've been radically candid. We can only know after we've solicited some feedback from the other person about how our communication actually landed with them.
Glasp: I see. That reminded me of a quote saying that people wanted to be valued, appreciated, trusted, understood, and so on.
Jason: Yeah, exactly.
Glasp: Thanks. Okay, then let's move on to your past career and Khan Academy, if that's okay. I'm very curious – you joined Khan Academy in 2010, and Khan Academy started in 2006. You were a founding member. How did you start working at Khan Academy? Did you know Sal before?
Jason: Not really. From 2006 to 2010, Khan Academy was basically a YouTube channel. There was a little bit of software development that was used with a very small number of people. But in 2009, Sal was trying to take some of the software stuff that he was doing and bring it to more people. He started an open-source project where people could contribute. That's actually how I got introduced to Khan Academy – as a contributor to the open-source project. In 2010 was the first time that Sal got any sort of funding to hire anybody. The first bit of funding he got was from Ann Doerr, who is a philanthropist in Silicon Valley. She heard or saw Sal in an interview or something like that and had a conversation with him. She asked him, "How are you funding Khan Academy?" He's like, "Out of my savings. I'm funding it myself. It's self-funded. I'm spending my own money to run it." It was at a point where he had a family, and they were like, "Can we continue to do this? How's this going to happen?" She was his first major donor. She went back and wrote him a check for $100,000 or something like that and said, "You shouldn't be funding this out of your savings." Shortly after that, maybe six months after that, he got funding from the Gates Foundation – a million-dollar grant. It was on the back of that million-dollar grant that he was able to start considering hiring and building a team. With that money, he hired one of his best friends from college, Shantanu, who was like the chief operations officer, COO type person. Then he hired Ben Kamens as head of engineering and me. Ben and I were working together at a company called Fog Creek Software in New York City at the time. We had been working together for five years at that point and had been contributing together to this open-source project. He reached out to us and said, "I'm looking to hire a team." Ben had been volunteering longer than I had, and he said, "Do you know anybody who might be able to do some product stuff?" Ben was like, "Well, Jason's actually been in the background helping me with the product definition, design, and things like that." We met for the first time over the phone. I was in New York, and he was in Silicon Valley. Ben and I flew out for a week to meet the team, to meet Sal and Shantanu. The rest is history. We collaborated together over the course of that week. Sal was pretty excited with the direction we were able to take things, and he offered us the job on the spot. The job was for significantly less money than we were currently making, and it required relocating to California. It was a little complicated of a choice from a family perspective. My wife and I had bought a house in 2005, and we were like, "What do we do? Do we sell the house? What if Khan Academy collapses? It might not go anywhere." The classic startup story – do we uproot our entire life for this? But I was lucky enough to have a partner that was excited for an adventure. It turned out to be an incredibly good decision, not only because Khan Academy turned out to be very successful but also because we loved living in California. It was a great change of life for us. It allowed my partner, Jillian, to change careers too. She was in sales up until that point, and now she's a therapist, a marriage and family therapist. She went back to school and started her new career in California. But that was part of the deal. I said, "If we move, I get to do something new, but you also get to do something new if you want."
Glasp: I see. I'm curious about your early days of hiring. In the early days, if you hire someone, it affects the culture a lot. The percentage, the ratio, is higher. I wonder, how did you find the right person in the early days? Because those people set the culture, right?
Jason: Yeah. I think the truth is we didn't always find the right person. What I will say is that one of the big advantages of Khan Academy – and I think there's a similar advantage at Radical Candor – is that there's a philosophical underpinning to the reason for the business to exist. That really helps at weeding out a big chunk of people who you wouldn't want working there. The idea of caring deeply about educating people for free all over the world, at its core, is this moral stance that high-quality education is a human right. So, it tended to attract people who already believed those things, which helps a lot in maintaining a culture that is really mission-driven. The fact that Khan Academy was a not-for-profit also helped because you knew when you joined that there were never going to be stock options. This was not the kind of thing where you're going to get rich. That filtered out a whole bunch of other people, especially in Silicon Valley, who would not have been great fits for the organization. But we made some mistakes along the way. I'm not going to say that we didn't. Especially because there were some things that we had to get good at that we weren't necessarily very good at, like fundraising. Sal is actually really incredible as a fundraiser, and I think he's talked about this publicly, but Sal is terrified of flying. So, it was literally painful for him to fly to do this fundraising. We tried a few times to find the right person to do fundraising for the organization, and that was really hard. One, we didn't have a great perspective on how to do it. Ben and I, on the product and engineering side of things, had a philosophical approach rooted in what we had learned from Joel Spolsky and the team at Fog Creek for how to hire, what we were looking for, and how to approach software development. But that was an area of the business where we didn't, so we had some false starts there. We hired some people, and they weren't quite the right fit. They were maybe used to fundraising for a different kind of nonprofit organization than the one we were trying to build. That's an example. We also made some bad hires on the product team – people who weren't exactly the right fit. But the benefit, in some ways, is that it was apparent pretty quickly that they weren't the right fit. I think that has a lot to do with the ethos of the organization being really clear to all the members of the team. I think the same thing is true with Radical Candor. There have definitely been some people that we worked with over the years that we thought were going to be great and turned out not to be great, both from a cultural fit or contribution perspective or just from a working style perspective. Radical Candor is a completely distributed organization. We have no offices anywhere in the world, and that takes a special kind of person to do that. You need a different kind of mentality to be effective in an organization like that. The long and the short of it is, I don't think the goal is to avoid mistakes. Yes, try to avoid mistakes wherever you can. I think the goal is to respond to mistakes with action as opposed to inaction or paralysis. I would say I have gotten a lot better at that over the course of my career. I have felt really empowered to take action at Radical Candor. Khan Academy struggled a little bit more, especially in the early days, with reacting quickly to problems in hiring, but it got much better at it as time went by. I feel like that's where the magic happens. You have to be willing, when you're growing quickly, to take a chance on people. You have to be equally willing to pull the plug if that experiment just didn't work out. One of the things that we've been trying at Radical Candor is more of long-term consulting to hire. We might hire somebody full-time as a consultant for three months or something like that. It's not just a trial period to see if they'll fit in the organization. For us, it's also for them. They get the opportunity to experience this, and we pay them really well. It's not cheap, but it's less expensive and easier to break up at the end of it if it's not working out than if you hire somebody outright from the beginning. I think more and more organizations are trying that.
Glasp: A quick question, by the way. When you interview people and you have a gut feeling, you see their skills, performance, and communication skills, but also you have a gut feeling about that person sometimes. Do you advise trusting that feeling? I'm curious about your opinion on gut feelings. Sometimes everything looks great, but my gut feeling says no. Have you gone through that situation?
Jason: Absolutely. Here's the thing about my gut – it's unreliable. What I've learned is that when I get one of those feelings, it's usually an indication to look underneath that feeling. Usually, the gut feeling is like, "I feel uncomfortable for some reason." Instead of making a decision based on that gut feeling, what I've learned to do over time is to try to understand the reason beneath the gut feeling. What is it that I am uncomfortable with about this situation? Can I name it? Can I get more specific than that? That is probably the biggest lesson that I've taken away. The second thing that I'll say is that I don't think you should make hiring decisions based on gut feelings. I also don't think a single person should make a hiring decision completely on their own. I think that hiring, promotion, and firing decisions should have at minimum two people's perspectives in there because there's always a risk that the source of that gut feeling is bias or prejudice against that person for some reason that I might not even be fully aware of. One thing that we instituted at Khan Academy to help with this was we had a rubric, meaning we had a definition of what we were looking for in people, and we asked people to evaluate against that rubric. That was number one, to help reduce the amount of bias. But we left a blank spot on the rubric that said, "What else did you notice about this?" The approach we took was that the results of the interviews that came before you were not available to you until you had completed your own interview. You couldn't look at what other people had said about a candidate until after you completed your interview with that person. What we noticed was that those gut feelings were often reinforced when other people had the same observation, blinded without knowing that the person before them had a similar feeling. That was a really helpful way to turn the gut feeling into data that you could actually respond to in a way that was fair to the person who was applying for the job.
Glasp: I think we experienced the same things that you mentioned just recently. That's why we are noting. You are the right person we should ask. By the way, you were working as a CPO at Khan Academy. How was that transition from Chief People Officer to Chief Product Officer?
Jason: There was no transition. I did it wrong. I did both jobs at the same time, which I strongly discourage. If I learned one thing, it's that I am not the kind of person that can do two big jobs simultaneously without it harming me, like my health. I was effective, but I was sacrificing my sanity to do both of those jobs. Here's how it happened, though. How it happened is sort of interesting. As we were growing, the team that grew the fastest was the product design engineering organization. That was the team that was growing the fastest inside the company. We had all these great hiring processes in place, these rubrics. We also had leveling, like how to consider someone for a promotion. We built up these mechanisms for hiring and managing people that were working pretty well to attract and retain the kind of talent we were hoping to have, not just in product engineering design but across the organization. I said, "The reason why this is working is because we've treated this as a design problem." We said an organization is designed to serve the people that it hires and to ensure the people that it hires can serve the organization's mission and successfully execute the tasks and duties that correspond to the role the person is hired into. That is a design problem. I was like, "We've shown a successful pattern here. I want to bring some of that to the rest of the organization. I want to treat people operations as a design challenge." To Sal's credit, he was like, "Okay, go for it." I felt like the dog that caught the car. Sometimes dogs chase cars, but they don't catch them, and that's good because if a dog were to catch the car, it would be terrifying, right? They'd be dragged along by the car. I was the dog that caught the car. It was really challenging to do both those things, but also incredibly rewarding. I got to build a great people operations team. I'm super proud of the work that we did together. We built a more robust recruiting function and things like that on the side of the organization that was really good for the organization's ability to grow and to evaluate and retain talent.
Glasp: Not sure if this is the right question because Khan Academy is an NPO, but what kind of metrics or KPIs were you following at Khan Academy, if you can tell?
Jason: In which role?
Glasp: As CPO, I mean, as Chief Product Officer.
Jason: We had a handful of metrics, like the standard ones that every software company has, like user growth, engagement, retention. But we also had learning goals. We had some metrics that corresponded specifically to how much people were learning. We had ways to evaluate how much learning growth was happening on the platform, and those were quite important. It turns out there's a virtuous cycle between engagement and learning growth to a point. One of the things that was cool about working at Khan Academy was that we were part of a generation of apps like Duolingo and others. We were very friendly with each other, like the explosion of Coursera-type organizations. All these product teams were talking to each other about how to measure these things and what it looks like to be successful. We always landed on the side of we'd rather be more conservative around engagement and not try to push engagement so hard and stick a little bit closer to learning first. Learning as a leading metric or at least equal to engagement. At some point, I remember – and I might be misremembering this story, so I apologize to my friends at Duolingo – but at some point, they went the other direction and focused far more on engagement, thinking the learning would follow. They were seeing the same pattern that we did, which is as engagement goes up, learning goes up. They made a decision to make the app easier, which increased engagement dramatically, and they saw a negative response where total learning went down. It stopped increasing as engagement went up. To me, that was a lesson. I remember talking to my team, my product team, and saying, "Someone's going to do it. Someone's going to push engagement at the cost of learning, and we'll learn something from it." If I'm remembering the story correctly, Duolingo tried it and wound up having to revert the change to keep learning and engagement tied more closely together, even if that slowed down their platform growth potentially. Those are the KPIs. We also had some internal KPIs around employee satisfaction, understanding roles and careers, and things like that. We knew that our biggest risk was losing talent to the next big startup. We had product-related metrics and employee retention and satisfaction metrics that we paid very close attention to because retaining talent was a big challenge.
Glasp: You have seen the growth from 100,000 users a month to tens of millions of users at the end. Have you found any inflection points or turning points where something went right?
Jason: There were a couple of occasions where that was true, but the interesting thing about Khan Academy is that interest in Khan Academy from viral stories didn't typically get to our users. Our users were primarily school-age kids who didn't care about news stories about Khan Academy or that Sal wrote a book. Those viral moments attracted donors, which was really helpful in fueling money into the company. They didn't help fuel usage. The thing that helped fuel usage was building great content and making it more accessible on the internet. One of the big learnings for me about Khan Academy was that we had to learn to treat Google as an interface to Khan Academy because most people who came to Khan Academy for the first time got there by typing a homework question into Google. We had to learn how to make sure our interface spoke equally well to students and robots so that it was easy for people to discover the content. Great content, number one. Discoverability of content, number two, were the things that actually led to growth. There were some moments where we got significantly better at discoverability that had real growth implications, but I would say most of it was hard-fought year-over-year incremental improvements, making things better, and then the law of big numbers being on your side. One percent growth on 20 million users is still a huge number of new users. If you're growing 10% year-over-year, which is small for a lot of platforms, it's still huge for education. Two million more students a year getting access to high-quality free educational content.
Glasp: Sorry, this may be a stupid question, but out of my curiosity, do you think Khan Academy would have been successful if they were for-profit?
Jason: I don't think that's a stupid question at all. I think it would have made it very hard. If we had to monetize early, it would have been very hard. There would have been a lot of things that led us away from the mission. If the mission is to provide a free world-class education for anyone, anywhere, free of charge, being a for-profit company would have challenged us a lot. I don't think it would have been impossible, necessarily, but it would have been incredibly challenging. Even as a non-profit, one of the things we realized over time is that if we wanted independence from our funders, we had to find ways to generate revenue. We just had to find the right audience to charge. Over time, the way we accomplished that was by charging partners like the College Board, which makes the SAT. We charged them to build free SAT prep on Khan Academy. They paid us to make a free product for their audience because one of their objectives as an organization was to make it so that there wasn't such an advantage for people with wealth in taking the SAT. That was one way we made money. The other way was by licensing to school districts. You can buy a textbook or license Khan Academy and get an interactive textbook. Licensing school districts, and there's a virtuous cycle because school districts could apply for grants to get Khan Academy as a license. It gave us more independence from the grantors because the relationship was us to the school district, not us to the granter. That helped generate some independence, both in terms of decision independence and also in terms of not having to raise large sums of money every year. It's a lot better to have recurring contracts with school districts. That's a much more straightforward way of meeting your budget every year, less likely to go wrong.
Glasp: Totally, I understand. You touched on AI and education in the beginning. Nowadays, people are using ChatGPT, Anthropic, Gemini for many things, and some teachers, professors, and students are using AI or large language models to cheat in tests or homework. How do you think education or learning will change with AI from your experience at Khan Academy and your current job?
Jason: I have some sympathy for the college professors who are like, "Oh, they're using AI to cheat the system, and that makes AI bad for education." I have sympathy for that point of view, and I feel like our systems shouldn't be so easy to cheat. If we really want to evaluate a student's learning, there's a good question in my mind whether writing a paper on your own and turning it in asynchronously for a grade is really a good evaluation of a student's ability. Maybe that should be more like a thesis defense as opposed to a graded paper where there's an interaction between the teacher and the student. The problem is our system of education doesn't scale that way. It's very hard to give that one-on-one interaction time and maintain the economics of the system of education. AI could help with that. Interestingly, a way to use AI would be to conduct those defenses. The teacher doesn't necessarily have to be there, but the AI, with a recording, could have that back-and-forth with the student. The teacher could still do the evaluation, but the AI could conduct the interview. All of a sudden, that opens up an entirely new avenue of how to better evaluate a student. Now take that example notched up one – the student speaks English as a second language. The AI can conduct that interview in their native language and translate the results for the teacher who only speaks English. Now you have more equitable access to one-on-one time for evaluation, a richer evaluation, and you allow students to perform in the language with which they are most comfortable. That is far better for evaluating students who speak English as a second language. The possibilities that AI opens up are tremendous in the world of education, but education will have to refactor itself substantially to take advantage of them. I don't think we are very much right now. I know Sal and the team at Khan Academy are actually working on this, but for the most part, the way people are looking at this is how to bolt AI onto the old way of doing things. That is likely to produce incremental gains but not step-change gains in education. We need to reimagine the world of education in a way where AI can do things that have been previously impossible to accomplish at scale at low cost. Conmigo is the Khan Academy AI partner. Essentially, what they've done is they've made an AI that does not give you answers, which seems silly, but one way to make AI educationally more valuable is to have it ask you questions instead of giving you answers. A very simple switch – that's at the core of what they've done. Now, they've done a bunch of other impressive things on top of that, but at the core of what they've done is they've said the goal of the AI is to ask questions that lead the person to understanding the query they started with, as opposed to giving the person the answer, which is by default what AI does. In part because the closest thing to a natural language AI that we had before this was search, I think a lot of people are sort of pre-programmed to treat large language model chatbot-type AI as a search when it can do so much more.
Glasp: Really interesting and insightful answer. Thank you so much. Since time is running out, do you have some advice for future founders, product managers, or someone like you, from your perspective?
Jason: Two pieces of advice. One is, do not try to build a company alone. It's not just about having co-founders. What you need is support. A co-founder may not actually provide the support you need. If you don't have any support outside of that, you will struggle mightily. Founding a company is a lonely, difficult job. You need a network of support. The advice people give is to make sure you get along super well with your co-founder and that they're your support. I don't think that's feasible in every case, and I don't think it's reasonable to expect that. The only reason that advice is sound is because it assumes the person doesn't have a support network beyond that. In Silicon Valley, a lot of the worst breakups I've seen were the result of people needing something from each other that the other person really couldn't provide. It frequently wasn't related to the work; it was often related to the relationship, expecting that they were going to be close friends or really. That doesn't always work for every business, and it doesn't always make sense. What you need is not a co-founder but a support network. I'm very lucky – Kim is a co-founder but also an incredible support. I have my partner, Jillian, who's an incredible support. I have good family support, and that has made this doable. That's tip number one – know who your helpers are, know who your support network is going to be. It doesn't have to be your co-founder. Tip number two is to have an experiment mindset. One of the things I've learned over my career is that people who burn out are the ones who are so certain about things, and when that thing turns out to be wrong, they're flummoxed. They get derailed and dejected. They get sad. Have a growth mindset. In practical terms, treat what you're doing as a founder as a grand experiment. Explicitly make it the case that learning how you're wrong is just as important, if not more important, than learning where you were right about something. Celebrate the moments where you've learned something and be ready to pivot. One of the things people misperceive about science is that experiments are designed to show that something is true, which is wrong. Good experiments are designed to disprove a hypothesis, to put the harshest possible test against a hypothesis and disprove it if possible. It's very hard to prove something is absolutely true. It's much better to prove that it's very likely to be true, and the best way to do that is to test it as harshly as you can. That is the frame you should have – testing the hypotheses you have as harshly as possible and expecting them to be disproven. The most common result of scientific experiments is no result – I do not learn anything from this, neither prove nor disprove my hypothesis. Be prepared for that. Find the kernels, the small things you can learn from, like how you structured the experiment could have been better, and use that to fuel the next experiment. Have that experimenter's mindset, be excited to be proven wrong. It's much less disappointing because you're going to be wrong all the time. If you're doing something novel, you're guaranteed to make mistakes. If you're doing something novel over and over again correctly, it's by blind luck. That's the only way you're going to be successful all the time. If you expect that, it doesn't derail you. I learned that at Khan Academy – so many assumptions I made were just wrong, proven wrong over and over again. I got really good at seeking data before making a decision and understanding that even with that data, the decision I was about to make had a high probability of being wrong. The most important skill I had was to pivot quickly once I learned that I was wrong.
Glasp: Thank you for the great advice. That means a lot. This is the last question. Since Glasp is a platform where people share what they are learning as their digital legacy, I'd like to ask you what kind of legacy or impact you want to leave behind for the future or future generations.
Jason: Let me tell you a story. I had a person working for me at Khan Academy who was incredibly talented. She was a designer. She would often come to me very frustrated that we weren't able to make big sweeping changes to the platform. It felt like the impact we were having was incremental. I said, "I understand the frustration. Education is like this big rock. It's creaky and old, very hard to move. It's like this huge boulder that we're trying to change, and it's granite, very hard stone. The frustration you're feeling, I understand. You're a needle trying to run as hard as you can against that rock over and over again, but the rock is going to dull you. You're not going to win that fight. You'll just end up being a dull needle." I said, "Instead, I want you to be water dripping on the top of that stone. Be implacable, tireless, and constant. The only thing that's changing the shape of that rock is water. That's the only thing that has a chance of changing the shape of that rock, nothing else. You could try to get a jackhammer against it, but as one person, that jackhammer is not significant enough. The other thing that could change the shape of that rock is huge collective action – millions of people with jackhammers chipping away at that rock. But the risk is that the rock shatters, and you destroy value. So, as an individual, the better strategy is to be the water – drip, drip, drip on that stone and change it over time. Have patience because what you're doing is making a difference. It might not seem like a huge difference, but every single drop, if you look at a rock with water running over it, where the water runs over it, life grows. Plants grow from the water running over it. It's hard to appreciate that when you're the water because the stuff is growing behind you, but it's happening. You're making a difference." She found that really helpful. I hope that maybe some folks in your audience find that helpful too because that is the legacy I want to have. I want people who come after me to say that Jason helped to gently carve a path for me or helped to patiently change the landscape of the company or the problem we were working on in a way that made it better for everyone along the path. That's what I really hope – a little better every day. Khan Academy took that approach – little better every day, and I've seen it have massive impact over time. Khan Academy seemed like this huge thing changing all the time, but honestly, it was those little improvements that we made along the way that cumulatively made a huge difference over the course of seven years while I was there.
Glasp: Thank you so much for sharing the story. Thank you so much for joining today. We really enjoyed the conversation and learned from your experience and insights. Thank you so much.
Jason: Absolutely, my pleasure.
