There there are basically three ways I think that behavioral economic can help companies. 1 is we call it the little tweaks, default sign up flow, and so on. The second one is more complex which is about adding motivation. And then there's the 3rd component, which I think is the most interesting. And that's about what is the role behavior Economics in the core product. So, Guy, we have dinner early with us, and then you're Power has thinker in psychology and behavioral economics.
You're a professor at Duke and the founder of the center advanced a Flint site. You've written a few Beller selling books And on top of that, you work with the world's best companies from Google to lemonade to motivate their customers, promote emotions, and behaviors that they need in order to grow. And what we wanna today is basically take all that wealth of knowledge and understand how startup founders can actually use them in building their companies and building their products.
And this is, you know, we've had many discussions on that in the past, and it's all been incredibly useful for me, and we wanna bring this also to our listeners So would be great, I think, for me and for people that are not sure to get the definitions for it out of the way. What is actually behavioral economics? And what's the economics part of this definition.
Yeah. So, you know, 1st of all, economics sometimes people think of it that it's only about decisions, about money, but it's actually not economics is a perspective on human behavior. Right? Economists get involved with how to set up taxation and education systems and all kinds of things. But economics also assume some very strong things. Basically, economists assume rationality.
They assume that people have all the computational power possible that there are long term thinkers, that we are long term thinkers, that there are no emotions. And it turns out that, you know, it's okay sometimes to make these assumptions but mostly it gives us misleading results. The thing about economics is it's both a descriptive theory and a prescriptive theory.
It's supposed to describe the world and describe some of it it's a prescriptive theory that tells you how you want to build the world. Right? So if you have a startup and you take the rational economic perspective, you would, for example, set up prices is a function of, cost benefit analysis. And behavioral economics comes in and says, you know, not so quickly. Humans are much more complex. Let's not assume perfect rationality. Let's instead put people in different situation.
Let's be empirical about it, and let's see how people behave. And it turns out that when you look at how people behave, it's very, very different than what rational theory predicts. So if we just look at pricing, for example, and I'll give you a story. I was in Australia. And I met a really wonderful bank there. I will not tell you their James, and they said that they are going to change the ATM fees. Right?
They say people come and get cash, and it was costing $2.50 Australian, and they're going to make it free. And they said, oh, people would love us. Because we're making it free. And I said to them that I think people would hate them if they do it for food because I said, look, you think to yourself that people would be grateful I think that people would say you bastard. Why did you charge me all these years? And I said, look, a lot of things about pricing is not about the value.
It's about fairness. And if you just reduce the price to free, you're basically telling people we could have charged you enough thing, we just didn't, and that would create a tremendous anger because of fairness, not because the service is not worth 250, It's because of fairness. And I said, what you should do instead is you should tell people, look, it really cost us 250 to provide you with this service. But if you don't have the money to pay for it, pay us less. It's up to you.
So I basically use the name your own price approach. This way you don't have to go to 0. Right? The rest that can happen is that people pay you nothing, but you're not communicating to people that this cost you enough anyway, they didn't do what I proposed. It was probably too late. By that time, they're already ready with the campaign and so on. It backfired exactly as I predicted. And in follow-up tests turned out that my approach would have been much better. Right? So we tested it.
It would have been much better. It would have changed completely the sense of fairness. Some people would have paid, by the way, not the majority would not pay 250, but lots of people would pay something for it.
And this is an example that you take the rational perspective, and it just leads you to the wrong conclusions if you understand the complexity of human beings, even for small decision, it matters a great and the interesting thing is that if you just took the dry economic theory, then their customers should have been ecstatic. Right? That's right. They should have loved That's right. They should have said, oh, we're going to join this bank. Go. Think about another example.
I'm just giving pricing examples. Of course, it's not just about pricing. But you remember what happened to Netflix when this is an old story, but you remember what happened to Netflix when they split their services? So Netflix started as a company that used to send DVDs to people at home, and then they had service on the side, but as the internet became faster and faster, the streaming service became more and more popular. And at some point, they said, okay. Let's split it.
I don't remember exactly the numbers, but let's say it was $14 a month. Instead of $14 a month for both streaming and DVDs, will make it $8 a month for DVDs and $8 a month for streaming. And Wall Street analysts were ecstatic about this. They said, oh, you know, revenues will go up. New people would sign up. A lot of good things would happen, and they announced it. And I think that close to a million people left them day. Now what happened was that it wasn't that the deal was not better.
It was better because most people did not need both. The DVDs and the streaming, they could use only one of them. So for most people, they were getting whatever they were getting before for less money, but what happened was that people felt it was unfair. Why are we increasing the prices for me? Now, by the way, you can increase prices. It's not that increasing prices is not possible, but you need to do it in a way that people feel is fair.
And when you do it in a way, it people feel fair, then you can increase price. When you just look like price gouging, people are really upset. And basically, sometimes people just cut their nose to spite their face. Right? They basically are so angry with the company. They're willing to not get the good service just as an act of revenge. And so if we try to simplify the difference between pure economics, kinda the dry numbers that make sense on paper, customers should be happy.
Most of them don't need both services, and so the price Morgan service is lower. They should be happy. That's dry. That's the computer answer for the problem. And the actual behavior people, that would be the difference between economics and behavioral economics. Right? This is what people many times miss that there's more in human behavior than just the dry numbers.
Yeah. That human behavior is a magical and fantastic and complex, and we're capable in all kinds of things like, you know, climbing mountains and writing books and doing startups, but our motives for doing all of these things are much more complex. Right?
If anybody of our listeners today would basically write down an equation of what gets them up in the morning and what gets them to be excited about what they're doing, it will not be an equation that is driven purely by how much money am I making in what will be the financial maximizing next step for Pete. And we have lots of other motivations and whatever works for us works for lots of other and therefore, we need to put this into our products.
And, you know, when we talked last and we're thinking about what to talk about today, you kind of asked me to think about the different ways in which behavioral economics can help startups. And I think there are kind of 3 buckets of help. The first one, let's call it tweaks. What do I mean by that? It's basically seeing something in the process that is just slowing people down. And think of it something like somebody is setting up the wrong default. Right?
You look at the sign up flow and you say, hey. You're asking people a question here that you don't really need to ask too difficult. It's too complex. People are dropping off. The story about Morgan donations. Right? You basically have Morgan donation is opt in. Pete have to go to the DMV and sign up for Morgan donation. Very few people sign up. You have it as opt out. People are assumed to be organ donors, and they have to sign something to out of it, almost everybody becomes an organ donor.
The default really matter. You know, many times when we set up products and sign up flows and so on. We don't truly understand how difficult things are for the consumer to And therefore, we set up processes in a way that is too complex or too difficult, and there's a lot of tweaks around that. Now it's not just default, it's a bit more complex than that. For example, if lemonade turns out that when you do a chatbot falls under what we call conversational norm.
You and I when we meet, you know, now with over video, but when we meet face to face, there's a discussion, right? I ask a question to answer you. There's a give and take, and you don't stop in the middle. Right? It'll be very strange if the 2 of us met. And all of a sudden in the middle, I would just spin up and go somewhere. And now this conversational norms gets people to continue responding to questions in a way that they wouldn't do in a regular form. Right?
It just fits on to, like, if you think about the chatbot asking you a question, it fits with this general metaphor of you know, somebody's asking me question. I need to answer. I can't stop in the middle. So there's a whole set of things that we can call them small tweaks. Then there are things that are more complex. And I'll give you one example from a study we did with a large HMO.
And this HMO send people a notification saying, you have a appointment to meet a specialist in 5 days, in 2 days, they send a message twice, please show up on time. And if you're not going to show up, please follow this link and let us know you're not going to make. Before they send this text message, there were lots of people that didn't show up. So they started Beller these text message to remind people. And they thought that people were not showing up because they forgot. So say, hey.
Let's remind people 5 days before, 2 days before, But even with those reminders, more than 20% of the people did not show up on time. Right? And it's very hard to run a Beller care system when 20% of people don't the way, if it's every day, 20% don't show up, it's easy, but changes over time. It's very hard to manage. So they asked us for Beller, and they limited us very much. They said the only thing you can do is change a text message. Right?
You could imagine that you could give people points and reputation and money. Like, the world of motivation is very broad, but they limited us only to the So we try different things. Some Pete, we reminded them that their family wants them to be healthy. Some people, we reminded them the name of their doc Remember, Doctor X is looking forward to seeing you on this day and this time. For some people, we reminded them about the nurse. Remember, nurse y is going to see you.
For some people, we told them how much it would cost the HMO if they don't show up. We reminded people about your own commitment for health. We did all kinds of things. And one other thing we told them was that some other patient could use their appointment if they don't show. So the first thing that happened is that all of our methods worked better than the control condition. Right? So that's good news. Right? Add a source of motivation. Everything worked.
The thing that worked the best was to say another patient could use your appointment. So the feeling that by not showing up, you're basically ruining it for somebody Beller? That's right. By not showing up and not cancelling. By the way, it's perfectly free to cancel. So this is a slightly more complex issue, because now you say, okay, people don't show up. We need to add motivation. And adding motivation is much more complex, and there's many more Right?
And if we were going to sit here for an hour and think about what other motivations could we add in a text message, we could probably come up with more idea. Right? I'm not saying that the right one was that. What was the differences, by the way, between them in terms of kinda give people the feeling of when you get it right? How much better is it compared to control? We cut the no show rate by more than half. That's amazing. That is a huge financial impact on such an organization.
Huge financial impact. I think it was something like 800,000 days of weight a year. That's amazing. That we saved them. By the way, we did something else. We had some people on the team that are much better mathematicians than I am, and we asked the question of what if we used AI?
What if we basically created an algorithm, what what I told you, another patient could use your appointment that worked on average for everybody, but we had a lot of different types of clinics and types of doctors and type of illnesses and so on. And we said, what if we could tailored the specific message for each person, how much better would it Pete? And the results we got are basically kind of the famous eighty 20.
More than 80% of the value from picking one message that was the best for everybody, but we could get some additional value by for individual. Right? So there is kind of, if you ask me, like, how much are people similar? Kind of mostly were the same. The same thing works on everybody, and there's some tweaks on the edges. Interesting. Yes. We see it in many companies that when you get it right, it's 80% of the value.
And then with more personalization, you can add additional percent, but it's basically the big things is understanding the core motivations and the big segmentation. That is a big part of the change. Yeah. So we said, okay. So there are basically three ways, I think, that behavioral economics can help companies. One is we call it the little tweaks, default, sign up flow and so on.
The second one is more complex, which is about adding motivation, right, how do we add motivation to people, somebody else, money, whatever it is. And then there's the 3rd component, which I think is the most interesting, and this is the one that is worthwhile for people to think about earlier in their agenda, and that's about what is the role of behavior economics in the core product or the business Morgan. And it's not true for everybody, but I'll give you maybe 2 examples.
So one example is a company called Shaper, and Shaper, the goal was to help people lose their way. You know, there's lots of ways to get people to lose weight. But if you start thinking about this from a behavioral economics perspective, you say the following. You say, okay. The struggle for health is a daily struggle. And you can't be healthy 5 days of the week and not healthy 2 days of the week with food, right, you can completely overrun everything you do.
By the way, the obesity epidemic in general, it's a 200 calorie a day game. If you can cut 200 calories a day systematic, basically, people lose a pound every 2 weeks. That's incredible. It's just really how to do it on a consistent basis. So people don't. But so we said it needs to be every day, and you need to start in the morning. You can start your diet. Only in the afternoon. So and then we said something like the foreign said, okay.
In psychology, there are basically models of recall and models of recognition. Recall is something that you think about yourself, like you walk around the house and you say, oh, let me think about my long term savings or my diet or my health. It happens occasionally, but not that much. Right? The idea that somebody will come up with a priority that is something about their long term well-being. It just doesn't happen that much.
Recognition happened when you see something and that reminds you of some. And by the way, in the digital era, we don't have that much recognition. Think about pictures those of us who don't have picture frames, how often do we think about our loved ones? Yes. We have all of these pictures on our phones, but they're mostly off. I'm good friends with your wife. Do you have a picture of her on your desk? But I have a digital frame. So the answer is yes.
So, you know, but when we have things in the environment, they remind us of things. By the way, religion, of course, has this right. So we said, okay. So we need first thing in the morning and the reminder. So we said, let's think about the bathroom scale. First thing in the morning, you see it every day. If somebody gives you 2 square feet of their prime real estate, take it. Okay. So we started with the bathroom scale, and then we started the bathroom scale.
I know it sounds ridiculous because, you know, what is there to study, but it turns out there's a lot to study in the bathroom scale. Pete are three things we learn. First, thing, it turns out it's really good to stand on the bathroom scale every morning. Not so good to stand on it at night. Now why? Yes. We weigh more at night, but it's not the reason if you channel is in the morning, you remind yourself that you want to be healthy, and then you eat a little bit less throughout the day.
You're not gonna change a lot in your behavior at night before you go sleep. That's right. If you stand on the scale at night, you just go to bed by the Morgan, you forgot the whole thing. Yeah. That reminder doesn't work for a long time, but if you have it just before breakfast, it's very Okay. That was point number 1. Point number 2, you remember the principle of loss aversion. Loss aversion is the idea that we hate losing money more than we enjoy gaining money.
And with weight, of course, it's the opposite. We hate gaining. We like losing, but because they're not symmetric, imagine somebody who goes up and down in their weight We all fluctuate. Right? It's based on how much salt we had and when we went to the bathroom and so on. Somebody fluctuates up and down, up and down. The ups are really miserable. The downs are slightly happy. But on average, it's not good news. Right? High misery, slight happiness, high misery, slight happiness.
Basically, we don't feel good about it. And the last thing is it turns out Pete expect their bodies to change very quickly. People say I've been on the diet since yesterday morning. Like, where are there? Nothing changed. Nothing changed. I hate my skill. Now the body takes 8 days to 2 weeks to react, but our theory is much faster. And by the way, it's true for lots of medical condition.
So what happened is you go on and die for 4 days, you stand on your scale and your weight fluctuates and went up by half a kilo. People say it's not working. Then they take a day off and eat some cheesecake, and then it goes down a little. So if you think about it, the scale is very confusing and demotivating. By the way, before the digital era, the scales were mechanical, and the needle was thick and we didn't really know our exact weight. Completely. And so you couldn't be as upset.
That's right. So what did we decide to do? We said, let's do a scale with no display, and let's give people the feedback on an app, but in trends of the last 3 weeks, let's not focus on the numbers. The numbers are confusing and demotivating. Let's give people the trend. And now we had another interesting question. Do we do 5 trends or 4 trends? So think about 4. Much worse, the trend is going up a lot, up a little, down a little, down a lot.
That's four points or 5 points, up a lot, up a little, nothing happened slightly down a lot's down. So what do you think? Is it better to have a 5 point scale or a 4 point scale? Is it good to have the middle? I think it's good to have the middle. I think the middle gives you leeway and helps you with motivation. You're absolutely right. And here's the thing. Imagine a year. 52 weeks. And imagine somebody is losing weight 12 weeks and nothing happens for 40 weeks.
Is that a good year or a bad year? That's a great year overall. Amazing year. In my age, a year that nothing bad happened, but this is where it's so important to understand the psychology because if people encode Nothing happened is bad news. A year when you lost weight 12 weeks and nothing happened forty times, you would feel as a failure, but it's an amazing year. And we need to give people credit for nothing bad happened. So we created a 5 point scale, and we celebrate the middle.
Congratulations. Very, very important. Interesting question. What are the words that you're using for the neutral or what are the colors that you're using? Because you wanna make the neutral positive. Right? So we gave the neutral green color. We made the decrease in weight in blue colors, and we made the increase in gray colors. So we separates the green and we give you extra extra points for it.
Now with the first study we did lasted about half a year, And, you know, we had the control condition, people who got the regular scale, we had our condition, and we went to call centers. And why did we pick call centers? We wanted people who are relatively low income, relatively obese, because that's some of the toughest population to change. Right? If you get people who are highly motivated, that's not as interesting.
And the people who got the control scale, the regular scale, a digital scale that sells your weight is 168.2. They gained 0.3% of their body weight every month, gain, gain, gain, gain, for the 6 months of the study. The people who got our scale lost 0.6% of their body weight every month, lost, lost, lost, lost, lost, and the big deal was adherence First, right, we had people who basically stood on the scale, 79% of the people stood up on the scale five times or more per week.
It was not frightening anymore. It was not aversive, and they did it. And then the last thing is that they understood better the relationship between cause and effect. Now here's the thing that came to me at the end of this. You know, a lot of times we think that the right thing to display is accurate information, and I'm all for accurate information, But I think that information does 3 things. It's about accuracy. It's about helping people understand the relationship between cause and effect.
And it's about motivation. And it turns out you can't always have all 3 of them. And for me, as somebody who wants behavioral change, Accuracy is nice to have, but not that important. The first thing I want is motivation. The second thing I want is understanding the relationship between cause and effect. And only the third thing I want is accuracy. I want to create interfaces that motivate first then help people understand cause and effect and then are accurate.
But if those goals conflict, I'm willing to give accuracy. Now the reason I'm talking so much about this, it's not that that scale is that crucial, but it's about the perspective of thinking about all the little details. Right? You look at the scale and you would say, what the boring product? What else is there?
The moment you take seriously human behavior, motivation, understanding, what do we do with variants, right, statistical you are great with dealing with veterans, people are not, then you start designing your product from the beginning, taking these principles into account. These are such amazing examples.
And I think that at the end of the day, the toughest thing that we're trying to always tell startups is that when you temple across principles that are important for your company, it is really critical to think about them early on rather than try to patch them onto your product later. So, you know, we have these thesis about network effects, and we keep telling founders, you know, think about your network effects early on.
Don't wait for when you've already completed your product and then you try to patch something on top. And I wonder how I can take what you just spoke about and translate that and to start up guidelines. So, you know, we say that our tweaks and there are more complex changes. And then clearly, the highest level or the ones where you get the most out of it is when you actually start thinking about behavioral economics when you actually design your core product. You don't wait later.
You don't say, oh, my emails are gonna be using behavioral economics, but my core product is my core product or you don't wait just for the tweaks. And so how would you recommend that people and I know it's a big question, but how have you seen the best founders try when they're starting in it. Let's start the early as possible. When you think about your core product, how do you see the best founders think behavioral economics in designing their main product.
What is the process that they're doing? How should they think about that? Yeah. So I think a really good example from this is lemonade. So I met shy and Daniel before there was lemonade before there was anything. And and they came to talk to me about insurance. And their question was, what are the biggest barriers for insurance? Right? So they didn't say, okay. Here's the product. Help us tweak it.
I did some of those things later on, but in the beginning, it was, what's the barrier for insurance? And I told them trust I said, and here's the situation with insurance. You as a consumer pay the insurance, pay the insurance, pay the insurance. At some point, something bad would happen, and you want them to pay you back. But they have your money, and they write the rules. And the question is, do you trust them? And if you don't trust them, are you going to be dishonest and exaggerate.
And I worked with quite a few US insurance companies, and we saw a lot of fraud. Right? So and people justified fraud because they said these are bad people. They're in conflicts of interest. They don't have my best interest in mind, and I'm going to cheat. People kind of rationalize cheating because the insurance companies are bad because they don't wanna pay. I said it's a really bad situation.
If you started from scratch, you wouldn't design something like but I said you have to understand that not trusting the insurance company creates vicious cycle. And what happened is that people lie to the insurance company. The insurance company spends a lot of money to adjudicate claims because they don't trust Pete. And the whole system is losing lots of money. Like, adjudicating claims is what's called a deadweight loss. Like, everybody loses money. So I said, let's deal with that.
And the way, and we talked about different options and the way we chose to do it, is to eliminate conflicts of interest. And that also was very psychological. Right? There's a version of psychology that says, trust me. I have your best interest in mind. I have done lots of research and conflicts of interest, and I know that it's not enough that when you have people with conflicts of interest, they often don't see things correctly.
So we said, let's change it from a 2 player game to a 3 player game. Let's have the consumer lemonade in a charity. And when people sign up, they pick their favorite charity. And they pay lemonade, and lemonade always takes a fixed amount as our profit. But we keep the money in the pool. So let's say you you you joined lemonade and you sign up and you chose the World Wildlife Fund as your favorite charity.
We take all the people that sign up under the World Wildlife Fund, and you're one pool and you pay money to lemonade. We take the fixed amount. We pay back claims. And if there's money left over in the pool and we design it so that There's very high likelihood there will be money left over in the pool, and that money goes to the worldwide life fund. Now think about what it happens. There are 2 things. One is we say, look, we're indifferent of whether we pay you back your claim or not.
We're indifferent. We don't make a different amount of money. We have a fixed amount. And on top of that, if you are cheating, who are you cheating? The non for profit. Yeah. And of course, you know, there's all kinds of tweaks. We remind people about this, and there's lots of other things going on. But if you think about this, lemonade started. I mean, lemonade is a wonderful product.
There's lots of people working on all kinds of amazing things, data, and so on, but it baked trusts into the business model from the beginning. Right? It says if you want to have an insurance company that, you know, we don't have a 200 year reputation and so on, trust is the main Currier, and we have to make a business model around trust. Yep. And that clearly works Beller for the guys.
And so if I try to generalize it for a second, I mean, lemonade is such a great example because it really was one of the biggest differentiators and what got them to be so loved by their customers.
If we now take a founder that comes up with a new product, then they wanna think, what's the methodology to try and basically incorporates the overall economics thinking into their product or into their company, should they ask themselves what are people's motivated should they ask themselves, where are the incumbents wrong in understanding people? What are the few questions that they can ask themselves that would really help them build a product that is based on these principles.
Yeah. So, you know, creativity is an np complete problem. Right? It's really hard to get a recipe for But in general, I would say the following. If we think about the 3rd component, right, so we have tweaks, we have motivation, and we have the big questions. On the big questions, I think they need to do 2 things. 1 is they need to ask what are they assuming about their users? What is motivating them? What is causing them to want some things versus something else?
And they need to make sure that they're explicit about those assumptions. Because a lot of times when you write these these assumptions, you realize maybe you're not as sure about the assumption as before. When you start writing them down, right. So you can ask the question of, you know, what is causing people to eat large Morgan, or what is causing people to not keep to a diet. You know, what is causing people not to repay debt? Whatever it is, what is causing Pete?
That I think is the first thing. I would not start with what are the competitors are doing. I would basically ask what is my model of humanity as it interacts with my product. I'll give you another example. Paying back debt. Yep. Right. What is causing people not to pay debt on time? So 2 experiments. 1, we said, you know what? Maybe paying that is just not that motivating. When you think about it, you say, yes, people should pay their debt.
It's really important, but maybe it's not that motivating because the end is 0. Right? At the end of paying that, you've got nothing. So we try to frame it differently. So we send letters to Pete. And we said, you already paid 20 percent of your car payment. Would you like this month to pay a little extra and pay more of your loan? 0 interest. To other people, we send a slightly different letter and we say you already own 20% of your car.
Would you like to pay a little extra this month and own your car? A bit more. Yep. A lot of interest. So from the negative to the positive and motivation. Owning. Right? Because if you say what is my motivation to pay my loan faster? 0. What is my motivation to my car? Much higher. Or here is another example. This is a recent paper in which they basically gave people their credit card statement, but they allowed them to pay some of the categories specifically. So imagine you get it online.
And for some people, I say, you know, you owe a $1000, the minimum 50, how much do you want to pay? But for other people, we show the specific things, and you say, which one of those do you want to pay? And you can pick I want to pay for all my food and electricity and so on. And it turns out that people pay a 12% more. Your balance when it's specific things because there are some things we don't feel like they should be on credit.
Completely. And I said on the board of this company that selling e commerce 1,000,000,000 of dollars on credit, and we could actually see a huge difference in the pay ratio for things that were perceived by people thing that should be paid off, like, when they bought things for their kids for school or when they bought the thing that were a must compared to the payback ratio on luxury stuff, which many times defaulted.
Yeah. To go back to the question of what to do is you basically want to write your assumptions. So if you're saying about paying debt, you could say, oh, I'm assuming that people want to pay their debt back. How much? What kind of debt? The moment you get into it, now you can say, oh, you know, I'm assuming that people are torn between paying debt and other needs and what puts the Morgan with this? What is the right way? So, basically, writing down what do you assume?
I think is step number 1. I think step number 2 is kind of like where the rubber meets the road. Basically go very slowly through the experience with your product. And at each point, ask yourself what's the friction and what's the motivation. Right? What is holding people back and what would give them extra motivation? So you get a letter that says you owe this amount of money and you need to pay, what's stopping you and what would motivate you.
And it's not an easy process, but those are the process I think we need to go through. And then the other thing I would say is that there's this belief in Pete out early and iterate. I don't know about your experience, but my experience is that people do get out early, but they don't iterate until it's too late. Many times that, you know, we clearly advocate for going out and iterating fast. Point is that, as you say, many people sometimes make going fast with, okay, we put out something.
Let's wait. And, clearly, without data James faster, and there's no point of going out fast. That's right. And often they permits are not that good, so it's harder to learn. You would think, at least, I thought the electronic world would give us amazing platforms for testing The reality is the testing is tough. It's hard to know how to create the test correctly. It's hard to run test. Sometimes it's require extra work. In a lot of startups, I find don't do enough testing.
They basically are so driven to get out quickly. The the slowness that is required by testing something is just very unappealing. So I think we don't test enough. And, you know, sometimes by the time people come to ask me for advice, it's it's a shame because I find that we could have done some things much, much Currier. So that's a great James.
We're we're basically saying before you design your product, you've gotta write down, and test your assumptions on what motivates your customers in the behavior in the areas where you're interested. And then, basically, you gotta take your product and, as you said, test it stage by stage and find out and make sure that you're addressing the right motivations in each step of the product. Yep. Fit like a microscope. You want to go over it incredibly slowly.
And at each point, you ask yourself two questions. What is holding people back and what would motivate them. Right? So what would get me to sign up and what's causing me not to sign up? What would Pete me to, you know, part with my money and what would get me To walk out of the store.
Yep. You're famous for speaking about rationality or let call And, you know, one of the things that I found that many founders have a tough time with is that and everybody, actually, is that we kinda so many times that the motivations of our customers are rational. And I think that even sometimes even people that try to follow structure of understanding motivations, they just get it wrong because many times they're not understanding the real motivations of people.
They assume that what should work works. And then they're kinda shocked later on to find out that people are not rational or that, textbook, hey. You press here. You make more money. Doesn't work because there's additional elements that are impacting people's behavior. How do you suggest that people manage to understand irrationality and kinda better really understand what are the motivations when they're not trivial.
Yeah. And that's very tough, right, stuff because you know, I've been, you know, reading this literature and contributing some to it for about 30 years. And the reality is that it's really tough for people who are not coming from social science to master this field.
It's always kind of baffling for me how, you know, nobody would ever do a startup in biology without having a biologist, but people have no problem doing a startup on behavioral change without somebody who knows anything about behavior change. It is kind of surprising, but I think people do need to get some expertise on that. And of course, everybody can do it. It's just a question of time. We all have access to libraries and books.
And so on, but the reality is that, you know, startups are usually created by people who are amazing at technology, not mastered the literature on behavioral change, even if they do behavioral change. And there's just not enough time that they have to master it and they have to help get help. It's just not the question. I think it's something that people have to reduce and say differently. I think there's a lot of areas where startups benefit a lot from overconfidence.
And I think that overconfidence about an having intuition about human behavior is not a healthy one. So when you come to present to VCs, I mean, there's lots of things where overconfidence is helpful. Being overconfident about your own intuitions about human behavior, I think, is very unhealthy, and people should go and find somebody on, get some PhD students in some area of behavioral that is relevant to what you're doing. Get them to help you for a couple of months.
You know, you don't have to hire a full time person, but getting some help on that is crucial because human behavior is really mysterious. For example, I've been doing this study for the last 4 years now trying to understand how companies treat their employees, how the employees feel about the company and what it means for the profitability of the company in the stock market. I'm doing it on public companies in the US. So for example, do you think that salary matters for profitability?
The companies that pay employees better, the employees feel that they are getting paid better is are those companies more profitable? Yes or no? Well, the way you were it. I assume that there is some link, but not very significant. Yep. So now, however, perception of fairness in salary matters a great deal. So absolute salary doesn't matter. Quality of coffee doesn't matter. What really matters is people feel that they can make mistakes honest mistakes are valued.
Virocracy, by the way, kills motivation. I mean, there's all kinds of things that quality of coffee doesn't matter. Quality of furniture doesn't matter. Employee benefits doesn't matter. Appreciation matters dramatically. I studied 80 different components, but when I go even to HR and I say, what do you think matters and doesn't matter? The wrong, right, our intuition about what matters is just not that good.
And this is why people dig deep into their intuition, they're not necessarily going to get it right because some of the things that motivate us are very different. And it's also very hard to intuit that when you're not in that situation. So we have something called the hot called empathy gap. And the hot called empathy gap is the idea that when you're in a cold rational state, you don't really understand how you would be when you're in the hot state.
And, of course, the easiest example to think about is sexual arousal. Right? When you're not actually aroused, you said to yourself, I always respect women. I'm using it as a man now. I always respect women always wear a condom, always did this. When people are aroused, they change, but it's not just about such or when you think about things outside of that situation, it's really hard to intuit what it would mean to be in that situation.
For example, if you're not looking for a loan right now, you don't truly understand. I remember when I moved, I moved to Duke a few years ago, like, quite a few years ago, when I came here, I, you know, I have my own health issues. So I go to doctors a bit more than I would have liked to, but in the beginning, I hated this thing that the doctor come to you and say, hi, Dan. I'm doctor or whatever. And I wanted to say, Professor Rielly, for you. Right?
Like, why did I'm Dan and their doctor number? I was planning on saying it every time. But the moment you're in this robe and your behind is picking up, you are this small. And in that situation, the power is in their hand, and I wanted to say it, but I couldn't bring myself to it. And after a few years, they all know me. They stopped, but it is incredible how it's hard to predict how it would in a doctor's office. By the way, here's something we did with doctors.
Doctors sometimes ask you at the end of the visit. Do you have any questions? People are always so embarrassed to ask. They don't ask questions. We change it to the doctor saying, I know you have lots of questions. Can you ask me your top 3? You're different. And the moment you understand the power dynamics, and how difficult the power dynamic is. Now right now, I'm sitting here next day. Oh, yes. I'll go and ask questions next time.
No. You feel this Morgan you feel you're taking the time and during the rush and you feel like you're taking the time from another patient and you don't ask questions. Completely. Thing for me is that many times we look at the founders and we say, can these founders really understand the customers and many times we call it the founder market Flint, meaning that you know, that's a founder that can actually understand the motivation of the customers.
But even when that's working, we don't always end up with a founder being in the same state of mind as the customer is at the point decision. And so I think many founders, many time come to us, and, you know, they say we've asked customers. We've had surveys. We've had studies. And, you know, how do you feel about these studies as ways of discovering people's motivations?
Because I always, you know, when I look at them many times, It's very clear that when the people were asked, they were not really in the same state of mind. Even if they're the right customers, they're not in the same state of mind. For them to be able to predict what they really want or what they really need. How do you think about that? Okay. So for that, we need a whole another session at some point because it's such a big question.
But in general, there are things that people can report and things that people can't. I'm quite comfortable asking people to report what they have done. I can ask you, for example, did you check social media yesterday? Right? Would you be aware of the exactly how much time? Probably not. But can you tell me if you checked it yesterday or last night or before you go to sleep? The answer is yes. People can report quite well on what they have done with some limitations.
I would be very hesitant to ask people for their intuitions about it. And here's just an example. It turns out that retirement savings, 401 k in the US, is basically driven by defaults. If when you come to a new job, they default you into 401 k. You save. If they don't, you don't. Or, you know, big differences. Some people do, but huge difference But if you go to people and you ask them why did you save and why you didn't save, nobody says it's a default. Right?
People have story By the way, the stories are fantastic. Right? If you wanted the right fiction, go for it. But people can tell you. I'm saving. I'm not saving. But when you ask them for their, why devatology difference perception of what really happened. They have wonderful stories. And basically, you know, the lots of times we act as a function of the environment that we're in, but when we tell our self a story, the story is not about the environment.
The story is about something intrinsic in us, and we make this story so quickly that we start believing in it. Right? So it's not that people lie. They actually believe in it. They have no idea. And on the Pete, they're trying to find out the story And the story that they find is a story about intrinsic attribution. I did this because I wanted to, and then you just tell ops something. I would not trust that. Now there's some exceptions. Right?
There's some Pete. Like, if you want to understand why some people are able to control their a one c or take meal time insulin, know, people who have to take meal time insulin, it's quite embarrassing to go with coworkers to lunch and basically take a needle out and inject yourself. Some people have managed to get out of It's really good to take the people who don't feel embarrassed anymore and say, what have you done? How did you introduce this to your coworkers?
How did you get them to explain it? How comfortable do you feel? Like, you can learn from those things, but I wouldn't ask people for reason. I would first go to the literature. Right? I mean, it's kind of also amazing, you know, how quickly people are to experiment and how slowly people are to go and read something in the library. Oh, that is so true.
We had a few mistakes that are very clear, like, you know, not iterating fast enough for getting wrong sources for what are people's motivation and not using experts that actually help you. Other additional kinda typical mistakes that you see that companies got wrong that they could learn from behavior Economics and actually do Beller? So there's a huge range of things in pricing. Right? Pricing is a magical amazing kingdom, and people make lots of mistakes there, but I think that's one.
I think a second big mistake is that people don't work enough on their user's mental model of their product. I'll give you one example for this. I went to visit Microsoft a couple of years ago, and I met the people who were working on Morgan. And they told me that people use Cortana for the weather and jokes, and that's it. And I asked them what can Cortana do? And you can imagine what they said, everything. I said, then you're not setting up Cortana correctly. And they said, what do you mean?
And I basically created a foreign study. Some people, we said, hey. There's something called Morgan. Ask it anything you want. And people ask for weather and jokes. Other people, we said, There's Cortana for meetings. There's Cortana for entertainment. There's Cortana for transportation. There's Cortana Morgan we gave them 10 categories. Ask it anything you want. Now people had lots of questions. Right? They would say show time. They would say cancel my appointment. When is my meeting?
I mean, they had all kinds of things. And the thing is that people don't have a mental model of what can Morgan do. This, by the way, also fits with our notion of everything digital is hidden, and we don't know what to ask. Right? And one of the challenges is if you do something new, what do people understand about what is this new thing? Now this is called the curse of knowledge because we know our product so well. We think that it's clear. Yeah. Of course.
This by the way is what happened to professors. Right? We teach. And then after a while, everything we teach seems obvious Right? So teach worse and worse and worse because we just assume more and more on the side of the people listening. So I would say that not helping Pete create a better mental model of the product and its use. I think that's another one. That's a big mistake.
Yes. I think another one is not thinking clearly about what is the pattern of use we want to create and how do we reinforce habits? So, you know, people have measures like daily active or weekly active or whatever. But the real question is, are people using the product at the frequency and timing that we want them And what is the model for repeated use that we have? And how do we create that?
And people tend to default for the looking at the dry KPIs and not thinking about what's really happening behind them. That's right. So, you know, is this, for example, a product that would be very useful to try and get people into a habit of checking every Monday morning? Or Sunday or Friday. What is the right cadence and what do we create in the environment to become a trigger for that? So there's really a range No. It's amazing, but what I wanted, and it's just you've done it Beller.
It's just give people the kind of the perspective of what are the things they should think of. I wanna ask you one word. Started touching before and something really I sometimes ask people to think that they are designing products for homosims. And I don't mean it in an offensive way. I mean, offensive to think about people as Homer Simpson, but I don't mean it as people are idiots, but I mean that people are have busy complex lives.
And therefore, we need to realize that the time that they're giving us is very, very limited. People have very, very limited attention and time. And when we design products, we often think that people care about the things we care about. Right? Pete, I have this new product. People would probably read the manual and, want to lose weight and people after all want to control their diabetes. But the reality is that people have very little time and very little attention.
And if we design with that in mind, I think we'll get much more progress. Yes. That's a perfect segue because for me, the kind of last question I have, which is really interesting for me, both as the next founder, but also as somebody that helps and supports founders today. And he started touching upon it before. It's what do you see from the behavioral economic perspective on the mindset of the founder himself other than, you know, not thinking that you know what you don't know.
But what do you think is the mindset that's required, or what's the behavior economics elements that we can take into the actual behavior of the founders themselves if they wanna become the most effective startup leaders how do they get from being, you know, the dreamers of an idea to being this great execution leaders in their company? What can they learn from the overall economy? To become more successful at that job. So, yeah, I would separate it into treating people and treating the product.
And I think on the product, you can't test everything. Like, you know, scientists want to test everything possible. I think good leaders are able to separate core assumptions of the product and figure out what to test and what to make assumptions about, and they can sequence it the right way. They can say, you know, what's the right building blocks and what to test when? That would be kind of number 1 and 2 for the product.
Here, the central assumptions we're making, I'm going to test this assumption first and then this, and then we'll move to smaller things that are less Morgan. Right? You can say, for example, in Lemonet's case, It was first. Let's test trust. Let's test the charity model. Later on, we can test things like pricing and whether we could get people to like lemonade more answer more question and so on. And then internal to the company, it's about being able to create motivation.
The big lessons we have for motivation is that appreciation is incredibly important. I would say that appreciation, fairness, and tolerance to mistakes. Are the 3 elements that Pete startup leaders to be the most effective in leading their companies. The thing about tolerating mistakes is about agents Right? It's about giving people the ability to move forward in the way that they think is right and allowing them to make mistakes as well. Right?
So you want people to move forward and when they're trying something new and interesting, whether it's working or not working, you want to reward them for it. You know, all companies basically when you talk to them, they want people to take more risk, right, and try to innovate. Well, they think they do. Or at least they say they do. They don't necessarily behave as if they do. Exactly. Now the question is how do you also get people to behave this way and not just to say it?
Yeah. And I think that, you know, it's such a good point because at the end of the day, I think one of the biggest as we've seen between a successful and non successful startup ecosystems is the acceptance of failure. Those where there's no failures or failures are not acceptable, then you guys less, you know, a lot less risk taking and a lot less entrepreneurship.
And similarly, probably the biggest difference between large companies and startups and the reason why startups still win is because in the large companies, you know, your best way to get your bonus and get promoted is not make any mistakes, which is clearly the most important you know, being willing to accept mistakes is the number one thing in terms of, encouraging innovation simply because most innovation James. Yep. You know, it's not about accepting stupid mistakes. Right?
It's about rewarding the process versus outcome. So, you know, academics. Right? We do studies. Not all studies work as we wanted them to. Right? Sometimes we're wrong. I try to encourage my team to be wrong at least 50% of the time. If you're right too often, it means you haven't tried things that are sufficiently interesting, right, because it's very easy to do studies that always work as you expected if you are assuming very boring things.
Completely. And then when Morgan larger companies, when people came to tell me about a failure, I had these two buckets of failures. They had good failures and bad failures. The good ones were those where people were pushing the boundaries and trying hard and working, but working properly and doing everything that they should have done. And they were Pete bugs or stupid mistakes. And at the end of day, it didn't work like many innovations.
And then bad ones are the ones where, you basically didn't write the code Beller, or you didn't test it. And so you put out a buggy product, whether it shouldn't be now, or you basically completely didn't follow the process, and these are bad mistakes. And the way I used to look at it is that the people that had great failures, they're the ones that got promoted as if they had successes because you could see that they were doing the right things. And, eventually, they're gonna get it right.
And the people that had the bad failures, they're the one that, you know, we could never accept simply because that's bad for the company. And you want to separate those. And one way to frame this which I try to do in my academic setting is to reward process versus outcome. So imagine you have a kid and, one day your kid studied really well for the exam, but the exam asked about some, you know, side issues, and your kid did not get a good grade.
And then another kid didn't study at all, but went to the exam and the exam asked him the one thing he knew and he asked it. Do you want to punish the 1st kid and reward the second? The answer is no. Because in the long term, what the first kid did is going to be the right thing and what the second kid did is not. So what you want to do in reality is we want to reward behavior, not outcome. And we often reward outcome and not behavior because it's easy, not because it's the right thing to do.
Completely. We sit in politics many James, right, and how do you respond to terror attempts versus successes and so on? It's many times most people deal with the outcome rather than with what was the intention. And I think this is a real benefit of startups as companies get big and bureaucratic they set up procedures and procedures are, you know, inherently trying to get things that are measured and, you know, so on. Startups are much more flexible.
And I think one of the benefit is to use the size of the startup to truly understand what people are are doing, not what the outcome is, and to reward correctly the right people. Yes. That is the benefit because you don't need to stick. You don't have the liability of past products or past process or past things, you can always go to the new thing. I wanna ask you one last question, a personal one. And, you know, it's something that I've always wanted to ask you. You're clear.
You're such a wealth of information. And, you know, you you seem to understand people better than most. Well, clearly better than most people I know. And kinda wonder though was after all these years and all that knowledge you have from behavioral economics about people's behaviors, how does that change your behavior, your interaction with people, your life, because it seems that you have, like, this ability to understand people so much better How does that display itself in everyday life?
So lots of ways. So, you know, first of all, you know, just to be clear, I do make some mistakes as well. You think about 3 types of mistakes, you think about little decisions, big decisions, and habits. Little decisions is going to the supermarket, and something is on sale and, you know, still with the Beller price, you wouldn't have bought it if it wasn't on sale, at the same price. Like, it was 20. Now 15. You wouldn't have bought it for 15, but now you buy. I still do this sometimes. Right?
So that's But assume you laugh at yourself when you do it. Right? Sometimes I notice it. Sometimes I don't. It's amusing, and that's fine. I am much more careful with big decisions. Like, if you think about big decisions, buying a house, investing in a startup, getting married, getting divorced, you know, those are the big things. I do understand better the motives, and then I do protect myself against mistakes.
And then the third thing is habits, and I spend much more time on my own habits because I know that habits are basically a way to take a lot of small decisions and make them better. And in total, have a big impact. Right? So if I buy the small coffee on the medium coffee size, I don't care so much. But when it comes to habits, that's incredibly important. And then on the professional side, I'll say the following thing. I look at the world and there's lots of sad things.
If you look at the world in general and there's lots of sadness, but what gives me optimism is that we are designing the world. And my sense is that if we design the world better, we would have better outcomes. So think about the physical world. We have planes and chairs and air conditioning. We really have kind of a fragile body physically, but with technology, we could do amazing stuff, cover great distances, be called resistant, all kinds of things like that.
Intermental world, I think we're kind of maybe even more helpless than in the physical world. We have a hard time thinking carefully, making the right to decisions. We don't pick right. We get confused. All kinds of things. Now in the physical world, it's very easy. I'm on the standing desk since we started, you know, I I need to sit down. You know, you need to sit down and, you know, I'll sit down Pete. We understand our physical discomfort and we build around it. Our mental inability.
We don't see it as clearly. But the good news is that I think technology being all around us means that we can have decision aids that follow us. In the old day, if you have a computer that gave you, a decision aid and then you went to the supermarket, you forgot what you decided to do. Now we have phones that travel with us. We have technology everywhere. I think they're really way to improve. I look at the sadness around the world.
Our inability to manage our money, our health, the health of the planet. Hate, I mean, all of those things, I think we could build tools to overcome those. We're not there yet, and we're just starting. My optimism come from technology. The technology is going to be pervasive and everywhere. And we can create tools that would get us to be very different. Like, if you look at where we should have been and where we are, there's a huge gap. The way to fix it is not to change people.
It's to change the interface between people in the environment in the same way that we build better chairs. People are able to sit longer. I think if we build better Money management tool, people would Management money better. Yes. And better health control, health motivation. That's a very positive note and the one that I believe in as well. And I think that's a great note to wrap on.
You know, this has really been fascinating, you know, every time we speak, I know that we can speak for many more hours. And, you know, you mentioned before that we may need another session. I may hold you up to it because I think this is incredibly useful for everybody listening.
And for those of you that didn't, I truly recommend reading Dan's books because he's done a phenomenal job at taking this very complex science and turning this into something that every startup founder and every person can understand and learn from and build better products and better companies and become better leaders I wanna thank you so much for being with us and wish you that, you know, your prophecy technology is gonna help make the world a better place will actually take place. Thank you.
Lovely to spend some more time with you until next time. Thanks, Dan. You've been listening to the NFX podcast. You can rate and review this show on Apple Podcasts, and you can subscribe to the NFX podcast on Apple Podcasts, Spotify, Google podcasts, or wherever you get your favorite podcasts. For more information on building iconic technology companies, visit NFX, dotcom.