Ep 75: Leadership Lessons with Intuit Founder & Silicon Valley Legend Scott Cook

One key thing is they call anything that isn't exactly the way it should be a defect, and then they send the organization to discover the root causes of those defects and fix them. But by labeling any deviation from perfect a defect, giving you a word, people are motivated to go, we're going to eliminate the causes of those defects. So that's why they have that andon cord on production lines, so they can stop the line immediately when any worker detects a defect. They can stop the line and then get to root cause immediately. It's more important to get to the root cause than to keep going fast. If you keep going, the root cause is unchanged and you just build more cars with that defect in it.

Oh, the knob would always be ringing. No, I have a wife, so I get that immediate attention, immediate recognition of defects. She's very good at this. That's true, that's true.

And you've seen a lot of startup founders over the years. Is that something that people very often get wrong? Is that hard to do?

One of the things you mention a lot of people are learning over time: what techniques or tactics can founders use to grow their skills with their companies?

One phenomenon we've seen play out again and again is as the company gets bigger, maintaining its culture of innovation and disruption seems to get a lot harder. How do you mentor CEOs to foster a culture of innovation within their teams, especially as the companies grow?

Okay, we have a good list to work on. Yes, the culture moves over time to be more operational, less risk-taking, less agile, and mistakes are expensive. Well, that's going to be self-fulfilling. As long as mistakes are viewed as expensive, then the culture is going to naturally become very risk-averse because people are going to save their butt, and thus your rate of innovation is going to slow way down. So I think one way to attack that root cause is by making the cost of mistakes much lower. And a typical way to do that, and it varies by industry and may or may not work in what you do, is to move to an experimentation culture where the organization learns to run fast, cheap experiments with a small blast radius. So then you can try more ideas because they're cheap. You can try them with less risk because the blast radius is really small, or you can actually de-risk before you run the experiment. So moving from a culture of decision by hierarchy or opinion to a culture of decision by experiment. And there's a whole bunch of experimental methods. The Lean Startup was the book that made this famous, but there's much more on top of that. We actually had a story in our background. We had Eric Ries, the author of The Lean Startup, before the book came out, come and speak to us a couple of times. In the second of those, we had him do speed dating coaching. We had four teams queued up for 15-minute coaching sessions. A team would come up, present their quandary in five minutes, Eric would have ten minutes to coach them. And I remember one of the teams said, hey, we're in the payroll division of Intuit. We do employee payroll. Through some evidence, we figured out that like half of the new customers who come to our website wanting to sign up want to do payroll that day, but you can't do that in payroll. In order to run a payroll, you've got to get all this history, learn the employees, their prior deductions, you get all this stuff, and it takes two or three days to get up and running to run a payroll. So we think we should figure out a way to run payroll the same day, right, checks the same day, and then correct them later after we do setup. But management just said no. In fact, we came from a startup that created some payroll businesses and we bought some others, and even our startup leadership said no, that's not how payroll works. We got it in and said no, that's not how payroll works. But Intuit said we'll run a survey to find out what customers think. So we ran a survey of existing payroll customers asking, would you have done paychecks on day one or did you want to set up first? And like 85% said no, setup first. Of course, this is payroll. But in our bones, we just feel we're missing an opportunity. So Eric said, well, this is easy. In a startup, you'd run the experiment tomorrow. You'd put up a dry test in your signup flow, giving people two options: the new customers, do you want to paychecks today and set up later, or set up now and paychecks later? And in the first option, you have to explain that they may have to issue corrections because after setup, those paychecks may need to be revised. And then you'll find out what people want. And so he said, 24 hours you should have this up. He said, okay, you're a big company, I'll give you 48 hours. Well, within 48 hours, the team did have that dry test page up in the signup flow. They directed a percentage of traffic there. Now, they hadn't built any code. They'd only built one screen to give the choice, and that made it a real experiment. And by God, they got something like 45% of people hit paycheck same day. With that experimental proof, then everyone in leadership said, yes, of course, let's build this. We'll take the months to redo our engines because the engines were never built for this, but it's now worth it. We can see. And that change then later rolled out, and they didn't roll it out all at once. They rolled it out in small sections first. That produced the biggest gain in new customer conversion of anything they'd tried in years. So that's an example of running, in this case, a very cheap, very fast experiment. And there's a bunch of techniques to do this. That's only one. It doesn't always depend on your scenario. But I think the biggest thing is if the mistakes are expensive, you will become risk-averse. You've got to get mistakes to be cheap whenever possible. And it's not always possible, but it is in many cases. Yeah, go to the decision by experiment.

Got it. So how do you talk about failure with a CEO as you mentor?

Do you have these knobs in your personal life at all for defects? I think you can apply otherwise.

No, I love when you talk about failure with CEOs. It's a key thing. You know, I think my favorite way to start a board meeting is to have the highlights and the low lights, like what's not working. Let's right away acknowledge and talk about it. This is really important because any board will know if you're just giving them the good stuff. The board will know you're telling them only half the story, and they know you're not being entirely honest. And you want to be known as being entirely straight with your board. So I think what you're talking about is a simple and brilliant technique: start with what's broken too. And you know, so finally, what's the piece of advice, Scott, that you end up telling every CEO you mentor, and why do you think it's so crucial?

Well, Scott, thank you so much.