Ethan Evans’ Post

At Amazon, Jeff Bezos told us to "not compromise for the sake of social cohesion." I saw him tolerate lengthy, exhausting arguments as a result. When I was a director, my VP constantly argued with Jeff and refused to back down. Their main, recurring argument was about an arcane economic model called the Prime Attribution Model (PAM). PAM mattered because it determined how the fee paid by Prime members was divided between business units. Prime Gaming (our business) was getting almost nothing, and we disagreed with how the model was designed. During one chapter of the debate between Jeff and my VP, the discussion went on for so long that the CFO stepped in and asked, "Jeff, do you want me to end this?" Basically, he was offering to scold my VP and shut him up so that Jeff would not have to be the bad guy. Jeff explicitly said no. The debate was drawn out and exhausting, but Jeff clearly showed his preference for a strong, lengthy debate instead of a comfortable, convenient compromise. In the end, we never got the PAM model changed. But, Jeff never shut us down. He continued to entertain the debate, and we kept revisiting it every time we had new evidence for a change. It was a years-long debate that never changed anything, but not shutting it down allowed every piece of each argument to be brought to light. The result was that a fully informed decision could be made, even if it was one that my VP and I disagreed with. Now, there is a point where rehashing old debates becomes impractical if it is simply perseverating on things that have already been decided. This is why we would reintroduce the debate when we had new evidence, not just whenever we felt like it. In between rounds of debate, we would “Disagree and Commit” (an Amazon Leadership Principle). We never felt like the PAM model was a good model, but we didn’t pout about it as we continuously failed to change it. We reintroduced our argument when we had something new to say, but we otherwise committed to its use. This is where Jeff’s advice of “Don’t compromise for the sake of social cohesion” and the LP of “Disagree and Commit” come together. The time for debate is when you have evidence; the time for commitment is when that evidence has been thoroughly heard, analyzed, and accepted or rejected. The result is a culture where good ideas are not missed because people feel like they can’t argue on their behalf and things also aren’t stuck in holding patterns when key stakeholders inevitably disagree. Readers- How has this worked in your organizations? Is debate encouraged or do people often compromise for the sake of cohesion?

At Amazon, Jeff Bezos told us to "not compromise for the sake of social cohesion." I saw him tolerate lengthy, exhausting arguments as a result. When I was a director, my VP constantly argued with Jeff and refused to back down. Their main, recurring argument was about an arcane economic model called the Prime Attribution Model (PAM). PAM mattered because it determined how the fee paid by Prime members was divided between business units. Prime Gaming (our business) was getting almost nothing, and we disagreed with how the model was designed. During one chapter of the debate between Jeff and my VP, the discussion went on for so long that the CFO stepped in and asked, "Jeff, do you want me to end this?" Basically, he was offering to scold my VP and shut him up so that Jeff would not have to be the bad guy. Jeff explicitly said no. The debate was drawn out and exhausting, but Jeff clearly showed his preference for a strong, lengthy debate instead of a comfortable, convenient compromise. In the end, we never got the PAM model changed. But, Jeff never shut us down. He continued to entertain the debate, and we kept revisiting it every time we had new evidence for a change. It was a years-long debate that never changed anything, but not shutting it down allowed every piece of each argument to be brought to light. The result was that a fully informed decision could be made, even if it was one that my VP and I disagreed with. Now, there is a point where rehashing old debates becomes impractical if it is simply perseverating on things that have already been decided. This is why we would reintroduce the debate when we had new evidence, not just whenever we felt like it. In between rounds of debate, we would “Disagree and Commit” (an Amazon Leadership Principle). We never felt like the PAM model was a good model, but we didn’t pout about it as we continuously failed to change it. We reintroduced our argument when we had something new to say, but we otherwise committed to its use. This is where Jeff’s advice of “Don’t compromise for the sake of social cohesion” and the LP of “Disagree and Commit” come together. The time for debate is when you have evidence; the time for commitment is when that evidence has been thoroughly heard, analyzed, and accepted or rejected. The result is a culture where good ideas are not missed because people feel like they can’t argue on their behalf and things also aren’t stuck in holding patterns when key stakeholders inevitably disagree. Readers- How has this worked in your organizations? Is debate encouraged or do people often compromise for the sake of cohesion?

At Amazon, Jeff Bezos told us to "not compromise for the sake of social cohesion." I saw him tolerate lengthy, exhausting arguments as a result. When I was a director, my VP constantly argued with Jeff and refused to back down. Their main, recurring argument was about an arcane economic model called the Prime Attribution Model (PAM). PAM mattered because it determined how the fee paid by Prime members was divided between business units. Prime Gaming (our business) was getting almost nothing, and we disagreed with how the model was designed. During one chapter of the debate between Jeff and my VP, the discussion went on for so long that the CFO stepped in and asked, "Jeff, do you want me to end this?" Basically, he was offering to scold my VP and shut him up so that Jeff would not have to be the bad guy. Jeff explicitly said no. The debate was drawn out and exhausting, but Jeff clearly showed his preference for a strong, lengthy debate instead of a comfortable, convenient compromise. In the end, we never got the PAM model changed. But, Jeff never shut us down. He continued to entertain the debate, and we kept revisiting it every time we had new evidence for a change. It was a years-long debate that never changed anything, but not shutting it down allowed every piece of each argument to be brought to light. The result was that a fully informed decision could be made, even if it was one that my VP and I disagreed with. Now, there is a point where rehashing old debates becomes impractical if it is simply perseverating on things that have already been decided. This is why we would reintroduce the debate when we had new evidence, not just whenever we felt like it. In between rounds of debate, we would “Disagree and Commit” (an Amazon Leadership Principle). We never felt like the PAM model was a good model, but we didn’t pout about it as we continuously failed to change it. We reintroduced our argument when we had something new to say, but we otherwise committed to its use. This is where Jeff’s advice of “Don’t compromise for the sake of social cohesion” and the LP of “Disagree and Commit” come together. The time for debate is when you have evidence; the time for commitment is when that evidence has been thoroughly heard, analyzed, and accepted or rejected. The result is a culture where good ideas are not missed because people feel like they can’t argue on their behalf and things also aren’t stuck in holding patterns when key stakeholders inevitably disagree. Readers- How has this worked in your organizations? Is debate encouraged or do people often compromise for the sake of cohesion?

At Amazon, Jeff Bezos told us to "not compromise for the sake of social cohesion." I saw him tolerate lengthy, exhausting arguments as a result. When I was a director, my VP constantly argued with Jeff and refused to back down. Their main, recurring argument was about an arcane economic model called the Prime Attribution Model (PAM). PAM mattered because it determined how the fee paid by Prime members was divided between business units. Prime Gaming (our business) was getting almost nothing, and we disagreed with how the model was designed. During one chapter of the debate between Jeff and my VP, the discussion went on for so long that the CFO stepped in and asked, "Jeff, do you want me to end this?" Basically, he was offering to scold my VP and shut him up so that Jeff would not have to be the bad guy. Jeff explicitly said no. The debate was drawn out and exhausting, but Jeff clearly showed his preference for a strong, lengthy debate instead of a comfortable, convenient compromise. In the end, we never got the PAM model changed. But, Jeff never shut us down. He continued to entertain the debate, and we kept revisiting it every time we had new evidence for a change. It was a years-long debate that never changed anything, but not shutting it down allowed every piece of each argument to be brought to light. The result was that a fully informed decision could be made, even if it was one that my VP and I disagreed with. Now, there is a point where rehashing old debates becomes impractical if it is simply perseverating on things that have already been decided. This is why we would reintroduce the debate when we had new evidence, not just whenever we felt like it. In between rounds of debate, we would “Disagree and Commit” (an Amazon Leadership Principle). We never felt like the PAM model was a good model, but we didn’t pout about it as we continuously failed to change it. We reintroduced our argument when we had something new to say, but we otherwise committed to its use. This is where Jeff’s advice of “Don’t compromise for the sake of social cohesion” and the LP of “Disagree and Commit” come together. The time for debate is when you have evidence; the time for commitment is when that evidence has been thoroughly heard, analyzed, and accepted or rejected. The result is a culture where good ideas are not missed because people feel like they can’t argue on their behalf and things also aren’t stuck in holding patterns when key stakeholders inevitably disagree. Readers- How has this worked in your organizations? Is debate encouraged or do people often compromise for the sake of cohesion?

Founder Mode: Jeff Bezos once threatened to spend every penny Amazon had on a Kindle project when his CFO tried to reign him in. CFO: "Jeff, we only have so much money." Jeff: "How much is that? I might want to spend it." I think the big benefit of being the founder is that you can risk the company in the way a CEO normally cannot. Back when this happened, around 2011, Jeff owned about 19% of Amazon, which meant for practical purposes no one could remove him as CEO. Thus, he could take any risk he chose in a way that an appointed CEO cannot. Jeff also audited deep into the organization at times. A new college graduate engineer on my team emailed Jeff about an idea. Jeff reviewed it with him and then assigned senior leaders, who were against it, to build it. At the same time, there are plenty of so-called manager mode CEOs and companies who do very well. Warren Buffet did not even found Berkshire Hathaway - he bought it. Larry and Sergey famously brought in Eric Schmidt. Starbucks was small time until Howard Schultz got there, 11 years after it was founded. Right now, everyone is all about "Founder Mode" after Brian Chesky's talk at Y Combinator and Paul Graham's blog about it. I've seen founders who were disasters and completely could not scale. They needed help. If they got it, their companies sometimes did well (it depended on who they hired). If they did not, those founders failed. It is great to point to the amazing founders who led their companies this way. I've had the chance to work with Jeff Bezos and to interview with Elon Musk. This much is clear - they are both smarter and more driven than most us will ever be. It is an oversimplification to say "Founder Mode" is only way or the best. First, there are counterexamples of "professional management" working out extremely well. Second, many founders who have a great initial idea do not have the "complete package" to lead like Jobs, Bezos, or Musk., What I do agree with strongly is that a leader must be willing to cut through levels and talk to whoever they need to in the organization. There is a difference between doing this to the point where other leaders cannot run their own areas though. The Bezos idea I think split the difference well was "Single Threaded Owners." This Amazon process tries to make sure that key projects have a single "owner" with founder-like ability to make decisions for their product. They have all the resources to execute. They are also asked and expected to act like "owners" in the sense of treating it as their own mini company. The system is not perfect, but it allowed Amazon to scale to millions of employees while still moving quickly. I hear that some of this agility has finally been lost as Amazon enters its 31st year. Even if this is true, the model worked a very long time. In summary, the best "mode" is the one YOU are capable of executing. Readers - founder mode? Professional managers? Single Threaded Leaders? Something else?