
Stop telling your teams to coordinate. According to Jeff Bezos, two teams told to work together usually deliver nothing.
That’s one lesson from my interview with Ethan Evans, a retired Amazon VP who now coaches executives through his company, Level Up. Ethan joined Amazon in 2005 when it had 14,000 employees and left when it had 1.4 million. He watched Jeff Bezos fight to keep a giant company moving at startup speed, and saw the company eventually lose that fight.
I wanted to talk to Ethan because anyone building or rebuilding an organization runs into antagonists: forces that stall progress. Ethan kept returning to the ones that got in Amazon’s way.
I hope you can enjoy the whole interview—Ethan is a wealth of wisdom. But if you only have 60 seconds, you can take these three points into your next leadership meeting:
Decide with 70% of the information. Waiting for 90% is how you lose.
Bureaucracy wins by exhaustion. Escalate after two rounds, not ten.
Two is greater than zero. Fund overlapping teams and remove the weaker one later.
In this two-part interview:
Part 1: What Bezos was really fighting, and why bureaucracy wins by exhaustion
Part 2: Jeff’s perpetual antagonist, what leaders reward without realizing it, and Ethan’s own antagonists
Every great story has a great antagonist. That’s the premise behind Antagonist at Work: I think executives underuse one strategic tool, keeping the antagonist in view so they can mobilize teams with focus and urgency.
That’s the context I brought into this conversation with Ethan, and I am grateful for his generosity in sharing his time and experience.
This conversation has been lightly edited for clarity and length.
What Leaders Are Wrestling With
Brad: Much of your public content is about helping people who work with the higher-ups. But flip the script. You’re six years out of Amazon and still in those circles. What questions are you seeing asked at the top right now?
Ethan: The biggest questions at the moment are, of course, about AI and the changes the technology will make in the world. Those are the strategic questions most people are wrestling with: Will AI change my business? Will AI erase my business? Will it enable my business? Can I afford it? How do I stay on top of it? How do I deploy it? There’s a tension between “AI is going to make everything better and more profitable; we’ll lay off so many people and still be great” and the people underneath saying, “We’re not seeing those kinds of gains.” We’re also seeing resistance: People who don’t want to use AI because they fear it will devalue them, or who use it in secret and keep the time they save for themselves while pretending they’re doing everything the hard way. So there are dysfunctions on both sides: executive fantasies about highly profitable, low-headcount, high-output worlds, and employee behavior.
Stepping away from AI, there’s the question about flattening organizations. I have people asking me whether they should give up the management track and go back to being individual contributors, because there’s an idea that managers are passé or superfluous, or that we’ll need fewer of them.
In super broad strokes, the tech industry and the non-tech industry are on different paths. Big tech, like Amazon, Meta, Google, and Microsoft, has been flat or shedding employees after a decade of wild growth. Other companies I talk to are hiring. Maybe more slowly than before, but they never hired at the rate of an Amazon or a Google. They’re growing and trying to expand; they don’t necessarily have a hiring freeze. The tech companies generally feel they over-bloated their staffing and that AI will mean they need fewer, so: don’t hire and flatten your management structure. If there’s a single avatar of this belief that people are copying, it’s Jensen Huang of NVIDIA [known for having dozens of direct reports and few management layers].
When the Company Is Its Own Problem
Brad: A friend of mine, a director at Oracle, told me they spend so much time talking about Microsoft, Google, and Amazon, and rarely acknowledge that their biggest problem is themselves. What do you make of that?
Ethan: Two things, and I’ll quote Jeff. Number one, he taught us to focus on the customer, not the competitor. Be aware of the competitor, yes, but our real job is to serve the customer. Amazon has become more competitor-focused over time, with AWS versus Azure and Google Cloud, and now AI, where many would say Amazon is pretty behind. So there certainly is some competitor focus.
Second, Jeff was aware that we were bad at things. He used to tell a story about visiting a warehouse and finding all these problems in the workflow. He said, in effect, we’re doing okay as a business; imagine how good we’ll be when we’re actually good at this.
Most employees of big companies would agree with the idea that it’s amazing we make so much money given how screwed up we are. I don’t hear employees at any level, in any company of a few hundred people and certainly thousands, say, “This place is a well-oiled machine, well organized, with quick decisions, and good policies.” You always hear politics, confusion, red tape, indecision, and so on. So most people believe what your Oracle friend describes: that we are in our own way more than our competitors are. If we were executing the way we think is somehow possible, we’d crush the competition without getting out of bed in the morning. There are cases where a competitor is a giant, but most of the time, in bigger companies, people feel that if they did things better they’d be winning—hands down.
Day 1 and the Enemy Within
Brad: What comes to mind is Amazon’s “Day 1,” Bezos’s shorthand for running Amazon like a startup no matter how big it got. How much was that an attempt to define success itself as the enemy?
Ethan: It was. The idea was that we’re still early, and we must keep a beginner’s mindset. Amazon did a lot of things to hold on to a startup feel, and they managed it until about 2015. Even though they were already very large, they had the culture and organizational architecture to stay pretty startup-like until 2014 or 2015. Then, from around 2014, they hired so much that every team of ten became a hundred and every team of a hundred became a thousand, and they lost that nimbleness.
That’s one reason they’ve laid off so many people: they’re trying to get back to lean teams. But ironically, Amazon teaches about one-way and two-way doors [Bezos’s terms for irreversible and reversible decisions]. It’s very hard to walk back through an exit-only door in a theater, the kind with no handle on the outside. I don’t think Amazon knows how to walk back through that door, and I’m not sure anybody does when you have 1.6 million employees [Amazon’s current headcount].
Brad: You were there when they had 14,000 people and left when they had 1.4 million.
Ethan: That’s right. In 2005 they were about 14,000, and when I left in 2020, about 1.4 million. Both numbers include warehouse staff, which is the bulk of it. They’ve just grown and grown and grown.
Brad: If the enemy is within, what was the biggest internal force leadership was constantly fighting at Amazon, from your vantage point?
Ethan: The opposite of Day 1 is Day 2. Day 2 is where you become more concerned with politics, with short-term profits, quarterly results, and optics, than with long-term value. You care more about maximizing your local situation than believing that if you do good work there’ll be opportunity. You become more self-centered.
But the biggest thing is conservatism. Jeff wrote extensively about speed mattering in business. He said you need to make decisions with 70% of the information you want. Big companies are very good at making right decisions, but they make them slowly. They become risk-averse and miss opportunities because they won’t move until they’re certain, and by the time they’re certain, some disruptor has seized the opportunity. Most companies, he felt, wait until they have 90% of everything they can imagine, and that’s too slow.
So I think the biggest thing Jeff fought for was moving at startup speed even though we were bigger than a startup. Amazon won that fight for longer than you’d expect, but started losing around 2015, and according to my friends there, ten years later it has more or less completely lost. That means too many levels of review, too much micromanagement, too many people asking about minutiae rather than saying, “How could we be bold?” Too much covering yourself versus taking risks. Jeff and his original management team, almost all of whom are gone, fought really hard to be bold. The 30 or so SVPs who run Amazon now fight really hard to be right. Jeff wanted people to be right most of the time, but willing to be wrong enough to go fast. I think they’ve lost that.
A Common Enemy
Brad: One premise I’m working with is that nothing rallies a team faster than a common enemy. Politicians do it all the time; some religious groups do too. It seems Amazon had that at one time, if the common enemy is Day 2, or certainty. What would you say to an organization dealing with the atrophy that just happens over time? Is there an enemy they could unite around?
Ethan: When Steve Jobs wanted to build the iPhone, he moved a group of Apple employees to another building and encouraged them to fly a pirate flag.
Jeff originally tried to open Amazon in China, and the Chinese government essentially regulated it so local companies could win. So Jeff looked at India and said, “We lost a one-billion-person market; we can’t afford to lose the other one.” And he told the people in India to be “cowboys.” For a while the Amazon India offices had cowboy logos and cowboy themes. And they’ve done much better in India. Amazon India is a viable business, holding its own, whereas Amazon in China is a joke with no path back.
Brad: Pirates and cowboys. That sounds more like an identity than an enemy.
Ethan: So is it an enemy or a vision? If I have to name an enemy, what Jeff taught is that the enemy is bureaucracy, in terms of speed. He also taught that it’s far better to escalate a decision and get a call from someone higher up than to circle and circle around a decision. He said bureaucracy creates victory through exhaustion: The person who sticks to the argument longest wins, because everyone else says, “To hell with it, I’ll do it your way so you’ll shut up.” He didn’t like that. He wanted people to handle their own issues, but if you’ve gone a couple of rounds with your peers, done the right things, and still can’t sort it out, don’t wait. Go ask.
I practiced this. I was in a meeting with Jeff and thought he said, “Do X.” Afterward, Andy Jassy, now Amazon’s CEO, who was Jeff’s direct report at the time, said Jeff didn’t say that. I was sure he had. So I asked Andy whether, at our next meeting with Jeff a few days later, I could just ask. Andy said, “Sure, go right ahead.” I framed it neutrally: “In our last meeting, some of us heard X and some heard Y. Which did you mean?” Jeff asked who heard Y. I said Andy did. Jeff said, “That makes sense. That’s what I said, and Andy heard me right.” The key was that even though Andy was right, outranked me by two levels and knew Jeff much better, he said, “Sure, go ask him.” I learned I had to listen better. But it took less than a minute.
If Jeff had an antagonist, it was being slow, and though he never named it too pointedly, his model for it was Microsoft under Steve Ballmer. The harshest thing Jeff ever said was calling Microsoft “the country club across the lake.” [Microsoft’s Redmond campus sits across Lake Washington from Amazon’s Seattle headquarters.] He felt they were so focused on making employees comfortable, with free drinks, free dinner, and comprehensive medical care, that they’d lost sight of doing business. In that era the big releases were Windows operating systems, and they were always two or three years late. Jeff’s enemy was slowness and bureaucracy.
Brad: Was that enemy galvanizing for everyone else, not just Jeff? Did teams rally around it?
Ethan: Yes. Jeff pressed for aggressive goals. One of the big leadership principles was “Bias for Action”: all other things being equal, we want to move. He taught one-way and two-way doors; I mentioned the drawback of one-way doors, but his point was that most things are reversible. Rather than sit around and debate, if you can test it or just do it, do it, and see what happens.
His other big thing was pushing decisions down. He had a model called single-threaded leaders, or single-threaded ownership, and that’s what allowed Amazon to scale: all these miniature CEOs running around. Give someone a clear charter and all the resources they need, and let them run on their own. Don’t make them coordinate.
He had a famous debate with a Microsoft executive he hired, who said, “Jeff, you have all these redundant teams building similar things. It’s a waste. Why don’t we harmonize everything into one central plan?” Jeff apparently thought about it a long time, then went to the whiteboard and wrote, “Two is greater than zero.” He said every time he told two teams to coordinate, they got locked in debate, and instead of getting two of something, he got zero; neither finished. He’d rather have two teams, and remove the weaker one later. So even though Jeff was very frugal, he saw funding overlapping teams as an investment, not a waste: “I’m not wasting money, I’m making money.” Whether he was right, we can argue, but that’s what he believed.
The conclusion of this interview will be sent out early to my Substack subscribers in two weeks. If you’re not already a free subscriber, please consider signing up.
Does This Content Resonate?
Every organization has an Antagonist. Most leaders can’t name theirs.
After serving as a founder, facilitator, CEO, and chief of staff over the past 30 years, I know it’s not due to a lack of intelligence. It’s due to a lack of visibility.
So I built a short assessment to help you find it. In eight questions and less than five minutes, you’ll get a sense of where the friction in your organization is coming from. After that, if you like, we can jump on a free 30-minute call to discuss more.


