Ethan Evans × Ryan Peterman · Amazon VP: How Corporate Politics Work And How To Win
Evans is already a VP and wants to move to a different role under his own boss. The boss keeps saying no, because Evans is doing well where he is, and the resistance runs for weeks. Then Evans gets a card to play: an SVP in another organization offers him a job. He goes back to his boss and does not say "let me move or I leave." He says, roughly: you know that if I give that man my word, I have to go; you know I really want this other role; you've said no several times; I have to give him an answer. What do you want me to do? [01:00:26]
The next morning he is in the new role. The boss has finally seen that the choice was never "keep him where he is or move him," but "keep him on my team or lose him." Evans' own verdict on the line:
In no way was I lying. But at the same time I deliver it like I'm an angel just fallen from heaven.— Ethan Evans · [01:02:33]
That is the whole three hours in miniature: what a sentence means and how it looks are two separate things, and organizations run in the gap between them.
Evans spent fifteen years at Amazon, retired as a VP in 2020, and now earns his living teaching people how to get promoted. He sets the terms in the first minutes: he is retired, no boss is going to ask him what he told Ryan, so he is immune. The previous episode covered the performance machine and PIPs; this one, at Peterman's request for even more candor, is about politics proper: reorgs, turf, backchannels, how people are eased out. Peterman, a former Meta engineer, plays the wincing engineer throughout. After each tactic he says some version of "this makes me a little uncomfortable," then asks the hard follow-up: what about someone with no leverage?
Two months before the episode aired, Amazon announced about 16,000 more corporate job cuts, which with the October 2025 round made roughly 30,000, framed as removing layers and bureaucracy. Evans' line that once growth stops, politics turns cannibalistic lands squarely in that moment.
Evans starts from a measurement problem. Managers' impact is hard to compare: your product makes money directly, mine is infrastructure, so who contributed more? But nobody can argue with "Ethan has 42 people and Ryan has 17." Headcount becomes the proxy for promotion, and empire building is the rational response to that proxy [00:02:26].
In his day, he says, the thresholds were whisper numbers: he made director with 22 people, and the unspoken bar later crept to 75 or 80. After he left, he was told some orgs wrote the director bar down as 90. Once a number is written, the shell games start. He relays what one boss told a leader:
We're going to put these people under you for 6 months, check the box, and give them back to the person we took them from.— Ethan Evans, relaying a boss · [00:06:54]
Promotions also queue. HR keeps a "forward-looking slate" listing, half by half, who is up in 6, 12, 18 months [00:29:37]. Early promotions usually have no quota, he says, but the higher you go the scarcer it gets: a principal or distinguished engineer costs the company around a million dollars apiece, "so we don't just want to hand them out like popcorn."
Reorgs work the same way. He compares one to a ship leaving the dock: the business reason is why it sails, but since it's moving anyway, you throw more on deck as it pulls away, rewarding the people you want to keep and starving the ones you'd like to see go. He adds one variable. When a company grows fast, people hire instead of fighting; Amazon went from 14,000 to 1.4 million employees during his time (last episode he said 10,000 to a million), and his own team grew from 6 to 800. Now that big-tech headcount is flat, turf can only be taken from colleagues, which he calls cannibalism [00:32:40]. His extreme case is the post office: one postmaster, and until that person leaves or dies, there is no job.
The episode's key phrase is one Peterman picks up on the spot: the polite fiction. Evans' example: if you want more scope, don't say "give me his team." Say "if a reorg needs someone to step up, I'm here, and I'm not planning to leave." Underneath, both mean "I'll take whatever's available." One sounds like a pirate, the other like a selfless servant, and he insists both can be true.
Peterman's reaction, hearing Evans replay his old "my career is very important to me" line, is candid:
I feel the effects of extortion a bit.— Ryan Peterman · [00:22:09]
Evans doesn't deny it. That is the point: everyone can see it, nobody can pin it down. The craft has three layers. Write the key sentence in advance, like memorizing a chess opening, and don't raise your career five minutes after a Friday outage. Be "interpersonally warm, professionally firm": the same words said with a smile are a request and said with a frown are a challenge. And the third layer is attitude. He recommends Arbinger Institute's Leadership and Self-Deception [01:05:57]. The book uses a fictional workplace story to draw one distinction: do you see the other person as a person or as an object? It also describes how you slide into the second: you sense you should do something for someone and don't, which the book calls "self-betrayal"; then, so as not to be the one in the wrong, you start seeing them as in the way, as deserving it. That self-justifying state is "the box." Evans takes the first layer: when you want a raise or a promotion out of your boss, the boss becomes a barrier across the road. He then offers the scene engineers know best as the counterexample, the thing he calls objectifying people:
I'm treating them like an API. I make the right call, I get the right result.— Ethan Evans · [01:06:41]
His alternative is to see them as someone also just trying to get by, and say plainly: here is what I need, what do you need, how do we make this work.
Backchanneling, he says, comes in the same two kinds. The bad kind says privately what you wouldn't say in public. The good kind finds the CFO before the meeting and asks what worries them about your plan. It works not only because information changes hands but because the other person feels consulted and heard. An executive funding a project is asking who will come through when the chips are down; with two similar proposals, even if the rival's is slightly better, the money goes to the one who asked first [01:47:57].
He holds that influence and manipulation use exactly the same skills and differ only in motive; Darth Vader had executive presence in spades. That matches the book: the problem is not what you do but how you see the other person, so the same chess move can be played inside the box or outside it. It also means no one can check from the outside. He tells you to treat people as people while designing every conversation as a chess game, and only he knows which side of the box he is playing from. Peterman presses on the sorest point: you always had leverage, and an ordinary engineer doesn't. Evans' answer is blunt:
I mean, if you really are that discardable, then you've not created much value.— Ethan Evans · [01:43:36]
Where he lands: leverage comes before wording. If your lever is long enough, you can afford to be polite.
Between orgs, the polite fiction becomes tactics. On defense, the "skunk trick": list every unglamorous duty nobody wants, say a "Yugoslavian tax engine," and have its customers show up to say how vital it is. Add the porcupine: when probed, hit back fast and publicly so the raider goes looking for a softer target. On offense, the best move is a real case that you'd grow the business. The slyest is to ask a struggling team, "Is there anything I can do to help?"
That's my opening gambit for I want to take it over.— Ethan Evans · [01:29:14]
For people under a weak manager, his counterintuitive advice: if you go over your boss's head, never go alone. He does the math from the skip level's chair. Believing you creates three problems: decide what to do with the manager, hire and train a replacement, and do the job himself in the meantime. Disbelieving you creates none. He once missed for a long time that a leader under him was mistreating the women on his team, until several accounts reached him one after another [00:53:00].
Going over your boss's head has a ceiling too. Evans has seen skip levels quietly move an engineer who was about to be fired, but has never really seen anyone promoted over their manager's objection: overruling the manager means telling them they're wrong and then making them work with a more senior you [02:40:40]. Either win your manager over or get out from under them.
Then the exit. When easing out a manager, he never says "fired" or "performance." He says something like "I don't think we've gotten off to the best start; would you like my help finding another option?" A smart manager replies "I'm sorry you feel that way," and the negotiation has begun [02:09:08]. The person leaving has leverage too: saving the company the time a formal process costs, the story they'll tell their coworkers on the way out, and the manager's wish to believe they took care of them. None of the three is in a contract; all of it is narrative.
Narrative also decides whether a bad actor can keep getting hired. Evans describes a newly hired chief-level executive who began harassing women at once; when the company asked an HR representative on his team about complaints, she started crying, because he had harassed her too. He was fired quickly, but the company had testimony, not a judgment, so reference checks got only dates of employment. Evans later looked him up on LinkedIn: the Amazon stint had vanished from his résumé [01:58:15].
"Saying 'we might have to fire you, do you want to quit first?' creates a legal problem called constructive termination." [02:09:08]
Supported, but overstated
In Green v. Brennan (2016) the U.S. Supreme Court describes constructive discharge as discrimination so severe that a reasonable person would feel compelled to resign, followed by an actual resignation. It is a rule for counting a resignation as a firing inside a discrimination claim, not a wrong in itself. Careful wording is good practice; the legal label is stretched.
"I know he harassed people, but there's no judgment, so if I say it he can sue me. That's why companies only confirm dates." [01:59:45]
Supported
In Randi W. v. Muroc (Cal. 1997) the defendant school districts themselves argued employers are better off confirming only position, salary and dates. The court held mere nondisclosure creates no liability, but letters praising a man "without reservation" while hiding known sexual misconduct could, where physical harm was foreseeable. Silence is safe; half-truths are not.
"Amazon's leadership principles say there's no bonus for headcount, yet after I left, some orgs wrote the director bar down as 90 people." [00:06:54]
Insufficient evidence
The first half is true: Frugality reads "There are no extra points for growing headcount, budget size, or fixed expense." The 90-person bar rests only on what his coaching clients told him; no document turns up. The public direction runs the other way: in 2024 Jassy ordered the ratio of individual contributors to managers raised by at least 15%.