The Price of Excellence Most People Won’t Pay

KEYPOINTS

• 01

  All successes are unique, but all failures rhyme: good people leave, innovation stalls, G&A swells, and the company starts fighting customers and partners.

  Kaz Nejatian took over Opendoor months from bankruptcy and found that exact pattern. His diagnosis is that a lack of active management lets companies atrophy: when the founders are gone and no large shareholder cares, firms drift into avoiding looking bad and become 'hosts for professional leeches' who drain the company while making management look good.

• 02

  Change the defaults that are killing the company violently — with no change management — so the change feels jarring.

  Opendoor had produced many 'seven-step plans over eight months' to return to the office, and none worked. At 9am on his first Monday Kaz told everyone the company was back in the office next Monday, and anyone not planning to come was thanked and let go. His view is that gradual plans let the old default win; aggression is the only way to reset a deeply held norm.

• 03

  Create strong repelling forces: the people you hire are the company you build, so make it easy for the wrong people to opt out.

  In week one he rewrote the careers page from 'a happy place where we all care about each other, here are our ERGs' to 'This will be hard.' He sent the company a blueprint designed to make anyone who disliked how he'd run things quit, freeing room for committed people to opt in. Rebuilding meant essentially hiring 'an entirely new company,' which he admits was very painful.

• 04

  The most dangerous employees are competent but not mission-aligned, and they should be helped to move on fast.

  On a 2×2 of competence vs. alignment, the competent-but-misaligned quadrant tilts the organisation toward itself through internal politics. They keep getting promoted because bosses don't see what they're doing. Kaz deliberately widened that rift early. Meanwhile he found roughly a dozen exceptional ICs who had kept Opendoor alive 'on their shoulders' and rebuilt the company around them.

• 05

  Truth over feelings: groups drift toward nice lies, and truth shrinks as rooms get bigger.

  People are truthful relative to who's in the room, so large meetings carry less truth — one reason big companies are usually average and small teams work. A new CEO must find the 'speakers of truth' and get them alone. He calls the alternative 'suicidal empathy': assuming everyone is thin-skinned and offended by facts, which he says is killing good companies.

• 06

  Meetings are a bug: they exist because people couldn't get the information they needed beforehand.

  Opendoor holds meetings mostly to broadcast, not to present for approval. It is considered impolite not to disagree, and impolite to stay in a meeting you have nothing to contribute to — people walk out mid-meeting. Meetings run louder and 'more ornery' than average, but Kaz frames disagreement as respect: all mistakes and all information belong to Opendoor, not to individuals.

• 07

  'Say the thing': say it now, about the thing not the person, and repeat it until you've been heard.

  Written at Shopify after sitting in meetings where everyone acted fine while 'the shape of the chart' said otherwise. People avoid being the one who 'farted in church.' The rule: 'this sucks,' never 'you suck'; others needn't agree, but must hear you. At Opendoor, failing to object to stupid things counts as rude.

• 08

  AI is less about replacing people than collapsing management layers and changing who wins.

  Cheap communication flattens 'managers managing managers' chains. Opendoor built an 'AI exoskeleton' around every employee, making them 3–4× more efficient: the last time it bought as many homes as last quarter, opex was over twice as high. It also has more engineers than when he joined.

• 09

  Companies go upside down one degree at a time, and acronyms are a warning sign.

  New pilots in clouds turn slightly, slightly, and emerge fully inverted. Acronyms create in-groups and out-groups so nobody can tell anyone else is inverted. Kaz also assigns Orwell's 'Politics and the English Language,' challenges anyone to find a sentence with 'leverage' that means anything, and insists people decide things, not teams — never 'Legal decided.' Similarly, execs running on dashboards are 'playing SimCity' with derivative facts (his Hayek-inspired 'map is not the territory').

• 10

  Strategy is overrated relative to product and execution; delaying decisions rarely reduces risk.

  Google's strategy wasn't meaningfully different from 25 other search engines — execution was. The real test is whether you can find 1,000 people who love the product. A/B tests often become management's 'safety blanket'; he mocks Google testing ~400 shades of blue. Action creates information, and pain tolerance — holding your hand over the fire longest — separates winners.