Reid Hoffman's 10 commandments of startup success

Executive overview

Most startup wisdom focuses on the idea. What actually determines whether a company scales is the discipline applied across hiring, speed, culture, and decision-making — and the grit to sustain all of it through repeated failure.

Reid Hoffman distils a decade of Masters of Scale interviews into ten commandments, each illustrated with concrete stories from founders including Brian Chesky, Mark Zuckerberg, Eric Schmidt, Reed Hastings, and Sheryl Sandberg.

The best ideas sound laughable at first — your job is to find the squirmy no, not avoid rejection.

Commandment 1: Expect rejection, but mine every no

• Rejection is universal; Catherine Minshew of The Muse was turned down 148 times before raising her seed round.
• The best ideas are contrarian — that's why the space is open.
• Distinguish quality of rejection from quantity: a "squirmy no" (friction as the investor reasons toward no) signals you're on to something.
• A flat, pacifying maybe wastes more time than a hard no.
• Tristan Walker (Bevel) identified mid-pitch the moment an investor stopped engaging and moved on immediately.
• Treat rejections as data: mine them for patterns, not validation.

Commandment 2: Hire like your life depends on it

• Brian Chesky spent four to five months hiring Airbnb's first engineer — deliberate, not reckless.
• Ask: would I hire 100 more people exactly like this person?
• Eric Schmidt's formula: persistence plus curiosity is the strongest predictor of success in a knowledge economy.
• Mark Zuckerberg's rule: never hire someone you wouldn't work for in an alternate universe.
• Sheryl Sandberg: cognitive and personality diversity matters as much as demographic diversity — complementary temperaments create resilience.
• At scale, Google reviewed all hiring centrally to prevent "glue people" — nice but low-value connectors — from accumulating.

Commandment 3: Do things that don't scale

• Paul Graham's advice to Chesky in 2009: your users are in New York and you're in Mountain View — go.
• Airbnb's founders knocked on hosts' doors, photographed homes themselves, and lived with users to map every touch point.
• One host returned with a binder of product notes — the team's first roadmap came directly from that visit.
• The 11-star experience thought experiment: map from "they opened the door" to "you're going to space," then find the sweet spot between the two extremes.
• Sam Altman: companies that try to scale before the product is genuinely loved have to generate demand artificially — that's the hard, usually fatal, kind of blitzscaling.
• Stripe, Dropbox, Airbnb all took longer to get the product right, then scaled with organic demand.

Commandment 4: Raise more money than you think you need

• Unexpected expenses are guaranteed; Mariam Nafisi's (Eve.com) domain negotiation ended with equity, a board observer seat, and Disneyland trips for a five-year-old.
• Unexpected opportunities are equally guaranteed — Nafisi's Minted pivot from branded stationery to crowdsourced design consumed nearly all original capital before the real idea emerged.
• Minted closed its Series A two weeks before Lehman Brothers failed; a month's delay would have closed the company.
• Take capital when it's available — you cannot predict when markets will freeze.
• Validate product-market fit before raising; then raise aggressively to fund the scale.

Commandment 5: Release early enough to be embarrassed

• Reid's most-tweeted principle: "If you aren't embarrassed by the first version of your product, you shipped too late."
• Mark Zuckerberg built a crowdsourced art-history study tool in days rather than studying — students' final grades hit an all-time high.
• Facebook ran ~10,000 simultaneous versions of the product at any given time; any engineer could launch a test to 10,000–50,000 users without management approval.
• Move fast and break things evolved into move fast with stable infrastructure — when bug-fixing cost more time than speed gained, the mantra had to change.
• Acceptable error rate: Hoffman told his chief of staff to expect a 10–20% rate of decisions he'd have made differently, in exchange for speed.
• Limit: embarrassment is the bar, not indictment. Launching so fast you generate lawsuits or burn capital without learning means you launched too soon.
• Users cannot always predict their own reactions — Facebook's photo-tagging feature was universally pre-rejected and universally loved in use.

Commandment 6: Decide, decide, decide

• Aviation training principle: make a decision and accept the consequences; indecision is worse than a wrong decision.
• Eric Schmidt structured Google around three weekly meetings (staffing Monday, business Wednesday, product Friday) so every decision had a known slot.
• Google's YouTube acquisition took ten days from discussion to close — preparation and clear ownership made it possible.
• The OODA loop (Observe, Orient, Decide, Act): the organisation with the fastest loop wins; slowness compounds.
• Provisional decisions: form an instinctive view, identify only what evidence would disprove it, go get that evidence — don't wait for completeness.

Commandment 7: Be prepared to make and break plans

• As companies triple in size, systems that worked at the previous scale need replacement — processes built for 9 people fail at 27.
• Sheryl Sandberg: small symbolic commitments (celebrating every birthday that day) matter when reversed; design systems for the company you'll become, not the one you are.
• Mark Pincus (Zynga): "I'll try anything and I'll kill anything and I'll kill it quickly — I'm not going to let killing an idea kill a winning instinct."
• Separate the specific idea (kill it if it fails) from the underlying instinct (preserve it).
• Ocean Spray: after a 49.9/50.1 vote on whether to sell to Pepsi, the losing side fully committed to the outcome — collective buy-in after a decision is as important as the decision itself.

Commandment 8: Don't tell employees how to innovate — manage the chaos

• Google's office design (four people per room, free food, casual environment) consciously recreated Stanford graduate school culture to sustain a sense of discovery.
• Engineers choose their own projects; product leaders must persuade engineers to join, not be assigned them — this filters for ideas worth building.
• 20% time: the real function was a check on managerial overreach — an employee under pressure could legitimately say "I'll give you 100% of my 80% time."
• Gmail, Google Maps, Google News, AdSense all originated in 20% time.
• Effective innovation leadership requires comfort with both humility (you don't have all the best ideas) and uncertainty (you can't schedule breakthroughs).

Commandment 9: Make every employee own the culture

• Reed Hastings' first company, Pure Software, dummy-proofed every process after each mistake — and ended up with only people who valued following process, unable to adapt when C++ became Java.
• Netflix's Culture Deck (now 10M+ SlideShare views) was originally a hiring tool — sent to every candidate so they could self-select out.
• Netflix frames itself as a sports team, not a family: high performance and honest feedback, not unconditional loyalty.
• Margaret Heffernan: internal competition (wanting colleagues to fail so you shine) destroys more companies than almost any other single factor.
• Counter-measure: include in performance reviews and compensation how much an employee helped people outside their own team.
• Interview signal: ask candidates who helped them in their career — if they can't name anyone, that's a warning.

Commandment 10: Have grit — follow the hero's journey

• Grit is not brute persistence up the same hill; it is one part determination, one part ingenuity, one part strategic laziness (conserving energy, minimising friction).
• Reid's board speech at critical junctures: "The road in front of you is super hard. You might die. But if you win, you'll be a hero. Are you a hero?" — people who don't resonate with that framing should leave the boat.
• The hero's journey requires knowing you might fail and moving ahead anyway.

Commandment 11: Pay it forward

• The long-term health of any startup ecosystem depends on successful founders reinvesting — as mentors, angel investors, and by encouraging employees to start companies.
• Linda Rotenberg (Endeavor): where big-fish entrepreneurs eat small-fish startups, the ecosystem stagnates; where they feed them, a multiplier effect takes hold.
• Mexico case study: the country's wealthiest business leaders signed up to support entrepreneurs after Rotenberg framed the choice as "aquarium vs. ocean."

Reid Hoffman: rapid-fire

• Billboard quote: "If I'm only for myself, what am I? If I'm not for myself, who will be for me? If not now, then when?"
• Most gifted books: The Startup of You, The Alliance, The Seventh Sense (Joshua Cooper Ramo).
• Most re-read fiction: Tolkien's Lord of the Rings — for its treatment of the hero's journey, power's corruption, and collective effort.
• Biggest mind-change: AI will have a ferocious, transformative impact — a view he resisted as an AI undergraduate.
• Best hiring rule: hold ideas loosely, share them widely with smart people — your asset is being in motion, not having a secret.
• Criteria for saying yes: (1) big and important, (2) referred by someone he deeply trusts, (3) leaves room for serendipity.
• Favourite failure: Socialnet, his first startup — every mistake he now warns founders against was made there first.