· Brian Halligan

Sequoia CEO coach: Why it's never been easier to start a company, and never been harder to scale one

A 20-year-CEO-turned-coach's operating rubric for scaling a company past 100 employees: hire slow, fire fast, pick spikes over balance, kill the silver-bullet fantasy, and over-communicate the shift from perspiration to inspiration.

90% confidence

Why this is in the corpus

Halligan has been running HubSpot through every phase from kids-table to scale-up and now coaches dozens of Sequoia CEOs — dense with operator-graded tactical content on hiring, DRI, EV>TV>MEV, next-play from someone seeing the same failure modes repeat across the current crop of founders.

Summary for skimmers

Sequoia in-house CEO coach Brian Halligan walks through his hiring rubric (blind references, spiky over three-out-of-four, avoid big-company hires), the CEO LOCKED+S profile, adult-vs-kids-table scaling transitions, and Halliganisms — EV>TV>MEV, eat-the-shit-sandwich-don't-nibble, next-play, kill-a-plant-with-two-waterers, silver-bullet-fantasy.

Briefing

What survives the editorial filter

This page should feel like a smart colleague already listened for you and left only the operating logic worth keeping. Not everything said in the episode makes it through.

Trust signal

Direct episode extraction

Guest type: practitioner.

Best used for

Decision-grade retrieval metadata not yet added for this episode.

Hold lightly

No explicit downgrade reason stored yet for this episode.

Principles

Durable claims that survive beyond the speaker's biography — each with explicit limits, transferability judgment, and evidence.

Principle

The big-company hire fails by impedance mismatch, not by skill gap

Hiring a fancy-title VP out of Microsoft / Google / Salesforce into a 50-500 person startup produces near-100% attrition not because the hire is unqualified, but because their operational expectations assume infrastructure the startup doesn't have.

Halligan cites HubSpot's own ~100% attrition on hires from Salesforce, Google, Microsoft; generalises to the McKinsey-Apple-Amazon-cohort pattern other founders have walked into.

There's just a massive impedance mismatch when you hire them... you definitely don't if you're 50 or 500 employees and they expect you to have your act together.Brian Halligan
We hired so many people from Salesforce and Google, Microsoft — like a hundred percent attrition rate on all those folks.Brian Halligan

Principle

Blind references are the highest-signal interview input — but most are box-checking theatre

The single most predictive hire-quality signal is a blind reference call with hard, specific questions. Most reference calls are already-decided theatre; the best ones use specific rubrics (top 1% / top 10% of employees; enthusiastically-rehire 1-10 scale).

Halligan receives these calls himself and says he can tell in the first two questions whether the caller has already decided — the box-checking refs never ask the top-decile or enthusiastic-rehire questions.

One of my favorite questions people ask is, would you enthusiastically rehire this person for that role?Brian Halligan
When they ask me something hard, like on a scale of one to 10, how likely you're to hire them again... or were they the top 1% of your employees? That's a good question.Brian Halligan

Principle

Homegrown talent is systematically underrated; promote before you poach

Internal VPs are routinely passed over for external hires who interview better because the external candidate has rehearsed the résumé story and the internal has been visible warts-and-all. Give the homegrown the shot if the finalists are close.

Halligan cites HubSpot (half the current management team has been there ~15 years) and Apple's bench of homegrown execs.

If you look at HubSpot, half the management team are folks that have been there for approximately 150 years.Brian Halligan
If it's pretty close, I think you give your homegrown a shot at it.Brian Halligan

Principle

Hire the spike, not the three-out-of-four: low-variance candidates lose to outliers

When interview panels produce a split-vote 4-out-of-4 / 2-out-of-4 candidate vs a unanimous 3-out-of-4 candidate, the correct pick is the spiky one. The well-liked average is a statistical artifact of weakness-seeking, not excellence-seeking.

HubSpot changed its policy explicitly on this; Halligan credits improved C-level hit rate to picking spikes and also to shrinking the panel from 8 interviewers to 4.

We changed it and we went with the spiker people... we went with people with weaknesses, we went with people with challenge stuff and that has worked out quite well.Brian Halligan
Almost every time we hired the three out of four, like the person with the least amount of weaknesses.Brian Halligan

Principle

Hire slow, fire fast — the default human bias is the opposite

Most CEOs hire fast and fire slow; the correct posture is the inverse. Halligan estimates ~50% of C-level execs are gone within 18 months of hire, so the hire deserves more time and the separation less.

Cross-validates with MongoDB CEO Dev Ittycheria's data point: 2 C-level turnovers per year over a 10-year span. Halligan says HubSpot was similar.

Hire slow and fire fast. People hire fast and fire slow. And if I had to guess, Lenny, within 18 months after you hire a C-level exec, at least 50% of the time they're gone.Brian Halligan
Dave, the CEO from MongoDB... two C-level turned over per year.Brian Halligan

Principle

Perpetual constructive dissatisfaction is a core trait of scale-up CEOs

The CEOs who sustain growth past 100 employees share a posture of constant, positive dissatisfaction — focused on the end state, rarely taking stock of what they've done.

Halligan learned the phrase from Noor Laurie Norrington, HubSpot's chair. He now sees it across the Sequoia portfolio and uses it as a positive descriptor, not a clinical one — paired with what he calls a humble generation of CEOs.

They're always a little bit dissatisfied with where they are and very focused on the end state.Brian Halligan

Principle

There's a massive tax on optionality when everyone else moves in 3-month cycles

Planning cycles have compressed from annual to quarterly. CEOs who preserve optionality (open one-way doors, delay irreversible calls) now pay a much larger tax than they did pre-AI: the org slows while competitors ship.

Halligan self-diagnoses this as his own pattern at HubSpot: "when things slowed down and there was churn, it was usually my fault" — sitting on one-way doors.

I think there's a massive tax and optionality when you can move this fast and try a lot of things.Brian Halligan
The planning cycles used to be a year. I think the planning cycles now are three months long.Brian Halligan

Frameworks

Reusable systems and operating models — including when they help and when they break.

Framework

LOCKED+S: Halligan's founder-quality rubric

Halligan evaluates founders on five traits: Lovable, Obsessed, Chip on shoulder, Knowledgeable, Student.

Halligan uses the rubric when Sequoia evaluates new investments. Cites Winston Weinberg (Harvey), James (Profound), Gabe (Rogo) as Student exemplars.

  1. L — Lovable: could a 28-year-old you crawl across broken glass to follow this person?
  2. O — Obsessed: evidence in their life of going deep down rabbit holes; deep founder-market fit, not a 6-month pivot
  3. C — Chip on shoulder: Sequoia criterion, boulder-sized chip to prove someone wrong
  4. K — Knowledgeable: deep domain command, not surface-level
  5. S — Student: learn-it-all not know-it-all; reads back decades of history
Use when: Early-stage founder evaluation (investor diligence, recruiter vetting, exec-team selection).
Skip when: Late-stage CEO succession where the bar is operational discipline rather than founder-grade conviction.
I call it my LOCKED algorithm. L is for lovable... O is just obsession... C is something I wouldn't have thought of, but this is a Sequoia thing, like chip on the shoulder... K is just for deeply knowledgeable about the domain.Brian Halligan
If I were to stick an S on it, I would say student.Brian Halligan

Durability: Durable — the trait set has been stable across Halligan's 20-year CEO career.

Framework

CV > EV > TV > MEV: the scale-up decision priority order

At scale, decisions should be evaluated in a strict priority order: Customer Value, Enterprise Value, Team Value, Me Value. Inverting the order is the most common failure mode of otherwise-competent VPs.

HubSpot codified this in employee reviews and management meetings. The failure mode shows up in quarterly employee NPS: a department collapses from 60 → 30 → -5 as the team catches on, and almost never recovers.

  1. CV — Customer Value: does this improve the customer outcome?
  2. EV — Enterprise Value: whole-company metric (retention, NPS, durable revenue)
  3. TV — Team Value: does this help my function hit its number?
  4. MEV — Me Value: does this help my comp, title, career?
  5. Rule: higher-priority values must not be traded off for lower ones
  6. Instrument: quarterly employee NPS by department; watch for collapsing scores
Use when: Mid-to-late-stage companies (~150+ FTE) where cross-functional externalities dominate.
Skip when: Very early stage (<50 FTE) where everyone is in the room and the framework is unnecessary overhead.
We always put on the wall, solve for EV over TV over me, and then we added CV in front of EV — solve for the customer first, then HubSpot, then the employee, then yourself.Brian Halligan
Sales net promoter score was like 65, 62, 68, 30... from 30 to negative five. They never actually recovered.Brian Halligan

Durability: Durable — the structural driver (incentive misalignment at scale) is not AI-era specific.

Signals

What appears to be shifting, for whom it matters, and what happens if you ignore it.

Signal

Top-of-funnel is shifting from SEO to AEO — buyer research starts in ChatGPT / Gemini / Claude

People are evaluating a product, they start in Gemini or they start in Anthropic or they start in ChatGPT... so your website's a lot less important.Brian Halligan
I think sites will change where you're going to have a really high quality avatar that knows everything about your products, knows everything about your company and your pricing and packaging.Brian Halligan

Opportunities

Only included where there is a buyer, a real wedge, and a plausible revenue path — not vague idea theater.

Opportunity

Personal knowledge-worker AI avatars that attend meetings on the user's behalf

I have like a Delphi clone... what I want is like, connect that thing to my email, into my granola, into my plot, and it knows everything about me. And then when I go to a meeting, I wanna invite that thing to my meetings... every knowledge worker will have one of these in 3, 4, 5 years.Brian Halligan

Lessons still worth keeping

Useful takeaways that did not fully clear the bar for durable principle status.

Lesson

Next play: absorb the error publicly, move on, do not compound it

When an unforced error hits, the next-play discipline — from Duke basketball coach Mike Krzyzewski — is to acknowledge it publicly, refuse to compound with overreactive decisions, and run the next play.

Halligan literally put Coach K's face on a slide at HubSpot company meetings after the March 2019 outage he describes as the worst day — he cried in front of the company.

There were times in HubSpot's history where we had the Mike Krzyzewski face on a huge slide in front of the company meeting saying next play because there was an unforced error and we need to deal with it and kind of move on.Brian Halligan

Durability: Durable — applies to any high-iteration environment.

Lesson

Flash tags: CEO words land as orders unless explicitly tagged

As the company grows, the CEO's casual comments in hallways, Slack, and email get interpreted as priorities regardless of intent. HubSpot built a tagging rubric (FYI vs Discuss vs Do This) on every CEO communication.

Dharmesh Shah (HubSpot co-founder) developed the specific tagging system. Halligan treats it as required discipline past the point of personally knowing every FTE.

Somebody inevitably would go home and build that thing and be like, Brian wants this as a big initiative... we needed to tag each email with how do you want me to get this done this week? Or is this something we should talk about? Or is this something that's just FYI?Brian Halligan

Durability: Durable — the pedestal effect scales with org size.

The Plays

Try these this week

Verb-first executable actions — each one tied to a stated outcome in the episode.

Next Play company meeting ritual: public error absorption with a recurring visual cue

Outcome: When the company has an unforced error, run a next-play ritual at the next all-hands: put the Coach K slide up, have the CEO own the error publicly, name what changed, then move to the next play. Same slide every time — recurrence is the ritual.

There were times in HubSpot's history where we had the Mike Krzyzewski face on a huge slide in front of the company meeting saying next play because there was an unforced error and we need to deal with it and kind of move on.
Brian Halligan
Triggered by events, not calendared. Used ~quarterly at HubSpot.
  1. 1

    Maintain a ready-to-deploy Coach K Next Play slide

    Same slide every time. Recurrence is the point; do not re-design per crisis.

  2. 2

    Trigger on any unforced error above a fixed severity bar

    Bar: customer-visible outage, missed quarter, major-hire attrition inside 18 months, security incident.

  3. 3

    Lead the all-hands: CEO names the error without hedging

    No passive voice, no spreading blame. If the CEO was the cause, say so.

  4. 4

    Name what changed — one specific decision or process

    No laundry list. One change that prevents recurrence. HubSpot 2019 outage: changed how software was deployed.

  5. 5

    Put the Next Play slide up; physically move on in the deck

    Slide transition IS the ritual. People associate post-error behaviour with the visual.

  6. 6

    Pair with a structural fix — never waste a good crisis

    Ritual handles the team; structural fix handles the root cause. Two moves, not one.

Scripts

ceo-opener

We had an outage yesterday. It was on me. Here's what happened, here's what we're changing, and here's the next play.

Before you start

  • · All-hands cadence triggerable off-schedule
  • · CEO willingness to own errors publicly
  • · Culture where named fixes actually ship
leadershipcrisis-managementorg-designseries-bgrowthscale

NDA + board-deck interview: screen C-level execs for challenger DNA in 30 minutes

Outcome: Before hiring a C-level exec, have them sign an NDA, send them the most recent board deck, and schedule a 30-minute unstructured conversation about it. If they are only complimentary, that is a red flag — you wanted a challenger, not a yes-person.

Parker Conrad has a good hack that I like: before he's got a C-level interview with CFO, Chief Product Officer or whatever, he has him sign an NDA and sends them the last board deck or the board memo or some important doc and he schedules a half hour interview with them and he just has a chat about the deck. And if they're just very complimentary and 'so great' and 'you're doing this great amazing thing,' it's a major red flag to him.
Brian Halligan
30-minute interview; ~2 hours per candidate including blind references.
  1. 1

    Send an NDA ahead of the interview

    Standard mutual NDA; scope covers the document you're about to share.

  2. 2

    Send the most recent board deck or a board memo on a live strategic question

    Real material, not a sanitised case. Stakes must be visible for the signal.

  3. 3

    Schedule a 30-minute conversation (not a presentation)

    Unstructured. Open with what did you think — do not hand them a framework.

  4. 4

    Listen for challenger DNA

    Red flag: this is great, you're doing an amazing thing. Green flag: concrete objection, thoughtful re-framing, asks you hard questions back.

  5. 5

    Cross-reference with blind references

    Independently call 2-3 blind references: top 1% of employees? Top 10%? Enthusiastically rehire 1-10?

Scripts

opener

Thanks for signing the NDA. I sent you our last board deck. Tell me what you think — where do you disagree with us?

Before you start

  • · Mutual NDA template ready
  • · A real board deck / memo
  • · CEO bandwidth for unstructured 30-min conversation
  • · Blind reference pipeline
hiringleadershiporg-designseries-aseries-bgrowthscale

Decision Moments

Actual decisions, real outcomes

Specific decisions narrated in the episode with their outcomes and transferable lessons.

March 2019: HubSpot had an all-day customer-visible outage; customers cancelled; Halligan took personal calls from furious customers. The first serious outage the company had ever had.

Did: At the next all-hands used the Coach K "Next Play" slide, cried publicly, named the error as his, named the specific software-deployment process change that would prevent recurrence, and moved on.Outcome: Deployment process permanently changed; HubSpot has not had a serious outage since; team recovered rather than compounding the error with recrimination.

Public ownership + one named fix + visual ritual bounds the error cost. Never waste a good crisis, but also don't let the crisis become the new identity.

Part of an emerging decision pattern across multiple episodes

After 15 years as HubSpot CEO and a near-fatal snowmobile accident, Halligan lay at the bottom of a cliff expecting to die overnight in Vermont.

Did: Decided, while waiting for rescue, to step down from HubSpot CEO role. Followed through: handed role to Yamini Rangan who remains CEO. Became Sequoia in-house CEO coach and started Long Strange Trip podcast.Outcome: Rangan has sustained HubSpot's growth; Halligan is now coaching ~30 Sequoia CEOs and running a peer-learning programme.

Life-altering events clarify founder-role fit. The right question for a 15-year founder is not "can I keep doing this" but "does this still match my harmonic motion" — and the answer should drive the succession decision.

Part of an emerging decision pattern across multiple episodes

HubSpot interview panels kept producing unanimous 3/4 candidates over split 4/4-and-2/4 candidates, and the 3/4 hires were systematically under-performing.

Did: Changed policy: pick the spiky candidate with some 4s and some 2s, not the safe 3/4. Shrank the interview panel from 8 to 4 to reduce anchoring.Outcome: Hit rate on C-level and senior hires improved measurably.

Unanimous 3/4 is the statistical signature of a candidate with no strong signal either way. Select for variance, not consensus.

Part of an emerging decision pattern across multiple episodes

HubSpot topped Glassdoor best-place-to-work and Halligan was #1 Glassdoor CEO, but customer NPS was ~25 while employee NPS was ~60. Two of four management-team hours went to employee experience.

Did: Shifted centre of gravity from employee-centric to customer-centric: changed management-team comp to retention + NPS (not revenue), put customer panels in monthly management meetings and board meetings, asked customers live "what do you hate?"Outcome: Customer NPS rose; employee NPS intentionally traded down 10 points to buy 10 points of customer NPS.

Employee-love and customer-love compete for management attention at scale; you must pick a centre of gravity, not drift between them.

Part of an emerging decision pattern across multiple episodes

Tensions surfaced

Contradictions and trade-offs the episode raises — judgment calls a thoughtful operator has to navigate.

Tension

Customer-centric vs employee-centric: a company can only have one centre of gravity

Companies optimise toward one centre of gravity — customer, employee, or investor. Best-place-to-work rankings and employee NPS can become a local maximum that crowds out customer outcomes.

HubSpot ran 2 of 4 management-team hours on employee experience and topped Glassdoor. Customer NPS ~25 vs employee NPS ~60 was the trigger to shift: changed comp (mgmt paid on retention + NPS, not revenue), put customer panels into monthly management meetings and board meetings.

We were number one in Glassdoor's best place to work. I was the number one CEO in Glassdoor. And as I look back at that, I'm not sure that's a good thing.Brian Halligan
I would give up 10 points of employee net promoter score to get 10 points of customer net promoter score.Brian Halligan

Durability: Open question — no universal answer; each CEO must pick.

Corpus connection

Where this episode fits for retrieval

What kinds of decisions this briefing is best pulled into.

Primary decisions

  • leadership
  • hiring
  • scaling
  • org-design