Lead Like You Own the Next Play: Brian Halligan's Playbook for Founder CEOs in the AI Era

Halligan embraces the description from a HubSpot board member that he lives in a "perpetual state of constructive dissatisfaction". He says the fast-growing CEOs he coaches share it: they rarely stop to take stock of what they have done and stay focused on the end state. He is also struck by how humble this generation of founders is compared with his own.

He runs peer groups that he jokingly calls the kids table, with companies under 100 employees, and the adults table, with companies over 100. The adults mostly want to talk about building an exec team and the next level of org design, and Halligan says they spend on average about half their time recruiting and interviewing. Many are surprised that their job is essentially to interview and hire, which he remembers from HubSpot's growth phase.

Halligan says CEOs dramatically overrate their ability to interview and their gut feel, and underrate a high-quality blind reference. He cites a MongoDB CEO who told him that, over about ten years, two C-level executives turned over per year. His own guess is that at least half of senior executives are gone within 18 months, which is why he advises hiring slowly and firing fast, advice he says almost nobody follows.

He shares tactics. Parker Conrad has a C-level candidate sign an NDA, reads the latest board deck, and chats for half an hour; if the candidate is only complimentary, that is a red flag because Conrad wants someone who will challenge him. Halligan also favors whiteboard problem solving over walking through a résumé, and in reference calls asks questions such as whether the person would enthusiastically be rehired, or how likely on a scale of one to ten the reference would be to hire them again.

At HubSpot, candidates were scored from one to four by eight interviewers. When one candidate got four fours and four twos, and another got all threes, they almost always hired the all-threes candidate, with the fewest weaknesses. The team changed course to prefer spikier people, cut the panel from eight to about four, and the hit rate improved.

Halligan warns that founders are blown away by a résumé from Microsoft or Salesforce and then discover an "impedance mismatch": a 50- or 500-person company is not as put-together as the hire expects. At HubSpot, he says, attrition among big-company hires was close to 100 percent. McKinsey hires, in his view, almost always fail because founders are skeptical of conventional wisdom and consultants tend to be conservative in outlook.

His preferred model is the 2004 Boston Red Sox, who ended an 86-year drought with a mix of inexpensive homegrown talent and a few veteran free agents like Pedro Martinez and Curt Schilling. Halligan says companies underrate their homegrown people and overrate shiny outsiders whose warts they have not yet seen. You want a mix, not all veterans and not none, and if it is close, he says, give the internal candidate a shot.

He notes that Brian Chesky has argued the industry over-rotated toward experienced management and delegation. Halligan thinks Chesky is mostly right, though most companies are still following the standard outside-hire playbook.

Asked whether CEOs are born or made, Halligan shares his "LOCKS" algorithm. L is lovable, meaning inspiring followership: would a 28-year-old business school graduate crawl across broken glass to work for this person? O is obsession with the problem, ideally deep founder-market fit rather than an idea from six months ago. C is a chip on the shoulder, a Sequoia emphasis, and K is deep knowledge of the domain.

The S, he adds, is student. He points to founders such as Winston Weinberg of Harvey, who are constant learners "like LLMs" and also study history. Lenny Rachitsky adds that extreme ambition, doing something most people call crazy, is another common thread.

On whether anyone can do it, Halligan says no. He describes a new breed of "five-tool" CEOs, like Bret Taylor, who can code, have taste and vision, and can also sell to investors, recruits and skeptical enterprises. That combination is rare, and he thinks AI is more likely to let people like him build things than to fill a founder's gaps in selling.

The skills new CEOs most need to learn, he says, are giving feedback, building a detector for being sold to, and inspiring people. The hardest recurring conversations are layering executives: telling a co-founder to step aside as CTO, or admitting a first sales leader cannot run the process. His remedy is peers: "misery loves company".

Halligan stands by the line from his MIT course syllabus that starting a company has never been easier while scaling has never been harder. He compares it to going from a few toothbrush options at the drugstore to thousands on Amazon. AWS made starting a software company easier in 2006, and he expects the number of new companies to mushroom over the next decade. The result is a lot of noise and competition, so standing out and accelerating is a distribution problem that founders have to learn quickly.

Rachitsky sums up the problem as breaking through the noise, which is essentially distribution. Halligan agrees and adds that founders did not grow up doing distribution, so they are all learning it on the job. In his view it is a learning game: the faster they learn, the better they do.

Halligan tweeted that sales may be the last job AI replaces. His reasoning: inside a company like HubSpot, AI already works well for software development, customer support and increasingly legal, but go-to-market has no canonical AI application yet. Enterprise sales, where trust builds between two "carbon-based life forms", will arrive late in the white-collar world.

He expects the top of the funnel to be turned on its head. Buyers will start evaluating products in Gemini, Anthropic's assistant or ChatGPT, which already know your site and your competitors, so your website matters less. Websites will instead greet visitors with a high-quality avatar that knows pricing and packaging, logs the conversation in the CRM, and hands off to a rep who brings an all-knowing sales-engineer avatar to every call. HubSpot, he says, has built one and it is working.

He also wants a personal clone connected to his email and meeting notes that can join Zoom calls as a participant. Looking at the AI-native companies he advises, he says the go-to-market motion is the same as it has been for 100 years, except sales engineers are now called forward-deployed engineers. The difference is mainly in the training and integration, and he thinks AEO, getting found in AI assistants, will matter first.

Asked what is most different about being a CEO now versus 10 or 20 years ago, Halligan relays Winston Weinberg's point that you can just do a lot more: what took a year takes two months. He thinks that is a danger, because companies hop to their second act before they have exhausted the first. He points to OpenAI doing many things and then, when Gemini arrived, refocusing on its core.

Planning cycles that used to be a year are now closer to three months, which puts pressure on CEOs to be faster, better decision makers. He recalls that when HubSpot slowed down, it was usually his fault, because a one-way-door decision was sitting on his desk. He used to value optionality, but says there is now a "massive tax" on optionality when you can move this fast.

The first Halliganism, borrowed from Google CFO Ruth Porat, is "when you have to eat a sandwich, don't nibble". Halligan expects a reckoning in private valuations in the next couple of years, with many companies forced into layoffs, and says the temptation to do small rounds repeatedly is a mistake. His advice is to tell everyone the bad news at once, because they are adults.

The second is "next play", from Duke coach Mike Krzyzewski, who yells to players to forget a missed shot and run the next play rather than compound the error. HubSpot used his face on a slide after unforced errors. Halligan recalls a day-long outage on the last day of March 2019, when customers canceled and he cried in front of the company, and says the crisis led them to rethink how they deployed software and there has been no serious outage since.

That led to "never waste a good crisis": most of HubSpot's good changes came from one, and they always swung the pendulum hard the other way. The outage also made them realize that as a CRM, not just marketing software, they were critical to customers' ability to do business.

The plant Halliganism says that if you want to kill a plant, have two people water it: it is either overwatered or not watered at all. Every CEO at the adults table becomes religious about the directly responsible individual (DRI). When a company is small, everyone is in the room; at scale, sales, services, forward-deployed engineers and product are separate, and anything important is cross-functional, so one powerful person must own it, even over people who do not report to them.

He also says there is no silver bullet, only lead bullets. He once thought one hire, investor or product release would transform HubSpot, but inside it was two steps forward, one step back, often triggered by a crisis. He says there is no one to rescue a founder CEO, not parents, VCs or advisers, which hits at the first crisis.

Asked if people start companies too casually, he says he talks many founders out of it, noting that Jensen Huang has said he would not start Nvidia again. Halligan says he would start HubSpot again, and is proud of it, but that it was not glamorous and demands deep obsession.

He also says CEOs of fast-growing companies are working harder than his generation did. People talk about 996, he says, but founders are working seven days a week and texting on Sunday nights, driven by a platform change they do not want to waste. He says he did not have work-life balance and does not recommend it, with Kareem from Clay the one outlier he knows.

The formula "EV is greater than TV is greater than me" stands for enterprise value, team value, and your own. Halligan says immature managers who did not scale optimized for their own team, such as a sales leader pushing bookings and leaving problems to service. HubSpot's quarterly employee net promoter score by department showed it: one sales score dropped from the 60s to 30, then to negative five, and he says teams that lose faith in a leader rarely recover. Later HubSpot added the customer in front: CV, EV, TV, me.

HubSpot tied this into reviews, criticized TV-over-EV behavior in management meetings, and gave out champagne awards to people who modeled it. Halligan quotes a Steve Jobs line that you work for Apple, not your boss.

He also says HubSpot was employee-centric early on, ranking first on Glassdoor, and he now thinks wanting to be liked or to be the best place to work is not a good CEO trait. He would give up ten points of employee NPS for ten points of customer NPS. They shifted by having customer panels at management and board meetings, asking what customers love and hate, and by paying the management team on retention and NPS rather than revenue.

Halligan says changing culture takes overcorrecting, and the bigger the company, the more obvious the signal must be. A CEO has to say the same thing over and over, and be careful what they say, because employees put the CEO on a pedestal: an offhand comment in a hallway can turn into someone's big initiative. HubSpot's co-founder Dharmesh Shah developed a tagging system for emails and messages to signal whether something was a request, something to discuss, or just an FYI.

He also shares that Shah never had a direct report at HubSpot, an arrangement he insisted on from the start.

He describes his own limits as the company scaled: the CEO job shifts from 90 percent perspiration and 10 percent inspiration at the startup phase to the reverse at scale. He says he had trust issues and trusted only a few people to be the DRI on important work, which he calls a scaling limit, and which he sees in every CEO he coaches.

Halligan closes by pointing to his Halliganisms post, which he says collects the mistakes he made in his 15 years as CEO, and to his podcast for CEOs, Long Strange Trip, named for his love of the Grateful Dead.