Dharmesh Shah Leadership: Stylized flat editorial waist-up portrait of Dharmesh Shah, middle-aged Indian man with short salt-and-pepper hair, rectangular dark glasses, clean shaven, navy crew-neck shirt.

Turn this article into takeaways for your work.

Summarize with ChatGPT

Summarize with Claude

Each assistant summarizes the article only for you and suggests best practices for your work.

Key Facts: Dharmesh Shah co-founded HubSpot with Brian Halligan in June 2006 after the two met at MIT Sloan, and he has held the title Co-Founder, CTO ever since, through HubSpot's 2014 NYSE listing under ticker HUBS and past $3.13 billion in full-year 2025 revenue. Yamini Rangan is CEO, and Brian Halligan, who ran HubSpot as CEO until 2021, stepped down from his executive chairperson role in May 2025 to become a non-employee director. Shah wrote and published the HubSpot Culture Code in 2013, a deck that has logged more than 5 million views and is still being revised today: the public copy on HubSpot's own site is tagged V41 and dated September 29, 2025. Outside HubSpot, Shah runs OnStartups.com, which grew out of his MIT Sloan thesis on software startups, and he now builds agent.ai, a marketplace for AI agents, alongside the simple.ai newsletter, which had passed 2 million subscribers as of this writing.

The Culture-as-Code Doctrine

Dharmesh Shah's Culture-as-Code Doctrine treats a company's values the way he treats software: something you design, ship, measure, and rewrite once it stops fitting the current build, not something you print once and hang in the lobby. Under this doctrine, a values document carries a version number the same way a product roadmap does, and it earns the same question a product manager asks about a feature: does this still serve the customer, where the customer is the employee reading it? The doctrine's other half is about who does the writing. Shah, the technical co-founder, kept authoring that code himself as CTO rather than handing it to a communications team or trading his engineering seat for the CEO's chair. Two decades after founding HubSpot, the Culture Code he first shipped in 2013 is on its 41st tracked revision, and he is still the one deciding what ships next.

Most technical co-founders don't get twenty years at the company they started. They take the CEO job and become full-time operators, or they get bought out, sidelined, or quietly replaced by a hired executive once the company needs someone who can run an investor roadshow instead of a sprint planning meeting. Dharmesh Shah did neither. He co-founded HubSpot in June 2006, never took the CEO title, and is still listed as Co-Founder, CTO on the company's own leadership page today, two decades and one IPO later.

Shah met Brian Halligan at MIT Sloan, and the two built HubSpot into a company that took inbound marketing from a scrappy idea to a NYSE listing in 2014 under the ticker HUBS. Halligan owned the go-to-market half of that story: he ran the company as CEO, handed the role to Yamini Rangan in 2021, and eventually left his executive chairperson post too, stepping down to a non-employee board seat in May 2025. Shah's half of the story runs in parallel and looks almost nothing like it.

HubSpot itself is a textbook case of the kind of low-end disruption that unseats incumbents: a marketing platform built for small businesses the enterprise vendors of 2006 had no interest in serving, priced and packaged for people the market leaders considered beneath them. Shah's specific contribution to that disruption wasn't a pricing model. It was deciding, in 2013, that HubSpot's internal culture deserved the same build-measure-learn discipline he applied to code, an idea few technical leaders extend past the product they ship.

The result was the HubSpot Culture Code, a slide deck Shah wrote and has kept revising since 2013. It borrowed its DNA from the Netflix Culture Deck that Reed Hastings and Patty McCord published four years earlier, but Shah treated the borrowing as a starting commit rather than a finished product. The version circulating publicly today carries a version number, V41, dated September 29, 2025, which is a strange thing to say about a document most companies write once and never touch again.

That instinct, treat the soft stuff like software, is also what's driving Shah's current chapter. Nearly twenty years into being HubSpot's CTO, he's building agent.ai, a public marketplace for AI agents, and writing the simple.ai newsletter on the side. This piece traces the decisions behind that arc: staying CTO, publishing a values document that made HubSpot accountable to its own words, and betting a second act on a technology most sitting executives leave entirely to their product teams.

Leadership Style Breakdown

Shah combines product thinking about culture with a durable technical role and room for independent experiments.

Culture as a Product: revising culture document with employee feedback.

Style	Weight	How it showed up
Product-Minded Culture Architect	45%	Shah's most consequential leadership decision at HubSpot wasn't a technical one. It was applying product thinking to the one part of a company that usually escapes it: culture. He wrote the original Culture Code himself in 2013, framed employees as the culture's customers, and treated the deck as a live release rather than a plaque. That framing shows up in a line he's repeated on his own blog: "culture is a product you build, and your people are its customers". The Code has been revised for over a decade, and the copy HubSpot publishes today is versioned V41, dated September 29, 2025, on the company's own culture-code page.
Durable Technical Co-Founder	35%	Most co-founding CTOs either move into the CEO seat as the company matures or get replaced by a hired technical executive once the engineering org outgrows a founder's management style. Shah did neither. HubSpot's own leadership page lists him as Co-Founder, CTO today, the same title he's held since 2006, through an IPO, a CEO transition from Halligan to Yamini Rangan in 2021, and Halligan's own departure from the board's executive chair seat in May 2025. Twenty years of title stability at a public company is unusual for any executive, and rarer still for the technical co-founder.
Second-Act Builder	20%	In the last few years, Shah has built a second body of work almost entirely outside his day job: agent.ai, a public marketplace for AI agents, and the simple.ai newsletter that explains how to use them. He built agent.ai through a separate LLC staffed with freelancers rather than HubSpot's own engineering team, explicitly to move faster than a public company's product organization can, a structural choice that says as much about his operating style as the product itself.

The 45/35/20 split isn't a coincidence. Culture-as-product is the load-bearing habit; staying technical for two decades is what let him keep authoring it personally instead of handing it to whoever held the CTO title next; and the AI agents work is the newest expression of the same instinct, applied to a technology most executives at his level would only fund, not build.

Key Leadership Traits

Trait	Rating	What it means in practice
Treating soft systems like engineering	Exceptional	Shah's default instinct, applied across the Culture Code, his blog, and his approach to AI tooling, is to take something people usually treat as fixed, values, habits, workflows, and rebuild it as a system with inputs, iteration, and a version history. He's said outright that "culture debt is much more insidious than technology debt", a sentence that only makes sense if you already think about culture in engineering terms.
Role discipline (staying in your lane)	Very High	Shah has had two decades of opportunities to angle for the CEO title, first as a founder with the standing to ask for it, later as a public-company executive with the tenure to justify it. He never did. HubSpot's own management page still lists him as Co-Founder, CTO, the same functional lane he chose in 2006.
Willingness to publish and be held to it	High	Publishing the Culture Code externally wasn't a low-risk move. It gave every employee, and later every journalist and critic, a written standard to measure HubSpot against. When Dan Lyons' 2016 book Disrupted argued HubSpot didn't live up to its own deck, the deck itself is what made that argument possible to make. Shah kept publishing it anyway, and kept revising it.
Building outside the day job	High	agent.ai runs on a separate legal structure with contractors instead of HubSpot staff, by Shah's own description, so he could experiment without pulling focus from HubSpot's core team. That's a specific kind of discipline: building a second body of work in public without asking your primary employer to absorb the risk of it.
Plain-spoken self-assessment	High	Asked what he considers his most significant contribution to HubSpot, Shah didn't point to a product feature. He pointed to the Culture Code: "of all the code I've ever written, I'll say this, that's likely been the most impactful". That's a notable thing for a CTO to say about a slide deck instead of a system he built.

The 3 Decisions That Defined Dharmesh Shah as a Leader

Three choices reveal the pattern: staying technical, maintaining a public culture document and protecting the core team from experimental work.

Choose Where Your Judgment Compounds: working on core design beside separately housed prototype.

1. Staying CTO instead of taking the CEO seat

When two people co-found a company, one of them usually ends up as CEO, and the other either finds a functional home or leaves within a few years once the operational demands of scale outpace what a technical co-founder wants to do all day. HubSpot's own history includes exactly that split: Brian Halligan took the CEO role at founding and held it until 2021. Shah could have pushed for it, or for the executive chairperson seat Halligan later held until May 2025. He did neither. HubSpot's management page and its investor-relations bio both list him the same way today they would have in 2007: Co-Founder, CTO.

Staying technical at that level of seniority isn't automatic. It requires deliberately refusing promotions that most executives in his position would take, because a CTO title after an IPO and $3.13 billion in annual revenue carries less institutional power than a CEO or chairman title, even at the company you founded. Kelsey Hightower made a similar bet at a different altitude, turning down VP titles throughout a career in cloud infrastructure to stay close to the technical work. Shah's version of that bet happened at founder scale, with founder equity and founder standing, which makes turning down the bigger title a more deliberate act, not a default.

What this shows: the CEO seat is not the only source of power in a company you founded. If your comparative advantage is technical judgment and your co-founder's is operating and go-to-market, staying in your lane after the IPO is a legitimate strategy, not a failure to advance. It only works if the co-founder relationship can survive one of you outranking the other in headline terms for twenty years.

2. Publishing the Culture Code, and refusing to let it go stale

Most internal culture documents stay internal, and the ones that do get published are usually static: a one-time PDF, a values page on the careers site, updated maybe once every few years when legal or HR flags something dated. Shah did the opposite twice over. He published the Culture Code externally in 2013, which meant HubSpot's own employees, competitors, and critics could all quote it back at the company. And he kept revising it: the version HubSpot hosts publicly today is tagged V41, dated September 29, 2025, meaning the document has gone through roughly as many public revisions as most companies give their pricing page.

That's the same instinct Marty Cagan describes in product organizations that actually work: an artifact only stays useful if someone treats it as a living thing with a customer, not a deliverable that ships once and gets forgotten. Shah applied that discipline to a values document instead of a product roadmap, which is a stranger and harder thing to do, because the customer, in this case, is the entire employee base, and every revision is visible to people who already have opinions about the last one.

Publishing it externally also created a cost most executives wouldn't accept: a public yardstick. When critics later argued the company didn't live up to its own deck, the deck is what let them make the argument at all. Shah kept the document public anyway, rather than pulling it or letting it quietly expire.

For you: if you publish your culture document externally, budget for the fact that it becomes evidence against you the moment reality drifts from it. Publish it anyway if you're willing to revise it as often as you'd revise a product feature, and pull it or rewrite it fast if you're not willing to keep it current. A stale culture document is worse than none.

3. Betting his second act on AI agents while still holding the CTO title

By 2025, Shah had spent nearly two decades as HubSpot's CTO, with a public company's resources and a hired product organization at his disposal. Instead of directing that organization toward his next big idea, he built agent.ai largely outside it, running it through a separate LLC staffed with freelancers rather than HubSpot employees. His own explanation, reported by the Boston Globe, was structural: "How do you experiment? How do you try things without distracting the core team?" By late January 2025 the platform had drawn 230,000 registered users and 250,000 newsletter signups; the standalone simple.ai newsletter he writes alongside it had grown past 2 million subscribers by the time this profile was written.

The decision to keep the CTO title while building a second, largely independent platform is a specific kind of bet. It says Shah still believes the most useful place for his personal experimentation is adjacent to HubSpot rather than inside a competitor or a fully separate venture, but that the experimentation itself needs to happen outside HubSpot's own product cycle to move at the speed he wants. That's a harder needle to thread than either fully committing to a side venture or fully directing company resources at the idea.

What this shows: a long tenure in one seat doesn't have to mean creative stagnation. Shah's model for staying relevant after two decades wasn't asking his employer to fund his next idea. It was building it small, in public, with a structure explicitly designed to avoid becoming a distraction to the team he still runs.

What Dharmesh Shah Would Do in Your Role

Apply Shah's approach by clarifying role ownership, maintaining the culture document and separating experiments from the core roadmap.

Keep Culture Accountable: accepting feedback and adding a revision tab.

Situation	Shah's Move
You're a technical co-founder feeling pressure to "grow into" the CEO role	Stay in the seat where your judgment compounds fastest, and let your co-founder own the title that matches their strength instead.
Your culture document hasn't been touched since it was written	Assign it an owner and a version number. If nobody can name the last time it changed, it's already stale.
You want to experiment with a new technology but your company's roadmap can't absorb the risk	Build it adjacent, on a separate structure, with people who aren't your core team, before asking the company to adopt it.
Your team quotes your stated values back at you when you fall short	Treat that as the system working, not as an attack. A values document nobody can hold you to isn't doing its job.

If you're a technical co-founder, Shah's clearest lesson is that the CEO title isn't the only way to keep your name on the company's biggest decisions. He kept veto-level influence over HubSpot's culture and technical direction for two decades without ever running the P&L. That only works if you and your co-founder have an explicit, durable agreement about who owns what, because an org chart alone won't protect a CTO's standing once headcount and investor pressure start pulling authority toward the CEO by default.

If you're a people or culture leader, the operational takeaway is narrower than "publish your values." It's: assign an owner, a review cadence, and a version number to whatever culture document you already have. Shah's Culture Code works precisely because someone, him, personally, keeps revising it against what's actually true at HubSpot today rather than what was true in 2013. A values PDF nobody has opened since legal approved it is not a culture document. It's an artifact.

If you're weighing a side project against your day job, Shah's structure is worth copying literally: a separate legal entity, contractors instead of your own team, and an explicit answer to why this can't just be a roadmap item at your current company. Drew Houston stayed in one CEO seat for 17 years at Dropbox by contrast, which is its own kind of discipline; Shah's discipline is compartmentalizing a second act so it doesn't compete with the first one for resources or attention.

Notable Quotes & Lessons Beyond the Boardroom

Shah's most quoted line about his own priorities came in an interview years after the Culture Code shipped: "of all the code I've ever written, I'll say this, that's likely been the most impactful". Coming from a CTO, that's a specific kind of humility: an admission that a slide deck about how people should treat each other mattered more than any system he shipped.

His writing on OnStartups.com supplies the clearest statement of the underlying philosophy: "culture is a product you build, and your people are its customers," paired with the warning that "culture debt is much more insidious than technology debt" because, unlike a bug, a bad culture hire's effects don't fully disappear once you remove them.

The origin story he tells about the Culture Code itself is less triumphant than the deck's reputation suggests. By his own account, the assignment came from Halligan, returning from a CEO forum where peers had pushed the importance of culture and telling Shah, in effect, to go figure it out. Shah's answer wasn't a workshop or an off-site. It was research: studying what actually predicted whether someone succeeded at HubSpot, then writing it down the way he'd document a specification.

That instinct extends to how he talks about agent.ai. Asked why he built it as a side structure instead of a HubSpot initiative, his answer was pragmatic rather than visionary: "How do you experiment? How do you try things without distracting the core team?" It's the same question a good engineering leader asks about a risky refactor, applied to his own career.

Where This Style Breaks

A published, versioned culture document is also a permanent record that critics can quote against you, and HubSpot has lived that risk directly. Dan Lyons' 2016 book Disrupted described his year at HubSpot in terms the company never fully rebutted, a "frat house" atmosphere and what he called a "digital sweatshop" with thin job security, and aimed a large share of that criticism directly at the Culture Code Shah had written, calling it propaganda that convinced young employees they were changing the world while enriching Shah and HubSpot's other investors. He also mocked the company's use of the word "graduated" as a euphemism for being fired.

HubSpot's founders answered on LinkedIn, citing a #4 ranking on Glassdoor's best-places-to-work list and defending the product, but the response left several of Lyons' sharper claims, including his ageism allegations, largely unaddressed. Even a reviewer broadly sympathetic to HubSpot's side of the story concluded Lyons remained an unreliable narrator while still landing real questions about inflated corporate-purpose rhetoric. Neither side's account fully settles the dispute, which is itself the point: once you publish the standard, you don't get to control whether people believe you're meeting it.

There's a second, quieter risk in staying CTO for twenty years rather than ever running the company: succession depth. HubSpot has changed CEOs once (Halligan to Rangan, 2021) and changed board chairs once (Halligan to Lorrie Norrington, 2025), but Shah's own role has never had to prove it can survive a transition, because he's never left it. A leadership model built around one person's two-decade tenure in a single seat is unproven, by definition, until that person eventually exits.

The second-act model carries its own tension too. Building agent.ai on a separate structure protects HubSpot's roadmap from distraction, but it also means a sitting public-company CTO is spending meaningful personal attention on a venture that isn't HubSpot's, funded and staffed outside the company he's paid to help run. That arrangement depends entirely on trust between Shah and HubSpot's board holding for as long as agent.ai keeps growing.

Learn More

Frequently Asked Questions about Dharmesh Shah's Leadership

Who is Dharmesh Shah?

Dharmesh Shah is the co-founder and Chief Technology Officer of HubSpot, a role he has held since co-founding the company with Brian Halligan in June 2006. He also writes the OnStartups.com blog, builds the AI agent marketplace agent.ai, and publishes the simple.ai newsletter.

What is Dharmesh Shah's role at HubSpot today?

He is Co-Founder, CTO, the title listed on HubSpot's own leadership page. Yamini Rangan is CEO, a role she took over from Brian Halligan in 2021. Halligan later stepped down from his executive chairperson role in May 2025 to become a non-employee board member.

What is the HubSpot Culture Code?

The Culture Code is a slide deck Shah wrote in 2013 describing HubSpot's actual operating principles rather than generic values. It has logged more than 5 million views, and HubSpot still revises it publicly: the version on HubSpot's own site is tagged V41 and dated September 29, 2025.

Why didn't Dharmesh Shah become HubSpot's CEO?

Shah chose to stay in a technical leadership role rather than take the CEO title Brian Halligan held from founding until 2021. That division of labor, Halligan running go-to-market and the company, Shah owning product, engineering, and culture, held for nearly two decades without either co-founder displacing the other.

What is agent.ai?

agent.ai is a marketplace of AI agents that Shah built outside HubSpot's core product organization, using a separate legal structure and freelance contractors so he could experiment without pulling focus from HubSpot's team. It had drawn 230,000 registered users by January 2025, according to reporting from the Boston Globe.

What criticism has HubSpot's culture faced?

Former employee Dan Lyons' 2016 book Disrupted described a "frat house" atmosphere and called HubSpot a "digital sweatshop," and argued the Culture Code itself functioned as propaganda. HubSpot's founders responded publicly, citing a #4 Glassdoor ranking, but left several of Lyons' specific claims unaddressed.

About the author

Victor Hoang

Co-Founder, Rework.com

Victor Hoang is Co-Founder and CMO of Rework. He spent 12+ years scaling B2B SaaS growth, building a lead engine that generated over 1 million leads and $10M+ in annual recurring revenue. Today he builds AI agents and MCP servers into Rework's products to empower customers across growth and operations. He writes about what actually works.

View full profile LinkedIn