Culture as a Product
Every company builds two products: one for customers, one for the team. Culture is the second one.
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 7
- Confidence
- 95%
Shah's central insight is that culture is not an abstract value set to be preserved but a product built for the team, whose customers are the employees. Once you accept that postulate, the entire product discipline transfers: you survey your customers, you run an NPS score, you log bugs, you triage them, you fix some and mark others 'works as designed', and you ship new versions. The job is never to preserve the culture — it is to iterate it, because the needs of the team change as the company scales.
Origin
Created by Dharmesh Shah at HubSpot when co-founder Brian Halligan asked him to 'go do culture' three or four years in. Shah built the Culture Code deck (16 slides originally, later 128+) by asking what a Python truth function predicting a new hire's success would look like. NPS itself is Fred Reichheld's metric, repurposed here as an internal culture measure.
Core principles
- 01You are not creating culture — it already exists. You are articulating the culture that is there.
- 02Culture is the product you build for your team; employees are its customers.
- 03No product person ever said 'I built this, it is awesome, I am done.' Founders say exactly that about culture, and it is the number one mistake.
- 04Culture exists to help great people do great things for the company — as the company scales, what they need changes, so the culture must change.
- 05Core values are the constraints you deliberately do not iterate away from; not everything in the culture is a core value.
- 06Aspirational statements are allowed if you explicitly label them as aspirations — they become self-fulfilling as new people join and behave as if they were already true.
How to run it
- 1
Reframe the task as articulation, not creation
Assume the culture already exists and is working. Your job is to describe it accurately enough that everyone knows what it is.
Watch out The word 'culture' is deeply overloaded and triggers a visceral negative reaction in many employees who associate it with posters on walls. Expect backlash on the first email.
- 2
Write the truth function
Ask: if I could write a function that predicted how successful a new person will be here, what would the input factors be? Those attributes are the culture code.
Pro tip Shah's original Culture Code deck was exactly 16 slides answering only this one question.
- 3
Run an internal NPS survey on the culture product
Ask employees, on a 0-10 scale, how likely they are to recommend the company as a place to work, plus the open 'why did you give that answer' question. Repeat it every quarter, forever.
Watch out If the top answer is 'the other people here', that is a self-reinforcing recursive loop that tells you nothing. Dig into what makes those people who they are.
- 4
Publish every response and triage the bugs
Publish the survey responses in full (quantitative and free text), categorise the feedback, and call the problems what they are: bugs in the culture.
- 5
Ship a bug-fix release at the all-hands
At the next all-hands, do exactly what a product team does with customer feedback: here are the bugs you identified, here are the ones we are fixing and when, and here are the ones that are 'working as designed' and we are not prioritising, with the reason why.
Pro tip Being willing to say 'works as designed' is what makes the process honest rather than a suggestion box.
- 6
Mark aspirations with liner notes
Where a statement is not yet true, keep it in the deck but annotate it as an aspiration — 'this is how we want it to be, we are not walking the walk yet'. Do not present untrue things as true, and do not remove things you want to become true.
Pro tip New joiners treat the aspiration as the actual culture and behave accordingly, which makes it true over time.
- 7
Separate federal law from state law
Keep a small core set of company-wide values that groups cannot opt out of, and let individual orgs optimise the rest for their context (e.g. support needs fixed coverage hours even though the company does not track hours).
Watch out Guard the federal layer hard. Leaders will make efficiency arguments to opt out of core values like transparency — 'may be right' is not enough to override a core value.
In the wild
Shah's very first email announcing he would survey the company about culture triggered a visceral backlash — still in his top three worst professional experiences. One response read: 'Dharmesh, here we're talking about culture, next we're going to have posters up on the wall about Excellence. HubSpot is not the company that I thought I joined.' He learned the word 'culture' was overloaded by decades of companies doing it badly, backed off the framing, and instead set out to describe what was already there.
→ The resulting 16-slide Culture Code deck grew into a public, continuously iterated artefact, with quarterly culture NPS surveys still running at 7,000+ employees.
HubSpot's belief in a flat, meritocratic organisation manifested as having no titles at all — sustained up to hundreds of people. Employees argued that titles are a signalling device for life beyond HubSpot, and that by withholding them the company was implicitly having to compensate people economically in other ways. After a day-long meeting the founders considered three options: hold firm on no titles, adopt classic titles, or let people invent their own ('Grand Poobah of X').
→ They shipped classic titles, reasoning that made-up titles are equivalent to no titles because they lack a benchmark. The core value (meritocracy, flatness) survived; the buggy manifestation of it was fixed.
Common mistakes
Believing your job is to preserve the culture
Shah calls this the number one founder mistake. You would never freeze the lines of code of a product; freezing the culture guarantees it stops serving the people it exists for as the company scales.
Building the culture product without talking to its customers
You would never ship a product without customer input. Culture decided in a founders' room without employee surveys is exactly that.
Stating aspirations as if they were facts
Employees immediately spot the gap between the deck and reality, and the whole document loses credibility. Label the aspiration instead of deleting it or lying about it.
Is it for you?
Best for
Founders and people leaders at a company past ~50 people who have been asked to 'do culture' and want a concrete, iterative operating system rather than a values poster.
Not ideal for
Very early teams (under ~10 people) where the culture is simply the founders' daily behaviour and a formal survey loop is overhead.
From the transcript
“so it's not creating culture it's articulating the culture so we know what it is”
“this uh thing that culture is a product period”
“builds two products one is the product they build for their customers and the other is a product they build for their team that's what…”
“Founders make is that they say oh we have an awesome culture and my job is to preserve the culture that is not the job”
“here are the bugs in the culture”
“so it's okay to say things that are aspirational just call them out”
From the episode
Zigging vs. zagging: How HubSpot built a $30B company
Dharmesh Shah (co-founder/CTO)