.png)
The Home of Culture Research: The Business of Culture I 2026 I Download Now
What a16z Learned About Culture From 500+ Investments and the Systems You Can Use Right Away
DEEP DIVE INSIDER PROFILES
.png)
After working alongside hundreds of technology companies, a16z has developed a clear view of culture: what matters is not what a company says it values, but the decisions people make when nobody is telling them what to do.
Andreessen Horowitz has had an unusually wide view of how company cultures develop. Its public portfolio directory now includes more than 600 companies, while the firm has built an operating platform around helping founders with the parts of company-building that appear once an idea starts becoming an organisation. Talent is a significant part of that model, alongside the broader work of helping founders move from small founding teams into companies with executives, managers and increasingly complex structures.
Across that work, a16z has arrived at a fairly unsentimental definition of culture.
Ben Horowitz has argued for years that culture is not the statement on the wall or the personality of the founder. It is the collection of assumptions people use to make everyday decisions when leadership is not there to provide the answer. In a16z's own description of its culture, the principle is even more direct: “A culture is not a set of beliefs, it’s a set of actions.”
That distinction changes where a founder should look when trying to understand their culture.
The interesting question is not whether the company says it values speed. It is what happens when somebody has enough information to make a decision but their manager is unavailable. It is not whether transparency appears in the values deck, but whether uncomfortable information travels upwards quickly. Culture becomes visible in hundreds of small choices long before it appears in an employee survey.
Horowitz learned some of this through experience at Loudcloud and Opsware. He has written that he initially assumed his own behaviour and values would naturally scale through the organisation. They did not. As the company grew, different managers created different environments around themselves, producing what he later described as a mixture of largely unintended cultures.
The lesson a16z now gives founders is to make important behaviours explicit before scale makes the gaps harder to see. That does not mean writing another set of broad corporate values. The more useful exercise is deciding what a value actually requires somebody to do.
If a company says it values customer obsession, what happens when a customer request conflicts with the product roadmap? If it values urgency, when is somebody expected to make the call themselves rather than schedule another meeting? If intellectual honesty matters, what should an employee be able to say to a senior leader without being punished for it?
Specificity turns culture into something people can use.
a16z has highlighted the same idea when examining Netflix's approach. Rather than relying on words such as integrity or communication, Netflix has historically described the actual behaviours it wants managers and employees to recognise. That makes the culture useful in both directions: the company can set an expectation, and employees can point to it when leadership fails to meet the same standard.
Hiring is where those decisions start becoming expensive.
a16z's work with early-stage founders repeatedly returns to the importance of the first people a company brings in. Early hires do more than fill roles; their behaviour becomes part of what later employees learn is normal. Its talent team therefore recommends treating recruiting as a company-building function from the beginning rather than something founders solve once headcount becomes a problem.
But there is an important qualification. Hiring for “culture fit” can easily become hiring people who think and behave like the people already there. Research discussed by a16z has shown the potential downside: similarity can create cohesion early while eventually limiting the differences in thinking a growing organisation needs. A culture becomes stronger when a company is clear about the behaviours it requires without assuming that everybody displaying those behaviours needs the same background, personality or way of approaching a problem.
References can then become more useful than a final box to tick. a16z's talent team encourages founders to ask former colleagues how a candidate responds to pushback and what environment allows them to do their strongest work. Patterns across those conversations can reveal much more about how somebody will behave inside the company than another interview designed primarily to confirm what the hiring team already hopes is true.
For executive hires in particular, a16z recommends creating the bones of a 30-, 60- and 90-day plan before the person even starts. Onboarding begins during the hiring process, with the new leader building relationships and understanding the context around the company before being expected to change it.
That is culture in a very practical form. Instead of telling somebody to “take ownership”, the organisation gives them enough information to understand what they own.
Clarity around ownership becomes even more important as the company grows.
Horowitz has written about “management debt”: organisational decisions that solve an immediate problem while creating confusion that compounds later. One example is dividing a function between two leaders without making it clear who ultimately owns the outcome. Employees then have to work out whose decision counts, while accountability becomes harder to locate when something goes wrong.
The cultural consequence is significant. A company may tell people to move quickly while designing an organisation in which nobody knows who has authority to make the decision.
a16z's wider company-building work increasingly points towards pushing useful decisions closer to the people with the information to make them. That only works when the surrounding system gives those people enough context and makes accountability clear afterwards.
This is also why culture cannot simply be written once and protected forever.
Some practices that are useful at 20 people become obstacles at 200. a16z has used the example of founders personally interviewing every employee. Early on, that can be a powerful mechanism for protecting the standard of the team. Eventually, the same ritual can slow hiring badly enough that the company has to decide which part of the original practice actually mattered and find another way to preserve it.
The behaviour may need to survive even when the ritual does not.
That is one of the harder lessons behind a16z's view of culture. Strong cultures are not necessarily the ones that change least. They are the ones that understand what they are trying to preserve well enough to change the surrounding system without accidentally losing it.
It requires leaders to keep checking what is actually happening further down the organisation. a16z recommends examining whether the behaviours people experience match the culture leadership believes it has created, then looking directly at the places where the two have started to diverge.
The most useful takeaway from hundreds of companies is therefore surprisingly concrete. Culture lives inside the hiring question somebody asks, the information a new employee receives in their first month, who is allowed to make a decision, what happens after somebody makes a mistake and which behaviours actually get rewarded when promotion time arrives.
The words still matter because people need language for what the company expects. But a16z's argument is that the language only becomes culture once the organisation behaves accordingly.
For founders looking for somewhere to start, that makes the task much smaller than “build a great culture.”
Look at the decisions people make repeatedly when you are not in the room. Decide which of those behaviours will help build the company you want. Then design the hiring, management and operating systems that make those behaviours easier to repeat.
That is where culture starts becoming something a company can actually use.
