Policy is vegetables for culture
Or, how I learned to stop worrying and love (good) policy
Don’t miss Jen Pahlka’s excellent article “Culture Eats Policy” and her conversation on Ezra Klein’s podcast picking up on topics from her well-reviewed book Recoding America (June 2023, Metropolitan Press).
Sure, “Culture Eats Policy” is punchy framing for a problem particularly rife in (but definitely not limited to) large organizations such as the government. And I would be the first to say that policy alone, without the right culture, is at best ineffective and at worst an active problem. But like a lot of simple models, the “culture eats policy” model is useful but very, very incomplete.
tl;dr
I’m here to argue that we should frame policy and culture differently in relation to each other, not in terms of relative superiority. Policy and culture have a dynamic and mutually supporting relationship. In some weird world where we have to choose one and only one of the two, sure: we will all pick culture. But in our world — the real world — we need both! Policy makes culture concrete, it reduces the cognitive load of learning and living a culture, and perhaps most importantly, with observation, it can continuously test a culture against the objectives of an organization so that it may be refined and improved. As such, policies can be used to create and nurture the culture that we want in our organizations.
To survive and thrive, culture needs policy.
Definition
If you ask four people what culture means, you’ll get six correct answers. That’s just how it is. So let’s establish some succinct working definitions before we go further.
Culture refers to the shared values and norms that guide the behavior of a group of individuals.
Policies refer to the formal rules and guidelines that dictate day-to-day work in an organization.
Making culture concrete
Cultural values can often be abstract and vague. Policies transform these into practical guidelines for day-to-day work.
Stripe says:
Seek feedback. We value intellectual honesty and look for other Stripes with the expertise to refine our ideas, challenge us, and deepen our understanding across the business.
The statement has two critical parts. “Seek feedback” is a succinct statement of policy. It is immediately followed by the cultural value that is reinforced by this policy, the value of intellectual honesty.
So when someone asks “so what does it mean to value intellectual honesty?” you can explain. It means you actively seek critiques of your ideas. It means experts are expected to spend time helping others refine their ideas. It means everyone needs to be comfortable getting their ideas challenged and challenging others’ ideas. (I’ve never worked at Stripe, maybe a Stripe will challenge me on this understanding.)
To take a couple of even more concrete examples, a policy like a daily customer complaint review concretizes a culture of valuing customer satisfaction. A policy like weekly spec reviews concretizes a culture of seeking and giving feedback.
Reducing cognitive load
Let’s continue with the example of weekly spec reviews. When a new team member joins, they’ve sat in the orientation and heard about the “intellectual honesty” value. But neither they nor their onboarding buddies need to think about how to embody the value from first principles. They simply adopt the practice of weekly spec reviews and they are done. This is a trivial and extremely common example where a policy has reduced the cognitive load of determining how to live a value.
Coming at this more broadly, I really liked this phrase from an Oso blog:
Opinion abstracts complexity
We can adopt a way of working — the “opinion” — without necessarily having to understand every nuance of why we work that way. Hordes of us in software engineering use ideas from Lean Manufacturing without having to know the origin of the word kanban. I can move my Trello cards from “in progress” to “done” all day long without once thinking about how my actions are reflecting a “we reduce waste” value.
More narrowly but still important is that using the word “opinion” suggests that we can think of a policy as being one possible instantiation of a cultural value, rather than “the” instantiation.
Continuously testing culture
This is a more subtle point to make, so we’ll dwell on this a bit.
Let’s go back to Jen Pahlka and “Culture Eats Policy”. Its obvious to note that the policies she refers to all supported something that was known to work well. Providing guidance that software should be built using “service-oriented architectures” is hard to argue against even today. The problem came from how the policy was written, interpreted, communicated, and enforced. That is to say, how “use service-oriented architectures” became “Thou Shalt Use An Enterprise Service Bus”.
The solution, of course, is not to insert more legalese into the policy to account for every possible misinterpretation. Rather, the core thing we need to identify is the desired outcome of the policy, such that the outcome supports the value of providing better service to the American public.
Imagine there was a section in the policy that explained the objective. E.g.,
We believe that service-oriented architectures are the best way to ensure that our services remain interoperable and composable, leading to faster delivery and ever richer services to better service members of the public. So we require all new software capabilities to be built using service-oriented architectures, e.g., by using an Enterprise Service Bus.
Now we can periodically go back to the policy and ask several questions. Has the policy been shown to support the objective of faster delivery and ever richer services and can we measurably say it has helped members of the public? How do we know the policy and outcome are connected? Faster delivery than what, richer services than what? Does the state of the art continue to be service-oriented architectures? And so on.
(Sidebar: remember the Bezos API Mandate? Did that policy not change and support the culture of Amazon?)
A brilliant example of this is the Agile Manifesto. Consider the first principle of the manifesto: Our highest priority is to satisfy the customer through early and continuous delivery of valuable software. Well, it says it right there. Our objective is to satisfy the customer (the cultural value here is something like “customer first”). Our policy requires early and continuous delivery of valuable software. We can now periodically ask: does our policy support our cultural value?
Having a clear objective and continuously asking whether a policy is achieving that objective is essential to ensuring that the policy doesn’t ossify into a meaningless or even harmful ritual.
Tying it all together
Let’s say your organization has a cultural value that says “We listen to our users”. How might we support that? With a policy that says “Because we value user feedback and want to solve real problems that they face, we will run at least 10 user interviews every month and address their top 3 pain points, thereby increasing our user satisfaction rating.”
What does that policy do?
It makes the value concrete. It answers the question “what does it mean to value user feedback?” It is now very clear that the organization backs the value with time and money.
It reduces the cognitive load of following the policy. You now don’t need to think “ok, so how might I listen to users?” The policy is the answer to that question.
It makes the policy testable. Having done all this, did we increase our user satisfaction rating? No? Ok, so do we suck at doing interviews, suck at solving the problems that we identify, or are we at the point of product maturity where we need to innovate beyond what users are able to imagine — meaning, do we need to change our values to prioritize innovation over responsiveness to user needs?
(By the way, yes — you can rewrite that as an OKR or a SMART goal. That’s just an implementation detail.)
One last reference to “Culture eats Policy”. Jen Pahlka says:
Changing the culture is possible, but it’s a long slog.
You know what will make it easier? Good policy.
So start by not reflexively dismissing policy. Reflect on the culture you want and design a set of policies that supports the creation and nurturing of that culture. Regularly test whether the culture is resulting in the outcomes you want and adapt your culture as needed, using policies as the tip of the spear.
Culture eats policy? Sure, but that’s because policy is vegetables for culture. It may not be the most appealing part of the meal, but you can’t not eat your vegetables. Why? Ask any mom.