As Armory continues to scale, we're crafting a "Roundabout, not Stoplight" culture because of Conway's law:
I spend my days interacting with Global 2,000 CIOs who are trying to figure out how to innovate faster. But the hard truth about unlocking innovation is that:
- Innovation stems from learning.
- Learning stems from failure.
- And failure means breakage of customer trust.
- Global 2,000 enterprises have an allergic reaction to breakage of customer trust.
- Ergo, an aversion to failure leads to a lack of innovation, which is then replicated across a company's systems, products & services due to Conway's Law.
How do we ensure Armory doesn't fall into this trap? And more broadly, how do we help Global 2,000 enterprise customers out of it? I believe the answer has two parts:
1) Learn to manage risk in more sophisticated ways by investing in a modern software delivery platform (like Armory's) which helps limit the blast radius when nuking the customer base due to code changes, which I write more about in this blog post.
2) Make experimentation "safe to try" at the edges of your organization: Creating a Roundabout culture enables this, and that's what this blog post is about.
From the very beginning of Armory, we've put a lot of focus on our culture. My co-founder Ben described v1 of "Life at Armory" on Facebook back in 2016 when we first started Armory.
We believe that the culture of a company is its operating system, and the OS is created jointly by all the people at the company.
I often hear founders and investors talk about the importance of culture, but it's always in very abstract ways, like "Culture eats Strategy for breakfast" (which is very real) but not very actionable.
I also see investors choosing to make or pass on investments because of a company's culture. For example, one seed investor pinged me recently, asking:
How well do you know [founder]? Can he run a company or even harder - rebuild one already running from scratch? It's a pretty big challenge.
He could really use the counsel and guidance of someone like you, but I don’t know if it’s time you want to dedicate or spend.
Yeah... I think EVERYTHING checks out but him.
What this investor was really saying to me was "I don't know that this CEO can build a company that will succeed, even though the technology is really good." He's unsure about the CEOs ability to create a successful Company Operating System.
So there's this incredibly important thing called "culture" that defines a company's operating system, defines which people are attracted or repelled to it, and defines how investable the company is, but at the same time has proven very nebulous.
And it's even more complicated because there's no one right answer. Facebook, Google, Amazon and Apple are all successful companies with incredibly different cultures.
And yet, it's obvious that some cultures are better than others -- especially as a company goes through different stages of growth. Uber's rule-breaking culture arguably served it very well to create an entire category and re-shape regulations governing the transportation industry, yet had a very, very negative underbelly to it that is now forcing Uber to completely reinvent itself from the inside out.
And there's another vexing element to culture, which is that the stronger it is, the more it's just lived, and the less it's typically documented and intentionally crafted. For example, each of our families had its own culture growing up. It's likely that you didn't write down that the culture was, but because it was so strong, you intuitively knew how to interact with your family members in the appropriate way for your family. Religions are also excellent examples of powerful cultures. Throughout history, humans have gone to war to defend their religious beliefs and the underlying traditions and cultures. In fact, opposing cultures (or those where members of one culture don't understand why other cultures act differently) is easily a top source of human conflict throughout time.
So there's no question that culture is a foundational, bedrock element that influences everything else around it. Like my co-founder Ben said early on, "Culture is what happens when the leaders are out of the room."
How, then, does one build a company with an intentional culture that scales while enabling continued innovation at it grows? And how do you answer questions like:
- How much defined process should a company have?
- How much should each employee be trusted with the company's money?
- How does a company ensure that each employee shares the same culture?
And many more fundamental questions that flow from this topic.
For the first time in my life, I've found a framework to help answer these questions in an actionable way.
It was Sue, our Director of BizOps, who found this framework, in the form of a book that was just published, called Brave New Work. And it's incredible.
Brave New Work provides a 12 step program, as it were, to examine and iterate on a company's culture. It's a process I highly recommend every founder do as early in the company's life as you feel you have bandwidth to do so because it'll never be as easy as it was yesterday. Every day the company grows and adds people is a day that the culture is being defined for you, whether you're intentional about it or not.
Here are two specific examples of how Brave New Work is such a mind-bender:
1) I've always wrestled with how defined to make processes at a company. One the one hand, smart people learn from each other, and part of the way they do that is by writing down what works, and then iterating on it. As I've often said, "Process is just writing down what works."
On the other hand, the more defined process a company has, the less flexible it becomes. It has more red tape and becomes more bureaucratic as it grows. This takes an incredible toll on a company as it scales, and keeps companies from re-inventing themselves.
Aaron Dignan, the author, illustrates this perfectly with his description of complicated vs. complex systems.
- A complicated system is like an engine or a watch. It's hard to figure out and requires deep engineering, but with enough effort, it can be solved for.
- A complex system is like traffic or the weather. One can have an illusion of control, but all the variables can't be fully controlled because it's a living system.
Aaron argues that modern companies (and especially startups) are complex, and not complicated, systems. Managers think "this reorg is finally right!" but in truth, they are trying to control a system that can, in fact, only be influenced.
Using this approach, the best companies do not institute endless processes & procedures to define their success, but rather, create a culture that pushes control and decision making to the edges of the organization and institute "Minimum Viable Policies".
And in fact, most industrial-age companies do a really poor job at this, because it used to be true that the output of line workers making widgets on a manufacturing assembly line could be measured and optimized through structure and process.
However, now that we live in a knowledge economy and are entering the Third Industrial Revolution, which is based in code, it's definitely no longer true that the output of a software engineer is optimized by forcing her to work in a specific, codified way that removes her ability to optimize her own output.
Optimizing for output has gone from complicated to complex, yet leaders at most Global 2,000 companies are stuck in a sub-optimal command and control structure.
2) So, how does one optimize a complex system? Aaron goes on to illustrate this using the example of the stoplight vs. the roundabout.
As he describes in his book:
He writes "What stands out for you about these two approaches? Did you notice that the signal-controlled intersection doesn’t require much thinking? Only compliance. The roundabout, on the other hand, requires drivers to be present and responsible for their own safety and the safety of others."
The lesson is this: When we stop telling people when to stop/go/turn and instead give them the ability to merge into traffic on their own, they also have the responsibility of avoiding collisions on their own. The most mind-blowing aspect of this change is what the data shows. Roundabouts:
- Reduce injury collisions by 75 percent
- Reduce fatal collisions by 90 percent.
- Result in an 89 percent reduction in delays.
- Are $5,000 to $10,000 cheaper to maintain each year.
- Work normally during power outages.
The book does an excellent job of outlining actionable steps a company can take to make their company more of a "Roundabout" environment and less of a "Stoplight" environment, with a series of thought-provoking suggestions like:
- Replace “Is it perfect?” with “Is it safe to try?”
- Give everyone the freedom to choose when, where, and how they work.
- Use the concept of a waterline to create guardrails around team and individual autonomy.
- Start distributing authority to the edge of the organization.
- Distinguish between reversible and irreversible decisions and treat them differently.
- Abandon singular job titles and descriptions for modular roles and role mixes.
- Allow people to hold multiple roles on multiple teams.
- Move from static to dynamic teaming—a marketplace of roles, teams, and projects.
- Move from an annual budget to a dynamic budget.
- Eliminate all status updates, project reviews, and other bureaucratic theater.
- Make org and team financials transparent and accessible.
- Make compensation transparent to everyone in the organization.
These are just a few of many, many "Company OS" suggestions the book recommends considering.
Armory is already living these principles, such as full internal comp transparency, and full internal financial transparency.
We "default to transparency." That doesn't mean that everything has to be transparent, but rather that we start with an assumption a piece of data will be transparent, and we have to work to justify permissioning it to specific audiences.
I find just "flipping the bit" to make transparency the default really helps drive the creation of a Roundabout culture, because it forces conversations around why something shouldn't be transparent.
For example, we have a one-sentence expense policy:
When you spend the company's money, do it in a way that's best for the company.
And then we have an "expense notices" channel where each tribal at Armory can see how we're spending our collective funds:
(PS – shout out to OneGraph for making our ExpenseBot possible!)
And that gives us an opportunity to celebrate pro-tips we can each use to optimize spend, like this:
If you don't have the bandwidth to read the full book, you can watch this 42-minute video which does a good job of summarizing its approach to being "complexity-conscious and people-forward" about crafting your company's operating system:
I'll be happy to share more about Armory's journey as we focus on scaling our company and culture as we grow. Let me know what you'd like to hear more about.
And here's the meta point kicker:
Due to Conway's Law, Armory has to live these cultural values ourselves to successfully build a software platform that enables Global 2000 enterprises to ship better software, faster. Want to dig in more? VentureBeat recently published this article on the topic.