Over the six years Ajit Varma was an “ad spam czar” at Google, he found himself getting into a routine: he would pour himself into an idea–and then have to gather opinions that would allow it to move forward.
“What happened to me was I spent lots of time presenting to execs in the company,” he says, “and they would basically change the direction of the product with 30 minutes of context.”
This, he says, was because the execs in question didn’t have a chance to think deeply about the product–which generates a “trend towards the mean.”
Then, in 2010, Varma started a data targeting company, Adku. There, naturally, he was part of a small team. It was a fantastic situation, he says, since they could iterate–gather information and change their product–free of the time sunk into intra-organizational persuasion he’d known before. But there was another problem: the startup didn’t reach the level of adoption to have the quality of data to allow them to iterate as fast they wanted to.
Then in April 2012 Varma joined Square and was soon leading the Square Marketplace team. From Varma’s description, it’s the organizational equivalent of Goldilocks finding the porridge that’s just right: at over 600 employees, $15 billion in payments processed annually, and millions of sellers onboard, Square is more mature company than startup. Yet many project teams are small, capped at 12 members, and also diverse. According to engineering lead Matthew O’Connor, they’re “full stack”: each team is responsible for developing their code, formulating their product, testing their code, deploying it, and running it in production.
Why does this matter? Because the structure and composition of an organization shapes the products it creates. Why? Because the structure–like whether or not there are outside managers to convince of a decision–and the composition–whether or not all the skillsets are present to make a decision–affect the rate at which decisions may be made. So if a company is optimizing for responsiveness, as opposed to efficiency–as Eric Ries and the Lean Startup would suggest–they’ll want to form their organizational structure accordingly. Which is what Square purports to do, as O’Connor, who was employee 14, explains.
“The original thought behind doing this, even when we were really small and the entire company was essentially the size of an individual team, was that we just knew that when we got bigger that this was going to be necessary for us to move fast,” he says. “And many of us had seen at previous companies what happens when you sort of divorce some of that responsibility and teams start having to interact with 10 or 20 other teams to be able to get their job done.”
So instead of getting divorced, Square made clever prenuptials.
Organism and organization have the same root: organ, from the Greek organon “implement, tool for making or doing.” Both an organism and an organization are systems of tools for doing things. So it follows, then, that organizations can learn from organisms, as Square’s full-stack ethos implicitly does, as the full-stack complies with the ecological insight into requisite variety: the idea that an organism needs to be as complex as its environment in order to operate effectively.
In other words, says University of North Carolina management professor Dave Hofmann, a set of complex information–like the kind that teams at Square are faced with–requires a complex set of knowledge:
The information that’s coming into Square is complex from a design perspective, from a product perspective, from an engineering perspective, from a marketing perspective. And so the organization is putting together those different specialized functions in basically an organism, that lower level team, such that the information comes in and all of the requisite knowledge required to effectively respond and process that information is right there within the team. So it happens much more quickly than in a traditional organization where information comes in and it has to be segmented out to the engineering department, the marketing department, and more.
In this way, as the knowledge (that’s the data) comes into the organism (that’s the full-stack team), the organism already has the range of expertise–in developers, designers, and project managers–to make quality decisions quickly.
Square Cash, the recently launched product that lets you email money, is representative of the responsiveness that the small team structure engenders. As Brian Grassadonia, the leader of the Square Cash team explained to us, the product began with three people and a thesis.
“We had a thesis that building a product on top of a foundational platform like email, using just the tool that people have in their pocket, primarily their phone, and their debit card, was very compelling,” he says.
Since they were such a tiny team, he says, a prototype of Square Cash got released within the company–and all of a sudden a product was born. They added another engineer, recruited some design talent, and soon got the team up to a dozen–but eschewed any further growth.
Why? Because, as Grassadonia has found, the more people you have on a team, the more relationships have to be navigated, which can lead to what he calls “decision atrophy,” that feeling of not wanting to make a decision because the need to be “getting buy-in from dozens of people and getting approvals three rungs up the ladder.” That decision atrophy, he says, leads to people making the easiest, lowest-leverage decisions while avoiding the more difficult ones. But by staying small and streamlined, decision atrophy is prevented.
“If you would’ve had a really, really large team, it would’ve been hard to kind of get everybody on board that an application was the right way to go,” he says. “We had been kind of invested in the email platform, and some people wanted to stay extremely true to that.”
But the small team size allowed everybody on the team to be involved with the decision–and thus let the product evolve. Like any good organism.
Hofmann, the management scholar, says that Square represents the far swing of the pendulum between “local customization and quickness” and “centralized efficiencies.” The question for the company, as it becomes bigger, is if they want to look across these teams and start finding best practices that can be communicated between teams, spreading the knowledge that years of running these full-stacks has generated.
“Now that’s going to smack of bureaucracy,” he says. “But some bureaucracy can actually be beneficial and help teams from having to re-create the wheel.”