Amazon's small-team structure means new projects get greenlit quickly and it's one of the secrets to the company's success. Here's how it works.

• Dr. Robin Gaster is the CEO of Incumetrics and a visiting scholar at George Washington University Institute for Public Policy.
• The following is an excerpt from his book, "Behemoth - Amazon Rising: Power and Seduction in the Age of Amazon."
• In it, he reveals how Amazon's structure allows for the quick formation of small, agile teams.

The following is a book excerpt.

Amazon is famously organized into small "two-pizza" (8 -10 person) teams. This directly adds flexibility and agility. Small teams can move quickly.

Amazon's two-pizza teams are agile, developed flexible inter-team structures, offer clarity or purpose, and are fast to innovate. They are also highly autonomous: the Prime Now team was able to launch its pilot on the iPhone first even though Amazon is generally an Android company. When the workload for a team becomes too big, Amazon may decide to break it into multiple smaller teams to keep their agility.

But teams mostly form because an employee sees an opportunity - a big one like Prime Now or one of the countless smaller ones. That employee is free to seek support and sponsorship inside or outside his own current team.

How that happens at Amazon seems quite different from most companies, which have clearly defined chains of command. For an idea to be implemented, it must move up the chain of command to the point where management is senior enough to act. If a decision involves financial resources or legal issues, those accountants and lawyers will need to be consulted. And all the way along, there are endless opportunities for vetoes. Often, one veto kills the project.

In contrast, Amazon offers "multiple paths to yes," as Bezos puts it.

Individuals can go outside their team or even their division to find a team interested in an idea. Teams can find any number of senior managers to sponsor and support a project.

Amazon has sometimes been called the "company of 100 CEOs," because there are so many possible pathways to greenlighting a project. It's not that Amazon makes innovation easy. It's never easy. But Amazon's structure is designed to get to yes.

Think of these proto-teams - ideas without even a team yet - as the seeds for 1,000 flowers that many eventually bloom. Amazon is diligent about fertilizing the soil and watering them with resources, in the sure knowledge that many - perhaps most - will fail.

These proto-teams - often even a single employee - are watered with enough resources to initially test whether there is something worth pursuing; that may be time off task to work on the idea, some hours or days from other team members, possibly other resources.

This period is pre-pilot. As described above, innovators start by working backwards to develop a clear vision of what the project is for, who it will serve, and why specifically customers will benefit. As it comes into focus, and more evidence is added that the project is worth pursuing, it may attract more resources from inside and outside the original team. That might include part-time help from other areas like logistics or human resources or advertising.

At this stage, the project is adding resources with dotted lines from existing structures. An HR person may be helping to acquire talent, but they remain within their existing HR team. Further information resources are easily available because of Amazon's previous decision to manage information flows via API; emerging teams can access information without needing permission.

At some point, the project/team gets the greenlight to move into pilot production. That will require significantly more resources, and the creation of a more focused team. Neil Ackerman explains that when he was starting to build a team for his Small and Light Fulfillment project, he searched for the team members he would need through the internal Amazon directory, and reached out to them for "many many meetings for coffee" over weeks of effort. Of the seven people he eventually asked to join the new team, three accepted, and he hired four from outside.

As the project becomes more successful, its place in the overall Amazon hierarchy becomes more settled.

It is no longer a pilot on an experiment; it's now a service within existing services. So Ackerman's Small and Light project became embedded as a successful tool for addressing the needs of the Marketplace. Over time, it attracted further use from Amazon Retail, and it's now a permanent service within the Amazon logistics network, reporting up through that chain of command, and funded through the standard OP1 mechanism. As projects solidify, teams become fully independent entities within the Amazon ecosystem. They have their own marketing, sales, engineering, and finance functions, so each product has its own profit-and-loss statement, and thus each has both autonomy and accountability.

So what's different here, and what differences does it make?

The main difference is that Amazon's internal environment is set up to encourage the constant formation of teams, and provides the critical early resources that let them flourish quickly. Management expects and even demands innovation, and the structure is set up to provide enough fertilizer at the earliest stages, sufficient support during the prototyping and testing phase, and enough clear pathways to final adoption. There are far fewer vetoes, and more widely distributed resources and pathways. And that matters. As Benedict Evans observes,

The structural advantage of them [teams], in Amazon at least (and in theory, at least) is that you can multiply them. You can add new product lines without adding new internal structure or direct reports, and you can add them without meetings and projects and process in the logistics and ecommerce platforms. You don't (in theory) need to fly to Seattle and schedule a bunch of meetings to get people to implement support for launching makeup in Italy, or persuade anyone to add things to their roadmap. This means not so much that products on Amazon are commodities (this is obvious) but that product categories on Amazon are commodities. -Benedict Evans

Of course, this picture is somewhat idealized. Small teams also compete for resources and projects, so from another perspective Amazon has created a Darwinian environment in which teams compete vigorously (sometimes viciously) on behalf of their projects and initiatives. Not everything gets funded, so there are naturally winners and losers. This hyper-competitive environment is another way Amazon pushes teams to innovate.

One final note. Small teams are only possible because Amazon has, since its earliest days, required that all teams share information electronically not only with other Amazon teams, but with outsiders as well. This key requirement makes it possible for hundreds of teams to connect - otherwise they would drown in information overhead, the sheer time and resource cost of communicating by manual methods like email and messaging.

Excerpt from Behemoth - Amazon Rising: Power and Seduction in the Age of Amazon reprinted with permission by Dr. Robin Gaster.

Dr. Robin Gaster is the CEO of Incumetrics and a visiting scholar at George Washington University Institute for Public Policy. He is the author of Behemoth - Amazon Rising: Power and Seduction in the Age of Amazon.