Amazon is one of the world’s largest companies and Jeff Bezosits founder and former CEO, one of the richest people. In a company like Amazon, productivity, Talent management and the synchronization of the equipment is very important and that everything flows in the best possible way a rule as curious as useful and interesting is responsible: that of the two -pizz teams.
The “problem” of the scale. As Explain Daniel Slater, World Innovation Culture Director at Amazon Aws, When Amazon began It was structured “to achieve rapid agility and offer constant value to customers.” Amazon was, at that time, an “inventing machine.” That is why hiring and organizational structure were prioritized. The problem is that Amazon, which started as an electronic commerce, climbed.
When the company was smaller, the teams “were more centralized and closely linked”, but as the firm began to grow this structure resulted in greater slowness and inefficiency. To face it “we radically change our technical architecture to what is known as Microservice architecture“, explains Slater.” We discourage our monolithic architecture in a vast network of unique and independent services “, which allowed to launch news and offers faster.
But that’s not all. According to Slater, to maximize the ability to remain close to users and their needs, at the same time that new products and services are launched, Amazon invented a new way of organizing the equipment. Instead of large teams, the firm opted for smaller and agile teams capable of taking advantage of the architecture of microservices. For this, a fairly peculiar rule was used.
The teams of two pizzas. The concept is quite simple and is that no team should be so large that more than two pizzas are needed to feed it. Not because A Amazon make a hole in accounting Buy a median of three more ingredients in the nearest Domino’s, but because “the smallest teams minimize the lines of communication and reduce the general expenses of bureaucracy and decision making.”
It is assumed that with Two pizzas They can eat about ten people, although that is an opinion that would be willing to fight with nails and teeth. In any case, “the two -pizz teams” is a somewhat searched form to refer to teams of less than ten people. This size “allows two -pizz teams to spend more time focusing on their customers and constantly experiencing and innovating for their benefit,” says Slater.


Does the team need three pizzas to eat? Well, then it’s too big | Image: Pexels
The advantages of small equipment. As they expose from Amazon, small teams increase “involvement and training” and reduce the known as Ringelmann effect (Individual productivity is reduced as the group becomes larger, and vice versa).
Citing the Hackman and Vidmar studyAmazon says that the smallest teams “also increase employee satisfaction” since “in large teams, individual contributions are less recognizable and individual involvement on specific areas becomes more diffuse.” These small teams, say from Amazon, also allow to experience faster and reduce the cost of failure.
A TEAM, A PRODUCT. How do you implement these “two -pizzas teams”? Easy: a equipment for each product or service. In the words of Daniel Slater:
“Instead of maintaining complex systems or solving problems that cover multiple services, business lines or customer segments, two pizzas equipment focus on a single service or offer, and only on customers that use it. This unique approach favors effectiveness and scalability.”
More pizzas. If the team grows, something likely if the needs increase, “we look for ways to divide it into separate equipment of two pizzas that work in a sub -section of the service with a single subprocess.” However, from Amazon they are aware that this structure is not for everyone and has its counterpart.
“Two -pizz teams are not perfect, and are not immune to limitations; an example is that with so many autonomous teams, of a single process, which They work quickly To meet the needs of their own clients, there is the risk of duplication and development in silos, “says Slater. It is also necessary to collaborate between teams so that rapid experimentation is effective, as well as for Share errors and learning.
Image | Flickr (Smithsonian Institution), Pexels
*An earlier version of this article was published in July 2025
GIPHY App Key not set. Please check settings