Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus convallis sem tellus, vitae egestas felis vestibule ut.

Error message details.

Reuse Permissions

Request permission to republish or redistribute SHRM content and materials.

Inside Day 1: How Amazon Uses Agile Team Structures and Adaptive Practices to Innovate on Behalf of Customers

A group of people hanging from a cloud.

In Amazon’s 2017 letter to shareholders, CEO Jeff Bezos tackled the question “What does Day 2 look like?” According to Bezos, “Day 2 is stasis. Followed by irrelevance. Followed by excruciating, painful decline. Followed by death. And that is why it is always Day 1.” Bezos’ thoughts on corporate mortality are substantiated by recent research finding that companies listed on the Fortune 500 from 2000 to 2009 had only a 63 percent chance of survival.1 The primary reasons for their failure were inflexible, non-agile business models and the lack of innovation. Thus, Bezos’ provocative shareholder letter comment was a call to all Amazon employees to stave off Amazon’s potential demise, maintain its Day 1 startup culture, and continue to launch innovative products and services. 

In fact, customer obsession and quick decision-making are key to avoiding organizational stasis and are essential for maintaining Amazon’s unique Day 1 culture. These two factors have enabled Amazon to rapidly go-to-market with its innovative products and services. Cameron and Quinn’s Competing Values Framework2 would describe Amazon’s Day 1 as an adhocracy culture type that is flexibly structured and externally focused on customers. Companies with an adhocracy culture value growth, variety, and stimulation, which encourage risk-taking, creativity, and agile ways of working. Compared to Day 1, Day 2 organizations are described as: (1) being too focused on following organizational processes even if the outcome is undesirable to the customer; (2) unable to quickly embrace external trends (e.g., cloud computing and artificial intelligence); (3) managing to proxies (e.g., using market research as a proxy for customers); and (4) making slow decisions.3 A Day 2 culture would be described by Cameron and Quinn as hierarchical, focusing internally on stability, control, and efficiency and valuing routine, formality, and consistency. See Table 1 for characteristics describing Day 1 versus Day 2 companies. 
golden fig 1.JPG
There are numerous empirical linkages between organizational culture type and organizational effectiveness, with a Day 1 culture as the most predictive of product innovation and team creativity.4 A hierarchical culture type has little success when it comes to innovative output. 

As startup organizations grow exponentially in size and complexity, how do their leaders and teams make the right decisions quickly to build customer-obsessed products and services while remaining Day 1? What is the optimal team organizational design for accelerating high-velocity decision-making and innovation? These are just some of the questions that Amazon and other high-growth startup companies face every day as they work to remain relevant and avoid organizational stasis. Having a Day 1 strategy alone is not sufficient for increasing innovation. Organizations must also have the right organizational structures and agile practices in place to facilitate customer obsessed teams and high-velocity decisions that subsequently result in risk-taking and quick implementation of creative ideas.5 

Amazon’s Agile Team Design

With rapid changes across technology, regulations, demand, and external trends, it is critical for companies to respond and adapt quickly in order to innovate faster than their peers. Companies are more likely to be successful staying close to their customers and adapting to changing conditions if they use flexible, agile team structures.5 Agile teams are small and composed of multidisciplinary employees (defined as employees in job families outside of their own and generally includes tech and non-tech employees) who break up large, complex problems into manageable modules and build solutions for each component that are then integrated into a comprehensive whole.6 Amazon has adapted a nimble “two-pizza team” structure to stay close to the customer, avoid distractions and competing priorities, and make quick decisions. The two-pizza team model is exactly what it sounds like—no team should be larger than what two pizzas can feed (about six employees). This implies that if a team is empowered to innovate and problem-solve on their own, they will be more successful in the long run. 

Various studies have supported that smaller, agile teams (usually four to six employees) innovate more and make quicker decisions7 compared to larger teams. The advantages of smaller teams can be attributed to higher levels of employee motivation and enhanced coordination.8 As teams get larger, employees feel less motivated and responsible for the output.7 Furthermore, coordination among team members increases for every member added since it requires understanding how to gather information in an organized efficient way.9 Acknowledging the downsides of large teams, the two-pizza team is designed to optimize for the most efficient work flow from a small, closely integrated group of employees. 

McKinsey & Company conducted a survey across a full range of companies, industries, and sizes to understand various characteristics and outcomes of organizations leveraging agile teams.5 More than 80 percent of respondents from agile organizations developed new products and services by having close interaction with their customers and field-testing new ideas and prototypes quickly to gather data and rapidly improve the product. Those leveraging an agile team design and practices were 1.5 times more likely to financially outperform others and 1.7 times more likely outperform their peers on nonfinancial measures.

Managing Single-Threaded Organizations with Two-Pizza Teams

A Single-Threaded Owner (STO) is a leader responsible for managing one or more two-pizza teams and protecting them from distractions. The STO approach guarantees that at least one person at Amazon will wake up every day trying to drive a single critical initiative forward without competing priorities. STOs also act as a bridge to larger organizational efforts. Amazon uses an annual strategic and operational planning process for teams to describe critical investments and goals for the upcoming year and the associated required resources. Coming out of the process, team goals are created that align to overall priorities for the organization. This process provides senior leadership visibility to proposed key investments. Because the business moves so quickly, Amazon uses weekly, monthly, and quarterly business reviews; roadmap reviews; and program reviews for STOs to audit against deliverables and update their plans. For example, individual teams will set goals that map to their development cadence (e.g., monthly, quarterly) and the STO conducts audits against these goals in the rhythm of the business. These mechanisms allow STOs to understand work required and allow independent, autonomous progress against key efforts. STOs also understand higher level organizational goals and are free to make the right tactical decisions on their own. Additional mechanisms ensure that teams are delivering the right results at a consistently high bar. 

Evaluating Amazon’s Agile Team Structures

STO and two-pizza team organizational structures are not always perfect. As a result, Amazon continuously evaluates their organizational designs and the impact of the team structures on execution speed and innovation. The Amazon Devices organization, responsible for building products such as Alexa, FireTV, Kindle, and other Amazon products, uses organizational network analysis (ONA) to inspect and audit its team structures. ONA is a way to visualize how communications, information, and decisions flow in an organization by analyzing the informal employee relationships.10 ONA results pinpoint strengths and potential opportunity areas through team collaboration and coordination. Amazon has used ONA to identify decision-making bottlenecks, unnecessary dependencies, and untapped expertise to improve decision-making speed and collaboration for innovation. When assessing innovation collaboration, Amazon used ONA to identify development connections. Development connections develop locally and refine big ideas so that they can be brought to life as prototypes, concepts, and solutions.11

A few recent ONA studies were conducted within the Amazon Devices organization to ensure teams were operating efficiently, maintaining Amazon’s Day 1 culture, and making high-velocity decisions. The purpose of the ONA was to understand how STOs make decisions and identify barriers that stand in their way of making smart and fast decisions and that lead to innovation. A short survey was administered to employees in several organizations within Amazon Devices to gather ONA data on decision-making and innovation. Sample ONA questions included: 
  • List a maximum of 10 people within the (X Devices Organization) reporting to (Y Leader) that you most frequently contact for sign off/reviews before making a decision necessary for execution.
  • List a maximum of 10 employees that you frequently turn to in order to gain input, direction, or to discuss a new or innovative idea within the (X Devices Organization) reporting to (Y Leader).
  • Which of the following has a significant impact on your team’s ability to deliver results faster? Select all that apply: (1) too many approvals; (2) difficult to align on priorities: (3) documents (emails, PR/FAQs); (4) errors/re-work; (5) meetings.
golden fig 2.JPG
Findings revealed positive insights about the STO organizational structure in which STOs were operating with little to no dependencies on other teams when developing new ideas (i.e., the intended organizational design). STOs were also acting as bridges across teams in the organization as a way to manage across organizational-wide efforts, while protecting their team (i.e., development connections) from distractions (see Figure 1). Barriers to STO teams moving fast were also identified as seen in Figure 2, with too many meetings, challenges obtaining approvals, and aligning on priorities. As a result, one specific action by all teams was to identify and eliminate unnecessary meetings and streamline approval processes, which resulted in quicker decision-making. STO leaders also met across teams to further align on priorities through their weekly business review process.
golden fig 3.JPG
The ONA data helped leaders of Amazon Devices identify a few highly utilized employees who could potentially be bottlenecks in the network as well as underleveraged employees. All of the ONA data was analyzed to load balance decision-making and collaboration for innovation development, and to build the right mechanisms to improve fast decision-making. In sum, ONA was an effective tool for assessing overall Day 1 organizational health to ensure the right informal team structures were in place to make high-velocity decisions and increase innovation development. The ONA data helped identify the right interventions to double-down on current strengths and address various opportunities to develop new innovations on behalf of customers.

Mechanisms and Adaptive Practices for Inventing and Simplifying at Amazon

Arena and Uhl-Bien12 describe how organizations become and remain adaptive through adaptive space. In the ONA findings discussed earlier, even Amazon’s agile-designed organizations require inspection and intervention to ensure employees’ creative energy is not stifled by the larger organizational system. Outside of formal organizational structures, Amazon uses various adaptive practices11 to enable “Think Big” interactions in order to address complex problems that are difficult to address without creating adaptive space. 

A signature adaptive practice Amazon uses to identify and solve customer needs is encapsulated in Amazon’s Working Backwards process. Before assigning budget to any new initiative, teams at Amazon test the idea by writing a press release, a set of FAQs, and even a product manual as if the product has already launched. The press release is a one-page narrative that uses customer-centric language to explain how the new product would look to the customer. It is written before building starts to confirm that the solution is the right thing to work on and to refine ideas. golden fig 4.JPG
With the press release, writers look into the future and imagine what they want a customer to experience and ideally, what they want them to say at launch. It is written for the customer and focuses on the customer experience. The press release also includes quotes from Amazon business leaders talking about the benefits of the new product. It takes a lot of hard work and collaboration to write a good press release, but Amazon has found it to be incredibly effective at testing ideas—building an enduring vision of the product or feature. Something really interesting happens during the Working Backwards process. The moment Amazon leaders sit down and start interacting with a press release, the idea evolves. The goal is not to get approval for the idea, but test its quality and make it better. Is it really customer-centric? Is this the right thing to pursue? Amazon is more interested in getting to the right answer than in getting buy-in on the proposal. The press release will go through multiple iterations with the goal to refine thinking and find the best customer-focused, innovative path forward. Only then does the team begin to invest in headcount or roadmap changes to build on the idea.

Amazon also uses FAQs to channel its customers and stakeholders to define how something will work and how best to execute. The press release describes the essence of the idea whereas FAQs describe the details. Solid FAQs uncover risk and help decide if a project is the right investment because big or small, every project is an investment. Virtually every product and innovation that Amazon has launched in the past 15 years—from Amazon Prime to Frustration-Free Packaging to Alexa and the Kindle e-reader—have started out with press releases and FAQs.

Like other companies, Amazon has adopted hackathons (i.e., co-labs) to combat business-as-usual thinking. Amazon hackathons vary in duration and purpose. Some run for a few hours, but the most common format is a day-long competition where people team up to invent and quickly prototype solutions to enduring problems or find novel ways to use existing systems and data. Participants then vote for their favorite solution, and awards are handed out to the winning teams. Hackathons not only generate new ideas for the hosting team, but they provide a sense of community for participants. Some example ideas developed from hackathons include a novel configuration of AWS services to prototype an automated hard-hat detection system for construction sites and an automated career growth planning tool for stronger career conversations.

While hackathons are effective for innovating on existing problems, grand scale invention often needs a different platform—one that is unconstrained by current project teams or lines of work. One of the adaptive practices Amazon has employed to foster this type of invention is a two-and-a-half day technology conference (i.e., adaptive summit). This is an annual gathering of technologists from across Amazon using Open Space Technology methodology to drive the agenda. On the first day, all participants are invited to propose sessions on any theme-related topic about which they’re passionate, with some guidance from conference organizers and senior Amazon technologists. On the second day, participants vote to converge on a subset of Day 1 topics to further problem-solve; these sessions are hosted by experienced moderators to foster quality discussions and drive the innovation process. At the end of the second day, participants vote on the top three innovations to be converted into detailed organizational-level project proposals and presented to senior leadership. 

Over the years, topics discussed at the conferences have included, “What happens when everyone has a 3D printer in their home?” and “The Age of Thought Computing.” The conference has also been the source of many new product ideas including a long-term AWS storage data archive and using computer vision in fulfillment centers—a predecessor to the Amazon Go store. 


To address the hard problems that are difficult to solve in the larger organizational system, Amazon has created adaptive space11 and used various adaptive practices.12 Some example Amazon practices include its Working Backwards process (i.e., press release, FAQs, etc.), hackathons, and adaptive summits, which result in multiple creative ideas evolving into customer-focused, innovative, revenue-generating products and services. 

The lifespan of corporate existence has been shortening1 and competition is Darwinian. The fittest organizations with chances for survival are adaptive, agile, and flexible to the changing needs of customers, technology, regulations, demand, and external trends. Consequently, agile organizations have the ability to innovate on behalf of their customers. Organizations that foster adhocracy produce the most innovation as compared to organizations with other culture types.3 Nevertheless, organizational culture is a strong, stable phenomenon that is difficult to change by leaders13 so organizations should foster adhocracy in order for leaders to successfully innovate.14

Amazon’s two-pizza team and STO organizational structures (i.e., agile structures) are designed so that employees stay close to Amazon customers, make high-velocity decisions, and avoid distractions and competing priorities. Multiple studies have supported that smaller, agile teams quickly produce more innovation.7 Amazon STOs manage two-pizza teams by acting as a bridge across organizational boundaries and using the right mechanisms to audit and inspect deliverables while delivering right results with a high bar. Amazon has leveraged ONA to inspect their organizational designs and ensure teams are making high-velocity decisions and producing innovation. ONA findings helped identified strengths, opportunity areas, and interventions for staying Day 1.

HR Professional Applications 

The following are practical applications for HR professionals to consider for helping their business make Day 1 a repeatable and embedded element of their organizational culture:
  • Assess and address organizational culture. Organizations require innovation to survive and the culture to support innovation. HR professionals should consider using Cameron and Quinn’s2 24-item Organizational Culture Assessment Instrument, a valid measure of organizational culture, to determine their organization’s current and desirable culture. If an organization has low levels of Day 1 culture, HR professionals could partner with their business by realigning their organization’s culture. Denham and Kaberon15 suggests using a Frame Stretching (FS) technique to help with making the cultural shift. FS helps leaders change their organization’s current frame with a new frame-of-reference so that leaders can reimagine their current situation and potential. 
  • Consider implementing agile team structures and STOs. HR professionals could partner with business leaders by starting small, launching a few agile teams, and gathering data to determine the agile team value and constraints faced.6 Determine if increasing team agility gets employees closer to the customer, increases innovation, and delivers positive financial results. Then scale up agility in other parts of the organization if benefits outweigh costs. It is also recommended that HR professionals partner with the business to design or identify STO roles in the agile teams. Below is a checklist to ensure single-threaded ownership:
    • Are they responsible for only one thing (program, product, platform, feature, or service)?
    • Are they able to make all/most of the decisions regarding the one thing they’re responsible for (absent of corporate/product priorities)?
    • Do they have a direct reporting relationship with all of the headcount required to execute the work?
  • Continuously evaluate organizational structures. HR professionals could partner with business leaders to assess the extent to which employees are making high-velocity decisions and are producing innovative products and services. They can use ONA as a tool to identify strengths, opportunity areas, and interventions for increasing execution speed, creating adaptive space, and appropriately leveraging talent.
  • Implement adaptive practices. Business leaders should find adaptive space opportunities and implement adaptive practices12 to foster a climate of creativity and produce high-quality creative ideas with the help from their HR partners. They could start small with hackathons and then spread more larger-scale practices (e.g., adaptive summits, mechanisms to get closer to the customer) across the organization.  
Beth Galetti is Senior Vice President of Worldwide Human Resources for Amazon.

John Golden III, Ph.D., is Head of Devices Talent and Leadership Development at Amazon.  He can be reached at

Stephen Brozovich is Head of AWS Talent Management at Amazon. He can be reached at

1 George, P. (2016). The scary truth about corporate survival. Harvard Business Review, December, 24-25. 
2 Cameron, K.S., & Quinn, R.E. (1999). Diagnosing and changing organizational culture: Based on the competing values framework. Reading, MA: Addison-Wesley.
3 Amazon Corporate (2017). 2016 Letter to Shareholders. Retrieved from
4 Hammond, M.M., Neff, N.L., Farr, J.L., Schwall, A.R., & Zhao, X. (2011). Predictors of individual-level innovation at work: A meta-analysis. Psychology of Aesthetics, Creativity, and the Arts, 5, 90–105.
5 Ahlback, A., Fahrback, C., & Murarka, M., Salo, O., (2017). How to create an agile organization. Retrieved from organization
6 Rigby, D., Sutherland, J. & Noble, A. (2018). Agile at Scale. Harvard Business Review, May-June, 88-96.
7 de Rond, M. (2014). Why less is more in teams. Harvard Business Review Onpoint, Spring, 13-14. 
8 Steiner, I. (1972). Group process and productivity. New York, NY: Academic Press.
9 Mueller, J. (2011). Why individuals in larger teams perform worse. Organizational Behavior & Human Decision Process, 117, 111-124. 
10 Cross, R., & Parker, A. (2004). The Hidden Power of Social Networks: Understanding How Work Really Gets Done in Organizations. Boston, MA: Harvard Business Review Press.
11 Arena, M. (2018). Adaptive Space: How GM and Other Companies are Positively Disrupting Themselves and Transforming into Agile Organizations. New York, NY: McGraw-Hill Education.
12 Arena, M. & Uhl-Bien, M. (2016). Complexity leadership theory: Shifting from human capital to social capital. People + Strategy, 39, 22-27.
13 Hogan, S.J., & Coote, L.V. (2014). Organizational culture, innovation, and performance: A test of Schein’s model. Journal of Business Research, 67, 1609–1621. 
14 Golden, J. & Shriner, S. (2017). Examining relationships between transformational leadership and employee creative performance: The moderator effects of organizational culture. Journal of Creative Behavior, October.
15 Denham, J., & Kaberon, R. (2012). Culture is king: How culture contributes to innovation. Journal of Product Innovation Management, 29, 358–360.
16 Ingham, A. G., Levinger, G., Graves, J., & Peckham, V. (1974). The Ringelmann effect: Studies of group size and group performance. Journal of Experimental Social Psychology, 10, 371–384.