Article | 10 min read

People: The Human Side of Innovation at Amazon

Where does innovation come from?

For Amazon – a company recognized for applying technology-enabled innovations to transform customers’ experiences and expectations – it may seem logical to think that technology is the source of innovation. While great tools and technology play their part, innovating consistently is not possible without people. People deliver the organization’s mission, whether that’s maintaining operational excellence, driving growth and expansion, or — as is the case at Amazon — being Earth's most customer-centric company.

► Download the ebook: People: The Human Side of Innovation at Amazon

To deliver on Amazon’s mission, our people live by the mantra, “Work hard. Have fun. Make history.” Amazonians come to work every day looking to build things that will make history – bold, audacious, game-changing things like transforming what customers expect in their eCommerce experience through Amazon.com, launching an extraordinary two-day shipping experience, changing the way customers interact with technology in Kindle and Alexa, or launching new industries in Amazon Web Services. We recognize that to make history – to continually innovate in spectacular fashion on behalf of our customers – is incredibly hard work, and so it should also be enjoyable to do.

People: The Human Side of Innovation at Amazon ebook

People strategy

Fostering the type of environment that enables Amazonians to work hard, have fun, and make history takes a lot of effort, but is vital to our success. We deliberately design (and regularly redesign) our people strategy in order to delight our customers and deliver results on their behalf. As part of that, we’ve developed and continue to improve specific mechanisms that operationalize the mindset and behaviors that fuel innovation and growth. These mechanisms include technology, tools, processes and practices across an Amazonian’s lifecycle – from their recruitment and hiring, to their day-to-day experience as an employee, their skills growth and career development, to their performance measurement – and the iterative improvements that improve organizational health over the long term.

People strategy

Working backwards

By working backwards from our people’s needs, we develop mechanisms that reinforce our core values so that our people can deliver customer-centric innovation consistently, at scale and speed. These mechanisms are effective for Amazon and while they may not be a perfect fit at every company, we’ve found that these methods, and the thinking behind them, can inspire other organizations seeking to strengthen their own approach to managing talent and advancing their own unique missions.

Working backwards

A flywheel for managing talent

Managing talent to best deliver an organization’s goals, meet employees’ shifting expectations, and scale as a business grows can be a challenge for any company. At Amazon, we’ve found that a flywheel can be an effective conceptual model to express how something operates. It’s a simple but powerful way to explain the intrinsic relationship between two areas that mutually reinforce one another – a virtuous cycle that requires a significant amount of force to get moving, but as you keep it turning, it builds enough momentum to continue on its own.

In his 2002 letter to shareholders, Jeff Bezos talked about the flywheel concept for the first time: “One of our most exciting peculiarities is poorly understood. People see that we’re determined to offer both world-leading customer experience and the lowest possible prices, but to some this dual goal seems paradoxical if not downright quixotic. Traditional stores face a time-tested tradeoff between offering high-touch customer experience on the one hand and the lowest possible prices on the other. How can Amazon.com be trying to do both?”

Employee Experience Flywheel

Lower prices, more visits

Bezos goes on to explain that lower prices lead to more customer visits, and more customers increase sales volume and the ability to invest in better customer experience. As the volume of customers grows, more value is realized from existing assets serving a larger customer base so prices can be lowered further, and the process repeats. Feeding any part of the flywheel effectively accelerates the loop. As Bezos said at the time, “We believe our ability to lower prices and simultaneously drive customer experience is a big deal.”

Lower prices, a better customer experience, more traffic, more sellers and greater selection all accelerate the flywheel. In Amazon’s employee experience flywheel, attracting and hiring top talent, providing them with industry-leading work experiences, investing in their growth and development, and constantly evaluating and improving our mechanisms accelerates the wheel (Figure 1). By providing Amazonians an environment and experience where their superpowers are intrinsically linked to delivering the best possible experience for customers, we’re able to deliver on Amazon’s mission while continually iterating and improving our approach to talent.

Lower prices, more visits

Employees innovate for customers

"The feedback loop enabled by the flywheel propels the ongoing evolution and improvement of our people operations so that employees can continue to best innovate on behalf of customers."

At their core, we believe that all Amazonians are builders. Who are builders? They’re individuals who are driven to delight customers. They’ll look at customer experiences, dissect what doesn’t work well about them, and seek to reinvent them — again and again. We seek those who question the status quo and, as our CEO Andy Jassy writes, “realize launch is the starting line, not the finish line.” ‘Builders’ are not limited to technical people. It’s a mindset and approach. Every Amazonian is expected to bring this builder mentality and behavior to innovating for customers.

Our process to hire builders focuses on three principles:

  1. Similar to many companies, we seek the right mix of capabilities, skills, and a diversity of backgrounds and experiences. Critically however, we also explore how well a person will thrive in Amazon’s unique and peculiar culture.
  2. To paint a more complete, data-driven picture of a candidate and to help assess candidates more fairly, we have a diverse set of interviewers assess the candidate through specific questions aligned with Amazon’s values, or what we call our Leadership Principles.
  3. Lastly, we believe that every hire is an opportunity to raise the performance bar at the company so that we never become complacent, and have implemented a specific Bar Raiser role to achieve that in the hiring process.

What does Amazon look for in a new hire?

In our hiring, we of course look for capabilities, skills, diversity of background and experience. However, another filter when recruiting and assessing applicants is whether an individual will thrive in Amazon’s unique and peculiar culture. It takes a particular set of traits to tackle customer challenges in the way Jassy describes, to be motivated and inspired by that work every day, and to thrive in that environment. When assessing candidates there are a host of human biases, well-established by psychological researchers, that can stand in the way. So, we’ve incorporated mechanisms into the hiring process that help ensure we hire people with a builder’s mindset while getting around those biases.

Our Leadership Principles are the foundation of how we operate.

It’s no surprise then that these 16 values — such as “Customer Obsession,” “Invent and Simplify,” “Learn and Be Curious,” and “Insist on the Highest Standards,” to name a few — are also the primary factors in whom we attract and hire. In the hiring process, we give every interviewer two Leadership Principles that are relevant to the role and ask that they evaluate a candidate against those.

There are carefully calibrated questions we use specific to each Leadership Principle that help us consistently assess a candidate’s behavior, and their approach to challenging situations based on their experiences. An interviewer assigned the “Ownership” principle might ask the interviewee to describe a time when they took on work outside of their comfort zone, or someone assigned the “Learn and Be Curious” principle might ask a candidate if they’ve ever invested their time to read up on a topic in order to gain a deeper level of expertise and do the job better.

By focusing on the Leadership Principles most critical for the role, and having a diverse set of interviewers assess the candidate through consistent and specific lenses, we arrive at an objective and data-driven picture to best assess the candidate’s fit with Amazon’s culture.

Amazon’s Bar Raiser Program

What makes Amazon’s interview process truly unique is our Bar Raiser Program. All interviewees at Amazon go through training and shadow at least five interviews before they’re qualified to interview a candidate on their own. Bar Raisers go beyond that with extensive apprenticeship-driven training to become experts on applying Amazon’s Leadership Principles to the interview process. Bar Raisers are people who are outside of the immediate hiring team or organization, and can look at the candidates through an unbiased lens.

Once the interview feedback has been consolidated and it’s time to discuss each candidate, the Bar Raiser helps ensure that an appropriately high standard is being used to assess candidates against the Leadership Principles, and that every hire effectively raises the performance bar at the company. Where a typical mindset might be “if the candidate is ‘good enough,’ hire them,” at Amazon we ask “would this person raise the bar for their peers in this role?” That is a high mark to meet for a new hire, especially when considering that their peers were also expected to raise the bar when they were originally hired.

Bar Raisers may call out particular strengths or weaknesses that the hiring team might miss. They help interviewees craft their questions and assess the candidates in accordance with our Leadership Principles. Most importantly, they have full veto power in the process, which enables them to be even more effective in applying our relentlessly high standards.

Once we attract and hire builders, it’s imperative that we enable them to build. If we want them to deliver customer-centric, iterative innovation, we must, as Jassy explained, “give teams the right tools and permission to move fast.” In order to provide our builders with a great experience, we equip them with the most effective productivity and collaboration tools possible, empower them to make decisions and control their own destinies as agile teams, and pair them with single-threaded leaders with clear focus, authority and accountability.

We have a special appreciation for productivity and collaboration tools at Amazon, and that’s become even more important in recent years. As physical experiences have been supplanted or complemented by digital ones, only by making the virtual versions of these as good as the real ones can organizations provide an effective and equitable experience. We take a virtual-first approach to developing and deploying our workplace technologies. Everyone can access the most effective tools, whether working in our Seattle headquarters, at a data center in Jakarta, or from home. A major part of why Amazon started AWS was to provide businesses with the right tools to help them experiment and build fast. Cloud infrastructure is our business, but we are also cloud consumers. It’s essential to enabling virtual workspaces with exceptional availability, reliability, and scalability that ensure our employees are productive.

More importantly, though, we create an organizational structure that supports rapid decision-making with few dependencies. Whether someone is a programmer or a program manager, working with code or working with people, builders are eager to iterate and innovate for the benefit of our customers. As organizations grow, expanding processes and communication may impede desired agility and speed. To address that at Amazon, we focus on ensuring that our teams have ownership, accountability and the right skills to execute, and that they’re small enough to move fast and stay close to the customer. This approach helps minimize bureaucracy and maximize the time employees can spend on innovating for customers. Smaller, nimbler teams can run fast, experiment early and often, fail fast, and rapidly iterate to drive customer value.

We also take a single-threaded approach to leadership, putting one highly skilled person in charge of a given project and making it their only job. Doing so results in sharper focus, clearer authority and accountability, greater creativity, and more ownership and engagement among team members. A single-threaded leader provides a strategic vision for their similarly dedicated teams and removes obstacles for them. “It’s hard for teams to be deep in what customers care about in multiple areas,” Jassy explained. “It’s also hard to spend enough time on the new initiatives when there’s resource contention with the more mature businesses; the surer bets usually win out.”

We give single-threaded leaders and their teams full autonomy, trust, and responsibility. Our collection of small, agile, and autonomous teams become subject matter experts. They know their customers’ needs better than anyone, spend all their time inventing for them, and are able to maintain that context and pace to keep iterating quickly. There are essentially a thousand startups at work at Amazon, enabled to do as much as is practical to develop, test, deliver, and manage new products and solutions for our customers.

Investing in learning and growth is crucial, not just for cultivating the capabilities necessary for innovation but also for meeting the expectations and ambitions of the builders Amazon hires. At Amazon, we have introduced instrumentation to operationalize this intent to provide training, challenges, and opportunities for expansion to our team members, and mechanisms to ensure our people learn and grow.

Amazon’s role in helping to grow and develop its people begins at their onboarding. We develop customized, job-specific learning paths for individuals. Then we support them in their jobs, using tools such as Wiki and video sharing tools, user forums, and news groups to enable self-service knowledge-sharing and problem-solving. To make sure that these aren’t just goals that could easily fall by the wayside in the course of day-to-day operations, we’ve integrated tools that ensure we walk our talk. One provides automated reminders to managers during an onboarding period to ensure that new hires have completed their necessary training.

"Our builders naturally value career mobility. Supporting that is a top priority and a differentiator for us as an employer."

While we hire talent for a specific role, we also willingly move them throughout the organization. Doing so enables our employees to acquire a diverse set of customer experiences that increases their capacity for innovation. We also want someone to take the path that best suits them so they remain excited about the opportunities at the company.

Amazon is known for its automated recommendation engines for products. Inside the company, we auto recommend promotions for our people via an algorithm that evaluates our annual reviews. This primes managers to elevate individuals, putting them in a position to have to make an argument against promoting a team member rather than relying on them to initiate the consideration.

Mobility means more than promotions though. It also includes upskilling, reskilling, and non-traditional career advancement. We support employees who want to move laterally, leverage their skills in new roles, build and manage teams, and launch careers in other industries. We offer formal and informal stretch assignments and job rotations to help people develop and broaden their capabilities with work outside their comfort zones. If a particular role isn’t a good fit or an employee has grown out of it, we make it easy for them to find and apply for other opportunities in the company.

We enable job mobility through some of the usual ways other organizations might, such as the use of internal job boards or encouraging informational chats with hiring managers. However, at Amazon, a person’s manager and the review process also play valuable roles. Managers understand that not only do they need to help make their people successful by identifying opportunities to uplevel skills or gain new skills, but crucially, as owners of a business area, they need to create the opportunities for their people to grow. Review cycles also help chart a development course. Amazon’s review process, called Forte, gathers feedback from between 7 and 20 peers who call out the three Leadership Principles and an Amazonian's super powers, which their manager can then use to hone in on strengths they can apply to upcoming projects, and areas for growth and development.

In order to drive customer-centric innovation, we set high standards at the organization and team levels, and we measure our progress against them. We also evaluate the performance of our talent strategies on many dimensions. However, what may be a little different at Amazon is that we don’t think that managing against these talent metrics is the People, Experience, and Technology (PXT) team's job. Rather, it’s the business leaders’.

"Our business success is directly determined by our people success."

Therefore, our business leaders own their talent metrics. Whether its recruiting and attrition numbers at the team level or the performance evaluation of a product manager, it’s the responsibility of the business leaders. Our HR teams partner with the business across all things talent-related, but the buck stops with the business owner. After all, it’s the business that’s responsible for creating the right hiring process, workspace or team culture.

Talent metrics

Alongside key financial metrics like revenue growth or operational metrics, a leader is managed and measured against a number of talent metrics. Depending on the role, those may include employee Net Promoter Scores (eNPS), engagement and employee well-being metrics, talent attraction and retention numbers, or results from our daily Amazon Connections survey to assess how our teams are doing and identify emerging issues.

As an example, an important talent metric for our business leaders is what we call “dwell time”: how long, on average, it takes an individual at one level to move up to the next level. We have a benchmark that leaders are measured against. If a team’s dwell time is longer than average, that may indicate that the leader isn’t taking enough risks to promote individuals in their organization and that could expose the organization to retention issues. If the dwell time is shorter than average, it’s possible that people are being promoted too quickly and the leader may not be holding the team to the highest standards. Neither result is inherently bad, but rather creates an opportunity for discussion and exploration to better understand the results.

Similarly, Amazon Connections – our daily employee survey – can surface important data quickly. Amazon Connections is a simple one-question survey employees need to answer when they log on each day. Questions often focus on ranking something on a scale of one to five to gauge agreement or disagreement, for example: “Our team is comfortable making decisions quickly” or “My manager does a good job on making a range of voices heard.” Data is shown to managers by area, and managers in turn often share the results with their teams in order to foster discussion around ways to improve and apply learnings. While simple, Amazon Connections has proven to be a very effective tool at helping us quickly assess various aspects of our talent and operations, and validate that we’re living our core values.

All of these efforts are part of a data-driven approach to ensuring that we’re creating the ideal environment for people to do their best work. Putting some rigor and process around the performance of talent management, and holding business leaders accountable for the results, ensures that our teams can apply their superpowers to solving customer problems, which yields business growth and success.

Success metrics are fuel for iterating and improving upon our talent strategies — attracting and hiring the best talent, empowering and equipping employees, delivering an equitable and effective employee experience, and growing and developing our people.

Just as we’re never satisfied with having done enough for our customers and are constantly iterating to hone their experiences, so too do we iterate our approach to people.

"We build new employee processes the same way we approach creating new customer innovations – by working backwards from a real employee need for everything we do."

We are always testing new solutions with employee test groups to determine what to roll out to the broader employee population, just as we do with customer products. We ensure every new employee system is instrumented with data collection so we can assess how well it is performing for Amazonians and continue to improve our tools and processes.

In this way, HR at Amazon is a mirror of our customer-facing teams: Amazonians are HR’s customers, and their role is to be obsessed with the employee experience. In doing so, they enable Amazonians to best work on behalf of Amazon’s customers. Both Amazon Connections and the Forte review process mentioned earlier are examples of innovations born out of employee needs, derived from a ‘working backwards’ process.

Key takeaways

People are vital to Amazon’s mission of customer-centric innovation. We deliberately design our talent strategies, working backward from employee needs to develop tools and mechanisms that enable our people to do their best work on behalf of our customers, while continuing to learn and grow. While Amazon’s approaches may not be a perfect fit at every company, they may inspire others seeking to strengthen their approach to managing talent and achieving their unique missions.

Here are some steps and processes to consider:

  • Use a conceptual model such as a flywheel to express your employees’ experience and help separate the impact of iterative innovation at each stage of their journey.
  • Incorporate mechanisms into the hiring process to help avoid human biases and ensure organizational fit; for example, parceling out important values and attributes for interviewers to look for separately, or using an objective person to validate interview findings and help raise the bar.
  • Put the right day-to-day tools, processes, and structures in place to enable your people to be the most productive and effective. At Amazon that means small autonomous teams with single-threaded leaders.
  • Explore mechanisms that can foster growth and development of your people. At Amazon, that includes tools such as automated reminders for training and algorithmic evaluation of annual reviews to suggest promotions, and policies that encourage career mobility, upskilling, and reskilling.
  • Consider mechanisms to more effectively measure your talent strategies, such as making HR metrics the responsibility of a business unit or team, or novel ways to assess programs and validate that the organization is living up to its values.
  • Make the best even better. Keep working backwards from the needs of your people to develop new mechanisms and improve on existing ones in order to continuously improve your employees’ experience.
Key takeaways like customer-centric innovation

About the authors

Choong W. Lee, Worldwide Head, Innovation Learning from Amazon

Choong engages senior AWS customer executives to help them apply insights and learnings from Amazon’s experiences, delivering customer innovations at speed and scale. Previously at AWS, he was the Worldwide Head of Migration Acceleration Program (MAP), leading a global team helping the largest enterprise customers accelerate their AWS migration and modernization efforts. Prior to AWS, he was a senior executive at leading technology and management consulting companies for two decades, including: Partner at The Boston Consulting Group (BCG), Global Head of Data and Integration Architecture Group within Accenture’s IT organization, and Technology Executive at Accenture. Choong holds a Bachelor of Science degree with Honors in Electrical Engineering from the University of Illinois at Urbana-Champaign.

Stephen Brozovich, Principal Evangelist, People Experience Team, Amazon Web Services

Stephen started his career at Amazon in 1999 as a web developer, and for the next thirteen years held technical roles that allowed him to observe the organizational patterns that accompanied Amazon’s growth and technical evolution. In 2012 he moved to HR to lead the Amazon Culture Program, charged with ensuring Amazon is still the most innovative, customer-obsessed company on the planet. Following a rotation with Amazon’s Executive Development team and three years running Talent Management for Amazon Web Services (AWS), Stephen returned to his focus on culture as a Principal Evangelist for AWS HR in 2019. His mission is to help AWS customers unlock innovation and invention within their own organizations.