Unpacking Amazon’s unique ways of working | Bill Carr (author of Working Backwards)

Nov 2, 2023 Episode Page ↗
Overview

Bill Carr, co-author of "Working Backwards" and former Amazon VP, discusses implementing Amazon's core practices like working backwards, single-threaded leaders, input/output metrics, and the bar raiser program. He shares insights from his 15 years at Amazon, covering process innovation and decision-making.

At a Glance
17 Insights
1h 33m Duration
14 Topics
8 Concepts

Deep Dive Analysis

Amazon's Unique Contributions to Business Practices

The Origin of Amazon's Process Innovation

Understanding Single-Threaded Leadership

Functional Countermeasures for Single-Threaded Leadership

Implementing the Disagree and Commit Principle

Deciphering the 'Leaders Are Right, A Lot' Principle

The Working Backwards Framework Explained

The PR FAQ Process for Product Innovation

Creating a Product Funnel, Not a Product Tunnel

Amazon's Approach to Action vs. Talk

Amazon's Flywheel and Input vs. Output Metrics

Cultivating Risk-Taking and Accepting Failure

Amazon's Bar-Raiser Practice for Hiring

Advice for Implementing Working Backwards Practices

Working Backwards

This philosophy dictates that when making decisions or thinking about problems, one should always start with what is best for the customer and then work backward from there to devise solutions. It prioritizes customer needs over internal constraints like financial or engineering limitations, operating on the 'article of faith' that serving customers well will lead to positive business outcomes.

Single-Threaded Leadership

This organizational model creates standalone teams with a single leader who has dedicated cross-functional resources reporting directly or indirectly to them. It shifts from a project-based to a program-based orientation, aiming to foster ownership, speed, and agility by reducing resource contention and allowing teams to focus on their specific domain with clear accountability.

Disagree and Commit

This principle mandates that team members have a 'backbone' to voice their disagreements and provide new information or perspectives during decision-making. Once the leader has heard and understood the dissenting viewpoint and made a decision, the dissenter is expected to fully commit to and support that decision, reflecting the understanding back to their own organization.

Leaders Are Right, A Lot

This principle acknowledges that effective leaders possess sound judgment, often developed through extensive experience, including learning from mistakes. It suggests that leaders who consistently make good decisions, weighing various factors and interpreting data effectively, inspire confidence and are more likely to be followed by their teams.

PR FAQ Process

A new product innovation process where teams start by writing an internal press release and a Frequently Asked Questions document for a hypothetical product or feature. This forces clarity on the customer, their problem, and the proposed solution, serving as a method to decide what to build and to filter out less viable ideas early in development.

Product Funnel vs. Product Tunnel

A 'product funnel' is a desired approach where many ideas enter at the top, but only the best and most impactful ones make it through to development, allowing for selection and prioritization. A 'product tunnel,' conversely, means every idea that enters is destined to come out, indicating a lack of critical evaluation and comparison against other potential builds.

Input and Output Metrics

Output metrics (e.g., revenue, customer growth) are the results, while input metrics are the measurable activities or factors that drive those outputs. Amazon focuses on improving input metrics (e.g., selection, customer experience, low prices) with the 'article of faith' that consistently improving these will naturally lead to desired output metric performance over the long term.

Bar Raiser

A crucial role in Amazon's hiring process where an interviewer, not the hiring manager or recruiter, runs the debrief meeting and technically has veto power over hiring decisions. Bar raisers ensure adherence to objective criteria (leadership principles) and behavioral interviewing methodology, acting as a check against urgency bias and maintaining high hiring standards.

?
What is the core idea behind Amazon's 'Working Backwards' philosophy?

The core idea is to begin any product or feature development by identifying and understanding the customer's problems or needs, then working backward to design a solution without initial constraints, trusting that customer focus will drive business success.

?
Why did Amazon develop the concept of 'Single-Threaded Leadership'?

Amazon developed it to address complexity and resource contention as the company scaled, aiming to foster ownership, speed, and agility by giving a single leader dedicated cross-functional resources for a specific program or area.

?
How should an employee practice the 'Disagree and Commit' principle effectively?

Employees should voice their disagreements by providing new information or a different point of view. Once the leader acknowledges understanding of their perspective and makes a decision, the employee must then fully commit to and support that decision.

?
What does the Amazon principle 'Leaders are right, a lot' imply about leadership?

It implies that effective leaders possess strong judgment, often gained through extensive experience and learning from past mistakes, enabling them to consistently make sound decisions that guide their teams in the right direction.

?
What is the purpose of the PR FAQ process in Amazon's product development?

The PR FAQ process serves as a method to decide what new products or features to build by forcing teams to articulate the customer, their problem, and the solution in a press release and FAQ format, helping to filter and strengthen ideas early on.

?
How does Amazon prevent a 'thinking to doing gap' where ideas are discussed but not acted upon?

Amazon avoids this by requiring ideas to be thoroughly documented and fleshed out through processes like the PR FAQ, preventing wasted effort on flawed concepts, and by assigning dedicated resources (single-threaded teams) to own and execute compelling ideas.

?
What is the difference between input and output metrics, and why is the distinction important?

Output metrics are the results (e.g., revenue), while input metrics are the controllable activities or factors that drive those results (e.g., customer experience, selection). The distinction is important because focusing on improving inputs provides a clearer path to long-term success than short-term efforts to directly manipulate outputs.

?
What is a 'Bar Raiser' in Amazon's hiring process and what is their role?

A Bar Raiser is an interviewer, independent of the hiring manager, who runs the interview debrief and ensures the hiring process adheres to objective criteria and behavioral interviewing. Their role is to act as a balance against urgency bias and maintain high standards for cultural fit and functional excellence.

1. Prioritize Customer Needs First

When making decisions or thinking about problems, always start with what’s best for the customer and work backward, as prioritizing customers is an article of faith that leads to sales, revenue, and free cash flow.

2. Implement PR/FAQ for Product Innovation

For new products or features, start by writing an internal, data-rich press release that clearly defines the customer, their problem, and your solution, then iteratively refine it through concentric circle reviews to strengthen the idea and select the best options.

3. Adopt Single-Threaded Leadership

Organize teams with a single leader and dedicated cross-functional resources, shifting from temporary projects to continuous programs with clear ownership and metrics; ensure a service-based architecture with well-defined APIs and implement functional countermeasures to maintain expertise.

4. Focus on Input Metrics

Shift focus from short-term output metrics (like revenue) to long-term input metrics that drive customer experience, such as selection, price, and delivery speed; rigorously measure and continuously improve these controllable inputs, avoiding compound metrics that obfuscate understanding.

5. Practice Disagree and Commit

Voice your disagreement with backbone, providing new information or a different viewpoint; once your leader demonstrates understanding of your perspective and makes a decision, commit fully to that decision, even if you still disagree, by focusing on the core idea and striving to make it viable.

6. Implement Bar Raiser Hiring

Appoint a “bar raiser” for each interview loop—someone outside the hiring chain—to lead debriefs, ensure adherence to objective criteria (like leadership principles) and behavioral interviewing, and act as a check against urgency bias, helping to maintain high hiring standards.

7. Create a Product Funnel

Design your product development process as a funnel, allowing many ideas at the top but filtering to fund only the best ones, rather than a tunnel where every idea is pursued, to ensure precious engineering resources are deployed effectively.

8. Separate Product Thinking & Shipping

Break product development into two distinct processes: first, a dedicated process for deciding what to build (e.g., PR/FAQ), and second, a separate process for efficiently and effectively shipping the decided products with minimal bugs.

9. Flesh Out Big Ideas Deeply

Avoid endless debates on vague concepts by thoroughly documenting and fleshing out big ideas before building; assign dedicated resources (even a single person or team) to deeply explore and evaluate compelling ideas, ensuring they are viable before committing to development.

10. Codify Leadership Principles

Develop a set of real, actionable leadership principles and create processes to reinforce them, such as using them as objective criteria in hiring and performance evaluations, to guide culture and decision-making.

11. Align Compensation with Long-Term

Structure compensation (e.g., stock-based) to incentivize long-term company success rather than short-term financial performance, and base performance reviews on actual contributions and input metrics, to encourage risk-taking and innovation without fear of immediate penalty.

12. CEO Drives Innovation Structures

Ensure the CEO is deeply committed to innovation and actively involved in reviewing new initiatives, creating dedicated structures and running interference to protect small, innovative teams from bureaucratic impediments and foster speed.

13. Commit Fully to New Processes

When implementing new processes, approach them with full commitment and discipline, understanding that they will be challenging initially and require several months to master, rather than trying them out lightly.

14. View Bar Raisers as Helpers

Hiring managers should view bar raisers and the interview loop as helpful resources, not bureaucratic hurdles, because their role is to assist in making the right hiring decision, which ultimately saves significant time and prevents negative team impact.

15. Avoid Compound Metrics

Do not create compound metrics by combining several important metrics into one, as this makes them meaningless and obfuscates understanding; instead, break out and manage each metric individually to understand actions and reactions.

16. Apply Scientific Management

Approach company management scientifically, experimenting with different ideas and hypotheses, implementing them, and iteratively improving based on what works, especially as the company grows and becomes more complex.

17. Cultivate Sound Judgment

Develop sound judgment through extensive experience, including learning from mistakes and observing others’ decisions, to effectively weigh different information and consistently make good choices that lead the team in the right direction.

If we served customers well, if we prioritized customers and delivered for them, things like sales, things like revenue and active customers, and things like the share price and free cash flow would follow.

Bill Carr

Most people confuse speed with velocity. And the difference between the two is that, you know, speed, velocity has both speed and a vector to it, meaning you're some specific destination.

Bill Carr

You have to look for what that is. And then your job is to then take that kernel and try to run with it and expand it and try to see how I can take that, that, you know, that idea, that concept, and then make it into something viable.

Bill Carr

Slow is smooth and smooth is fast.

Bill Carr

The data doesn't, you know, rarely makes the decision for you. What is happening is then a lot of judgment and interpretation of the data weighing that weighing various factors to then come to a decision.

Bill Carr

PR FAQ Process for New Product Innovation

Bill Carr
  1. Start by writing a press release describing the new product or feature, focusing on the customer, their problem, and the solution.
  2. Ensure the press release is written in a way that highlights customer value and is data-rich, not hyperbolic.
  3. Define the hypothetical launch date in the press release to provide directional cues on complexity and timeline.
  4. Develop a Frequently Asked Questions (FAQ) document to address potential questions and concerns about the product.
  5. Begin with low-fidelity writing, then share with a small group for feedback.
  6. Iteratively improve the PR FAQ by sharing with increasingly wider groups within the company, gathering feedback at each stage.
  7. Use this iterative review process to strengthen the idea and determine its viability, understanding that not all PR FAQs will reach the highest levels of leadership.

Implementing Amazon's Bar Raiser Hiring Practice

Bill Carr
  1. Select a department to pilot the process, choosing individuals who care deeply about hiring, are good interviewers, and have high standards.
  2. Appoint one person on every interview loop as the 'Bar Raiser,' ensuring they are not the hiring manager, recruiting manager, or a direct report of the hiring manager.
  3. Train Bar Raisers properly on the process, objective criteria (e.g., leadership principles), and behavioral-based interviewing methodology.
  4. The Bar Raiser runs the debrief meeting, ensuring adherence to the process and objective criteria.
  5. The Bar Raiser acts as a balance against the hiring manager's urgency bias, ensuring candidates meet high standards and fit the company culture.
  6. While the hiring manager makes the final decision, the Bar Raiser technically has veto power, though a good Bar Raiser guides the hiring manager to the right decision through discussion rather than using a veto.
2003 to 2007
Period of significant product and process innovation at Amazon During this 4-year window, most innovative products (Kindle, AWS, Alexa, Echo, Prime) and processes (except one) were developed.
15 years
Length of Bill Carr's tenure at Amazon Joined 5 years after Amazon's founding in 1999, stayed until the end of 2014.
500
Number of items on the S-team goals list (around 2007-2008) A list of the most important company goals harvested from operating plans.
10
Number of S-team goals with a financial metric Out of 500 goals, only this many had financial metrics like revenue or gross profit, indicating a focus on input metrics.
Up to 10 hours
Typical weekly time commitment for a Bar Raiser This role can be a significant time investment.
1999
Year the Bar Raiser hiring process was established at Amazon Established early in the company's history to maintain hiring standards during hyper-growth.
3-4 years old
Amazon's age when the Bar Raiser process began The company was in a hyper-growth phase, leading to the need for structured hiring.
2004
Year Jeff Bezos wrote about a 'puck' that would become Echo The technology wasn't available at the time, but the idea predated the product's launch by about a decade.