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

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.