Working Backwards, by Colin Bryar and Bill Carr
My nuanced summary of the principles behind Amazon's wildly successful culture of innovation.
I’m excited to kick off 2023 with my notes on this excellent book. The authors share some really interesting stories, particularly within Part 2, which are well worth the full read. Otherwise my favorite topics include:
The “Bar Raiser” hiring process (chapter 2) - essential if you want to be better at hiring.
The mechanics of how written narratives shape their work (chapters 4 and 5) - this alone makes me want to work at Amazon.
The useful DMAIC - Define, Measure, Analyze, Improve, Control - approach to metrics (chapter 6), borrowed from Six Sigma, which guides their data-informed culture.
If you enjoy the summary below, I highly encourage you to buy the book either via Amazon or Bookshop.
Introduction
This book explores the practical methods used to reinforce Amazon’s cultural goals of a) putting customers first; b) being willing to think long-term; c) being eager to invent (and willing to fail); and d) taking professional pride in operational excellence.” [The authors then provide their backgrounds, including Colin's 2 years as Jeff (Bezos)'s full-time technical advisor, and Bill's tenure setting up Amazon’s Kindle products and other digital businesses.]
PART ONE: BEING AMAZONIAN
Chapter 1: Building Blocks - Leadership Principles and Mechanisms
Amazon’s leadership principles began with Jeff, as he would do the job of new employees alongside them while they got up to speed. His early emphasis was to make each customer interaction perfect, since any "one bad customer experience would undo the goodwill of hundreds of perfect ones." [E.g. he would proof-read customer service emails dealing with any new topic before it was sent out, until he no longer needed to.] As the company grew, the HR team interviewed existing leaders to distill the essential qualities of an “Amazon Leader,” which has further evolved into the leaderships principles which are all woven into the day-to-day practices of the company:
Beyond individuals calling out deviations, these principles are reinforced by mechanisms of the company which create the conditions for success. The foundational mechanisms are:
The Annual Planning Process: Detailed plans are created 1x per year, and include a) assessment of past performance (goals achieved or missed, lessons learned); b) key initiatives for the coming year; c) detailed income statement; and d) requests (and justifications) for resources. Plans are reworked until they align with top-down initiatives. Midway through the year, plans are adjusted based on the peak season results and trajectory of the business.
S-Team Goals Process: Senior leadership selects the handful of initiatives & goals from annual planning which are most important to achieve. (Teams still strive to achieve all of their goals; S-team goals are just prioritized ahead of any others). Notably, a) there will be dozens if not hundreds of S-team goals across all of Amazon in any given year; b) they are aggressive (actual expectations are that only 75% will be completed); and c) they are detailed. Review meetings focus on deep diving into details around execution.
Compensation Plan: Equity is emphasized to ensure senior members are aligned with long-term value creation. This encourages collaboration & frugality, and is well aligned with other leadership principles.
Chapter 2: Hiring - Amazon's Unique Bar Raiser Process
Amazon's approach to hiring reflects a mindset of raising the bar with each new hire, and their process is designed to “raise the ratio of hits to misses significantly." Each interview loop has an elite interviewer - the “Bar Raiser” - whose sole purpose is to maintain quality within the process. Some tips:
Typical biases within hiring include: a) urgency bias (e.g. there is pressure to fill a role, which leads to desperation); b) confirmation bias (focusing on positive elements blocks out negative or contradictory signals); c) groupthink (particularly when feedback meetings are unstructured) ; d) personal bias (looking for people like us)
The cost of a bad hiring decision is always worse than waiting for the right candidate who raises the bar for your team. (At scale, hiring badly redefines your culture in undesirable ways.)
Rule of thumb: Aim to have your team feeling a healthy sense of intimidation by the consistently rising standards as your company grows.
Most startups lack a formal process and training for interviewers, which feeds the negative cycle of bias and poor decisions. This is amplified when they’re growing quickly.
The Amazon “Bar Raiser” program has the following steps when hiring for any role:
Job Description: reviewed by the Bar Raiser to make sure it is clear, specific, and focused. Probe for aspects the hiring manager had failed to identify. A good JD keeps interviewers and recruiters aligned.
Résumé review: straightforward when the JD clearly identifies the intent of the role. For example, looking for indicators of someone being a builder and inventor and flagging indicators of someone being a coordinator only.
Phone Screen: 45 minutes of guided questions around the skills required for the role and examples of how the candidate demonstrates Amazon’s leadership principles. Only move forward if the screener is inclined to hire. (It's a mistake if they move forward when uncertain.)
In-House Interview: aim for 5-7 individuals, and must be "carefully planned, prepared, and executed." Always includes the hiring manager, the recruiter, and a Bar Raiser, with all participants trained on the interviewing process. The panel should not include anyone who will become the direct report of the candidate (weird power dynamic), nor anyone “more than one level below the level of the position the candidate will hold” (unable to truly assess).
Written Feedback: Everyone submits detailed notes. Aim to be close to verbatim: take notes along the way, and then block out 15 minutes post-interview to tie it together. Oral feedback (during the debrief) is unacceptable. Force each participant to take an independent stand on whether to hire or not.
Debrief/Hiring Meeting: Led by the Bar Raiser (trained and evaluated by the Bar Raiser Core team). Sequence: a) everyone reads all written feedback; b) ask if anyone, having read, would like to change their vote; c) facilitate a conversation around ways the candidate either a) meets the bar; or b) falls short. End the meeting with a hire / no hire decision by the hiring manager, validated by the bar raiser. (Bar raiser vetos are almost never used)
Reference Check: (now de-emphasized). Collect 4-5 references of managers, peers, and subordinates. Ask, “If given the chance, would you hire this person again?” Or, “Of the people you have managed or worked with, in what percentile would you place this candidate?”
Offer Through Onboarding: assume the candidate is being pursued, so keep the hiring manager and team members involved. Check in at least once per week, send a thoughtful gift, and get to know them better to understand the factors which affect their decision on the offer.
When quality is maintained at each stage, all participants rise to the level of the expectations placed on them. Hiring managers, in particular, “redouble their efforts to source, recruit, and hire candidates who were Amazonian.” Tips on participating in this process:
Hiring manager mindset: listen to learn from the loop participants, and speak infrequently. Avoid selling the other interviewers on the candidate.
Interviewer mindset: help the hiring manager gather data and make an informed decision.
Behavioral Interviewing tips: a) seek examples of what they've done in the past to personally contribute to solving hard problems or addressing work situations similar to what they'll face in your role; and b) understand how they accomplished those goals to see whether their methods align with your desired principles.
Seek details through STAR: a) what was the Situation; b) what were you Tasked with; c) what Actions did you take; and d) what was the Result. Distinguish between what the interviewee personally accomplished vs. what their team did.
Avoid open-ended questions; design specific questions which map to specific principles, and probe into details. Example interview questions:
“Can you give me an example of a time when your team proposed to launch a new product or initiative and you pushed back on their plan because you didn’t think it was good enough?" (Example of Highest Standards.)
“What was the toughest call on Project X, and who made it?”
“If you were assigned to work on a different project instead of Project X, what would have changed about Project X?”
Learn how to maintain control of the interview, e.g. ways to cut someone short and move on to the next question.
Tweak the model to get the results and quality you need. If you have too many (underqualified) people reaching interview loops, try conducting 2 phone screens. If you're seeking a more gender diverse team, try offering a phone screen to every female applicant.
Chapter 3: Organizing - Separable, Single-Threaded Leadership
It is typically expected that larger organizations require more coordination to address dependencies between teams, leading to slowed innovation. This occurs in software (as more teams innovate within a shared codebase, thus increasing the need for careful review and cautious QA) and within organizations (where project approvals and allocation of shared resources requires working through layers of people on different teams). Unfortunately, this "not only slows down the pace of innovation but also creates a dispiriting second-order effect: disempowered teams [...] frustrated by the slow pace of delivery that was beyond their control."
The solution lies not in improving communication but rather in eliminating communication between teams. At Amazon, they iterated toward the Two Pizza Team rules, which evolved into the “Single Threaded Leader” approach.
Be small, with no more than ten people. While this size works well in product development, this constraint was later relaxed for initiatives that required more to be successful.
Be autonomous, with no need to coordinate with other teams to get their work done. Take time upfront to identify and remove dependencies from upstream areas (especially in software, which requires moving toward a microservices architecture). Alignment with company-level goals is guided by the annual planning process.
Be evaluated by a well-defined "fitness function." Carefully identify the metrics which matter. [More on this in chapter 6.] Note: the "sum of weighted series of metrics" approach was later abandoned for simply reviewing the individual metrics which matter most.
Be monitored in real time. Take time upfront to instrument metrics to have real-time feedback. Better-instrumented teams are better at course-correcting when mistakes arise, which makes being wrong less costly than going slow.
Be the business owner. You must be "responsible for all aspects of your area of focus, including design, technology, and business results." This means only your team is accountable for results.
Be led by a multi-disciplined top-flight leader, with "deep technical expertise, [ability] to hire world-class software engineers and product managers, and excellent business judgment." This was later relaxed, as these leaders were too rare and hard to find. Instead, they found a balance within the matrix organization model where each team member (e.g. software engineer) directly reports to their functional manager (director of software engineering, responsible for their career growth) and indirectly reports to their two pizza team manager. “This meant that individual two-pizza team managers could lead successfully even without expertise in every single discipline required on their team.”
Be self funding, in that their work would pay for itself.
Be approved in advance by the S-Team.
The most important part is having a team (and a leader) which is single threaded. They do NOT split their time across different competing initiatives, but are singularly focused on one primary initiative and are not dependent on others for pursuing success.
"The STL delivers high-velocity innovation, which in turn makes Amazon nimble and responsive even at its now-massive scale. Free of the hindrance of excess dependencies, innovators at every level can experiment and innovate faster, leading to more sharply defined products and a higher level of engagement for their creators. Ownership and accountability are much easier to establish under the STL model, keeping teams properly focused and accurately aligned with company strategies."
Chapter 4: Communicating - Narratives and the Six Pager
Guided by the goal of making S-team meetings as good as possible, there emerged glaring issues with powerpoint presentations: a) lack of complete information (often presented verbally); b) too much time spent on the visual artifact instead of the thinking around it; c) bias toward effective presenters rather than strong thinkers; and d) issues with the flow of the meeting when questions would lead the presenter to jump around. [This was informed by a Tufte essay on The Cognitive Style of PowerPoint , which explained how bullet-point lists are bad for complex ideas.] Adopting the narrative form was controversial and took time to take hold, but eventually stuck and substantially increased the quality and quantity of effective communication across the company. Key tips:
6 page narratives (with no gimmicky formatting to cram more in) are a good baseline for 60 minute meetings: Allocate 20 minutes of reading upfront, with 40 minutes of discussion.
Narratives are highly efficient ways to transmit thinking, especially when you give meeting participants the time and permission (or expectation) to read it carefully before discussing.
Make the "key tenet" explicit so everyone knows the principle that drives the thrust of the document. [The book provides a complete illustrative example narrative document.]
Consider including an FAQ section which anticipates and addresses "counter arguments, points of contention, and statements which might be easily misinterpreted." This shows how thoroughly you've considered the ideas.
After everyone reads the document, resist the temptation to walk everyone through the sections. Ask instead for feedback. Dive deep with the goals of uncovering the truth about the proposed idea and making the document better.
Have someone familiar with the topic (not the presenter) capturing detailed notes.
Maintaining high expectations of quality for six page narratives leads to highly productive meetings around well-formed ideas which shape how the company pursues its work. Jeff's frame of mind while reading a narrative is to "assume each sentence he reads is wrong until he can prove otherwise." Similar expectations of critical thinking for people reviewing the narratives ties everyone together around the success or failure of initiatives.
Chapter 5: Working Backwards - Start with the Desired Customer Experience
The PR/FAQ is one of the most important narrative documents. It reinforces "customer obsession" by clearly anticipating PR-worthy problems & benefits, and forces thorough thinking with a corresponding FAQ to cover the most salient internal & external questions. [This was more successful than mockups, API guides, and other formats they experimented with over the years.] This format requires critical thinking about the things customers care about the most, which then become the focal point of discussion. [The book offers a complete illustrative PR/FAQ for a smart mailbox.] High-level tips:
PR/FAQ is particularly important for new, category-defining products (like the Kindle), where you must anticipate and think through all factors influencing the customer’s experience.
Distill it down: <1 page (PR section) and <5 pages (FAQ). Longer documents only show you’re avoiding hard decisions about what is important or not.
Useful prompt when reviewing a PR/FAQ: “So what?” "If the press release doesn’t describe a product that is meaningfully better (faster, easier, cheaper) than what is already out there, or results in some stepwise change in customer experience, then it isn’t worth building."
Expect most PR/FAQs to not become a product that gets to market. This is a feature of an effective system. Leadership is about "deciding what not to do rather than what to do."
When you identify a large market opportunity with hard feasibility challenges, consider Jeff’s perspective: “We shouldn’t be afraid of taking on hard problems if solving them would unlock substantial value.”
An effective press release has the following:
Heading, which names the product in a way your customers will understand
Subheading, which describes the customer and the benefits they gain
Summary Paragraph, which summarizes the product and the benefit
Problem Paragraph, which articulates the problem, from the customer's point of view
Solution Paragraph(s), which details how it simply and easily solves the customer's problem
Quotes and Getting Started, which elaborate on the content with quotes from a hypothetical customer and from your company's spokesperson, and ends with a clear call to action.
"The FAQ section is where the writer shares the details of the plan from a consumer point of view and addresses the various risks and challenges from internal operations, technical, product, marketing, legal, business development, and financial points of view." A good external FAQ addresses questions around how the product works, how much it costs, and how/where to buy it. Topics for internal FAQs:
Customer Needs and Total Addressable Market (TAM): how big is the need, and what are they willing to spend on a solution. Filter out people who don't meet the product constraints.
Economics & P&L: unit economics of the product, rationale for the selected price point, expected gross margin, investment required to build the product (across all dimensions).
Dependencies: other factors - especially third parties with their own agendas and incentives - which may influence success or failure.
Feasibility: "What level of invention is required and what kind of challenges are involved in building this new product?" This could be product engineering, customer UI, third-party dependencies, risk management of the upfront investment, etc.
Chapter 6: Metrics - Manage Your Inputs, Not Your Outputs
It's easy to feel awash in data yet still lack visibility into key parts of your business. Distinguish carefully between input metrics (leading indicators you can directly influence which have big downstream impacts) versus output metrics (lagging indicators which measure results you ultimately care about, but cannot directly control). Amazon’s perspective was informed by the Six Sigma process improvement method called DMAIC:
Define: Select and define the metrics you want to measure. These should emphasize the inputs (leading indicators) you are able to change to achieve the desired results. Details matter when finding the precise input metric that will most reliably lead to the outputs you care about.
Measure: Find unbiased ways to collect information and measure your input metrics correctly. Finance is an unbiased "call it like they see it" partner to make sure this is done well. Find and invest in the tools to collect and present the information in a usable format. Don't compromise here! Dig deep into potential problems to verify you are seeing what you think you are seeing.
Analyze: Separate signal from noise and reduce (or at least understand) overall variance. Find the causal links between your actions and subsequent results; make the business predictable.
Improve: Push business performance as high as it can go. (Importantly, this is relevant only after the three prior steps, otherwise you’re flying blind and subject to thrash.)
Control: Monitor your processes to ensure they're operating normally and performance is not degrading. At this point, processes which are well understood have decision making logic that can be encoded in software are candidates for automation.
Throughout this process, maintain a core operating philosophy of continual improvement. The most important thing is simply to get started and let the process improve over time. More tips:
Adjust your metrics as you learn about their limitations or unintended effects. [Story of measuring selection offered to customers, which eventually became the percentage of detail page views where the products were in stock and immediately ready for two-day shipping. This a) properly weighted more popular items and b) the customer's successful experience of knowing they could get the item quickly.]
Example bias in measurement: taking a snapshot of your inventory at a point in time may skew product availability for when shipments are received and processed (as opposed to when customers are shopping).
Audit your most important measurement methods periodically: "Assume that over time something will cause it to drift and skew the numbers."
Be relentless about discovering the root causes behind any surprise or perplexing problem with the data. Amazon’s Correction of Errors process: a) write a document describing the problem or error; b) drill down on what caused it by using the Five Whys to get to the root cause.
Charlie Bell (AWS and Ops guru): “When you encounter a problem, the probability you’re actually looking at the actual root cause of the problem in the initial 24 hours is pretty close to zero, because it turns out that behind every issue there’s a very interesting story.”
Watch out for when noise starts to obscure the signal; in particular when you misattribute your actions or a defect in a process (signal) to normal variation (noise).
Conduct "Weekly Business Review" discussions where leaders review the most important metrics for their business. Key features of the "deck" being reviewed are that a) it's a data-driven end to end view of the business (and follows the customer experience); b) it's mostly charts, graphs, and data tables; c) contains only the best, most informative metrics you have identified to-date; d) emphasizes emerging patterns, with results compared against prior periods; e) includes both near-term (trailing 6-week) and long-term (trailing 12 month) trend lines [see example]; and f) is peppered with anecdotes and exceptions (elements which fall outside some standard or usual situation).
Make the WBR as effective as possible by: a) using consistent and familiar reporting to speed interpretation; b) focusing on variance (expected or status quo performance is passed over with "nothing to see here"); c) having business owners present and explain their own metrics; d) focusing only on operational tactics, NOT strategy; e) treating variances as learning opportunities (e.g. "it's not the Inquisition"); and f) ensuring smooth transitions between teams (every minute counts).
Take the time to review anecdotes and understand the root cause behind edge cases. It almost always identifies issues or adjacent problems that need to be solved. [Example: when reviewing the top 10 products with negative contribution profit, the root cause for each was usually completely different and led to improvements across areas of the business.]
[Story on the power of anecdotes]: Observing a CS agent who knew of customers complaining about delivery damage on a new item with faulty packaging led to creating an "Andon Cord" for customer service. This would immediately disable purchasing for the product and alert the category manager when flagged. "It turns out that the Amazon version of the Andon Cord empowered the right people, those on the front lines who were talking directly to customers. It surfaced serious issues as soon as they were noticed. It proved once again that giving employees the right tools to solve problems and relying on their good judgment is a powerful combination."
Bottom line: "Output metrics show results. Input metrics provide guidance." It's useful to track revenue (output), but it's even more helpful to see signals within new customer counts, new customer revenue, existing customer revenue, etc. which help you prioritize the right actions today.
PART 2: THE INVENTION MACHINE AT WORK
Introduction to Part Two
Developing your company’s capacity for invention requires patience, frugality, and a willingness to "suffer the string of failed experiments necessary to get there." That said, invention is most important where differentiation matters, especially when it comes to offering a meaningfully better customer experience. Beware of only staying within a "skills-forward" approach to innovation, where you lean into existing core competencies. It's important to "exercise new muscles, never mind how uncomfortable and awkward-feeling those first steps might be." As your company grows, expect that failures will grow in scale as well; otherwise success wouldn’t have moved the needle.
Chapter 7: Kindle
[Detailed story of carving out a major success within digital media - and transforming the publishing industry - through the dedicated Kindle device for reading e-books.] Key takeaways:
When facing a big competitive threat (e.g. Apple eating away at CD sales with iTunes), respond by deciding how to organize an appropriate team and who should lead it to achieve the right result. Avoid jumping straight into what to build in response.
Fundamental tenet: change and adapt to meet shifting consumer needs, otherwise you're doomed. (Don't be Kodak.)
When they decided to prioritize Digital Media as a new business, they reported straight to Jeff. If left as a part of the physical media business, it would have never been a priority.
Avoid “copycat thinking,” especially in rapidly changing competitive landscapes where you’ll always be behind fast moving competitors.
"The six-page document and S-Team goals allowed Jeff to stay aligned on all major retail and marketplace programs and give feedback in an efficient manner [...] and the PR/FAQ process enabled him to spend weeks or months to gain alignment and clarity at a high level of detail on each project" within digital.
Single-threaded teams meant Jeff would a) know that teams were staffed correctly to deliver; and b) know that he wouldn’t need to waste time arbitrating resource conflicts.
Author’s mindset as his scope shrank as the team grew: "It was not a demotion. It was a signal that we were thinking big and investing in digital for the long term."
Trusted in the results of a well-run PR/FAQ process, even if it shows your existing core competency is insufficient. [E.g. for Kindle, they found they "needed to build an e-book store [...] deeply integrated with a reading device."] Making their own device was the only path to control the customer experience and maintain an edge over competitors.
The bar raiser process (to ensure new hires would be aligned with the leadership principles) proved essential as they hired leaders outside of their existing zone of competence.
Prioritize product decisions which lead to a better customer experience, even if expensive. (E.g. including Whisperenet, betting on E-Ink).
Chapter 8: Prime
[Detailed story of launching the Prime subscription for free shipping, which "single-handedly - and permanently - raised the bar for convenience in online shopping."] Key takeaways:
Long-term thinking and customer obsession helps when embracing the hard reality that your $800M investment in the current fulfillment network (circa 2004) wasn’t good enough.
Honest review of declining YoY sales growth - despite growth in online commerce still accelerating - showed a need to make a more dramatic change with long-term potential.
Prime was built on top of multiple experiments: a) Free Super Saver Shipping (which introduced "slow and free" as a shipping option); b) improvements to real-time telemetry around the time it would take to ship a product after order was placed (the controllable portion); and more.
Be wary of any either/or proposition where both results are mediocre. Figure out what a good result would be, and work backwards to find what needs to be true to deliver it.
Amazon’s core criteria for new marketing initiatives: 1) Be affordable (no financially unsustainable approaches); 2) Drive the right customer behavior (to buy more from Amazon); 3) Be a better use of funds than directly improving the customer experience (by lowering prices, improving selection, or improving convenience.)
Learn from adjacent examples, but through the reality of your business. [E.g. Loyalty programs were promising, but in retail "giving away either product or shipping fees always has a cost."]
Encourage bottom-up innovation: the idea for an "annual subscription fee in exchange for free shipping for a year" was initially proposed in a 1-page narrative by a principal software engineer as a solution to the bug-prone methods for reconciling shipping fees for super saver shipping.
Celebrate "strong general athletes," people who are "customer-obsessed, inventive, long-term thinkers who take pride in operational excellence and embody the leadership principles."
Be wary of the "institutional no," where a company chooses not to pursue something hard and risky because it seems safer to stay the course. (The cost of these errors of omission often go unnoticed until it’s too late.) A.k.a. “Don’t be Kodak.”
It’s useful to “walk the store” regularly (Jeff would do this every weekend) to note bugs, issues, and ideas for improving the customer experience. (e.g. "We should offer free shipping for expensive jewelry.") Exploring orthogonal ideas creates a library of rigorous thinking to build on.
With risky and controversial projects (like Prime), it’s important that people are willing to disagree and commit to the path forward.
Useful general framing, used for Prime: let’s "build a moat around our best customers."
Maintain awareness of how prior projects can be repurposed to move quickly on new initiatives. [E.g. Prime’s rapid launch built upon a) Fast Track shipping (which took 2 years to build and enabled precise estimates about shipping time); and b) a new subscription platform for an upcoming DVD rental product.]
Chapter 9: Prime Video
[Detailed story of how they worked through years of missteps, technical challenges, licensing struggles, and more to eventually launch Prime Video, the streaming service bundled with Prime.] Key takeaways:
When someone fails at a key project, respond with: "Why would I fire you now? I just made a million-dollar investment in you. Now you have an obligation to make that investment pay off. Figure out and clearly document where you went wrong. Share what you have learned with other leaders throughout the company. Be sure you don’t make the same mistake again, and help others avoid making it the first time."
Collect examples of different ways to launch products, etc. to prompt a future breakthrough. [E.g. they eventually remembered how Netflix side-stepped the cold start problem for their nascent streaming service by including it for free with the DVD rental subscription plan.]
Leadership must give the team explicit (or at least implicit) permission to consider a target investment figure to make their project a success. [E.g. until Jeff asked for a plan for investing a similar $30-40M into content, the team assumed that figure was a complete non-starter.]
Bundling streaming video with Prime brought strong differentiation in different markets: "any competitor might launch a Prime shipping clone, or they could potentially launch a new Netflix-type service, but it was unlikely that any one of them would be able to do both."
The bar raiser process (and emphasis on long-term performance targets only) helped them build an Amazon Studios team which was aligned with the “Amazon” way (e.g. rigorous approaches to green-lighting the most promising projects).
Chapter 10: Amazon Web Services (AWS)
[Detailed story of how they launched AWS, starting with developer-facing services which directly benefited the Amazon ecosystem, and then expanded into general purpose tools.] Key Takeaways:
Create conditions to be surprised by your customers. [This was especially true with early services they gave to developers, initially using Amazon product data. The resulting high adoption and innovation showed it was a promising area to explore.]
Key moment: realizing that ALL tech companies were struggling with problems Amazon had already figured out at scale. (The main challenges were well known; there were no web services available to do the 'undifferentiated heavy lifting.')
Engineering teams pushed back on the PR/FAQ process, but the process prevailed in the end. "With the volume of scale we wanted to achieve, unless the service was built right with the initial release, teams would spend all their time keeping the system running and would not be able to develop any new features." The process helped them achieve alignment early on.
For example, the working backwards process led them to adopt the "cost following" pricing model (based on usage, which would always keep their revenue aligned with their costs) instead of a subscription model. Articulating the cost drivers early on made it easy.
Conclusion: Being Amazonian Beyond Amazon
The principles in this book can be valuable and applicable within any company, of any size. To get started, consider [following the key section headers in the first part of the book.]
Appendices
Appendix A: Interview Feedback Examples
[Elaborating on the Bar Raiser process, the authors provide an illustrative example of weak feedback vs. good feedback. The latter had verbatim questions & answers and frank commentary on the candidate's performance in each section of the interview.]
Appendix B: Sample narrative Tenets and FAQs
Tenets help a) ground the reader back in the most important factors (usually discussed and debated last time); and b) guide decisions around key tradeoffs that need to be made. Example tenets (all quoted):
We don’t make money when we sell things. We make money when we help customers make purchase decisions.
When forced to choose between building something that’s convenient for customers or convenient for ourselves, we’ll choose the former.
We don’t let defects travel downstream. When we notice a defect, we will not rely on good intentions to solve the problem. We’ll invent and build systematic methods to eliminate that defect. [This later turned into a leadership principle.]
FAQs are great to queue up "issues for discussion or highlight important points or risks in your argument." Example FAQs (all quoted):
What were the biggest mistakes we have made last period, and what have we learned from them?
What are the key inputs for this business?
What is the single biggest thing we can do to move the needle in this business, and how will we organize to do just that?
What are the top reasons we should not do what we’re proposing today?
When push comes to shove, what are the things we won’t compromise on?
What’s hard about the problem we are trying to solve?
If our team had X more people or Y more dollars, how would we deploy those resources?
What are the top three new initiatives, products, or experiments our team has launched in the past X months, and what did we learn from them?
What dependencies do we have in our area today over which we wish we had control?
Appendix C: Timeline of Events in the Book
[Reference timeline.]