Reading Notes: How Google and Amazon Do Product

Core of building a viable product at Google and Amazon

1. Define user needs and success metrics;
2. Deliver the most important needs at minimum cost;
3. Ship, get data feedback, set the next iteration goal.

Steps to build a product

1. Pick the right product direction. A good product serves a real need shared by many customers. Your job is a unique, meaningful way to meet it; everything in delivery should serve that mission.

2. Write documents Including a press release, a living FAQ, and a functional spec.

3. Design user experience Work from the user's view with design, iterate, and build clear, intuitive UX. Keep asking questions so design stays tied to the mission.

4. Start testing As lead, own bug triage and decide what ships vs. what must be fixed pre-launch.

5. Prepare to launch Know what success looks like — define metrics for product success.

6. Launch Launching well is more than uploading files — you need marketing and PR plans.

Chapter 1 — Mission and strategy: finding the right need

1. Actively solve customer problems Don't fixate on competitors and react passively.

2. Serve a real need shared by many customers When you amplify a problem, you reach more customers; solving it helps more people — that's how winners grow.

Building a strong mission:

Why a mission? Answer: A strong mission early reduces conflict overall.

How to build a mission?

1. Sparks interest To attract and align people, the mission must interest them.

2. Offers concrete, directional principles Direction in the mission clarifies the outcome you want.

3. Reflects a representative product or service, not everything at once.

Example: Amazon's personalization recommendations team mission: "Delight customers more." It fits all criteria.

How to set the right strategy:

Strategy is a rough plan using the company's unique strengths to win target users. One paragraph on how your product will stay more attractive than competitors long term for those customers. Cover three things: customer, company, competition.

Example

IMDb mission: Inspire video viewers.

Context:

• "Inspire" = background, exploration, even understanding friends' tastes.
• "Viewers" includes movie viewers; YouTube, Hulu viewers are in scope.
• For 20–40 working adults drowning in content, inspire them to choose, watch, and understand deeper.

Strategy

Working adults have limited time but rich networks, strong opinions, disposable income for content. We provide rich related info to inspire viewers and prompt feedback — a virtuous cycle for all users.

Summary: what IMDb should offer, why this company, competitive landscape, differentiation, and why this segment.

Chapter 2 — Product definition

1. Requirements

   1. Define users and needs; analyze competitors; be better and differentiated
   2. Brainstorm solutions
   3. Put product in customers' hands; listen

2. Press release A strategy-derived article for alignment and transparency.

   • Six elements:
     1. Product name, launch date, target customer, problem solved, how (keep it brief)
     2. From the user's view: what it is, when it ships, why it exists

3. Living FAQ Records disputes and important details.

4. Wireframes or flowcharts Visual product description for clearer discussion.

5. Presentation deck For VCs or execs: market opportunity (user scale), revenue opportunity (solution value), durable advantage (hard to copy).

   • Order:
     1. Current user state
     2. Then expand and hit key points fast

6. Functional spec

   • Includes:

     1. Intro (mission and strategy)
     2. Goals and priorities
     3. Use cases or user scenarios
     4. Prototype or wireframes
     5. API
     6. Timeline
     7. FAQ and open questions

   • Traits:

     • Why this product and what to build
     • Defines terms in the doc
     • If a goal seems off-mission, explain why it's a goal

   • Use cases vs. scenarios Use case: brief statement of required user actions. Scenario: narrative of how the user experiences the product.

   • Capacity planning Rough estimate of future usage for engineering. Build a table for annual/quarterly capacity. Estimate storage (posts, images, sizes) and traffic (visitors, dwell, pages per user). At Google also egress (data out) and ingress (requests). Most apps egress >> ingress; user-upload apps may be ingress-heavy — plan accordingly.

   • Timeline List major milestones: feature complete, trusted tester release; if engineering unknown, be conservative.

7. Product review Get team buy-in; hunt edge cases together.

8. Customer concept testing

   • Ship experimental features to some users; monitor usage
   • Show concept/prototype to existing or potential customers; collect feedback

9. Naming, pricing, revenue forecast

   • Three pricing bases: cost, value, comparison.
   • Comparison pricing:
     1. A sensible comparator;
     2. Assume elastic market — price down → volume up, price up → volume down (often true).
   • Revenue model:
     1. Show stakeholders this is a big opportunity — 3-year monthly revenue forecast
     2. Model combines market research, consumer intuition, math — deep insight
   • Estimate share of total market:
     1. Segment share (SMB, mid-market, enterprise)
     2. Reach via marketing: budget / keyword CPM rough estimate
     3. Revenue = price × growing users per period

Chapter 3 — Winning on user experience

• UX

  UX (User Experience) is how users complete tasks and how information is presented and optimized.

• Personas Named roles with salary and goals; use them to evaluate design.

• Evaluating design

  • UX checklist
    1. Most important task for the user?
    2. Simplest solution?
    3. Information architecture sound?
    4. Clear, understandable design?
    5. Consistent standards?
    6. Fewer clicks; can actions be reduced?

  Primary personas over everyone help prioritize.

  • Methods:
    • The more you ask of users, the less they can and will do — steep drop-off
    • SHE: Simplify, Hide, Embody
    • Rank content blocks by revenue potential
    • Personalize and be real-time; detail within reason
    • "Banner blindness": users ignore top-center like "whack-a-monkey" ads
  • Experience
    • If discoverability or usability is uncertain, test with real users
    • One primary button per screen
    • Different styles for different priority actions
    • In 3–4 step flows, show current step and total steps

• Working with designers

  1. Feedback as "As a [user type], I want…"
  2. Build consensus with questions; repeat business goals and priority when goals conflict

Chapter 4 — Project management: one simple plan, kept current

• A simple plan: task list + engineering estimates (time to complete). Sort by agreed feature priority; assign to team.
• Don't pad estimates per task; pad when setting release date.
• Balance load with eng lead; adjust tasks per release date.

Chapter 5 — Testing: personally review test plan and cases

• Column for bug severity if test fails (usually levels 1–4).
• Preconditions: setup before test.
• Postconditions: app state after task.

Chapter 6 — Quantitative analysis: three data types to collect

• Quant data supports or defends judgment.
• Goal metrics: progress toward targets.
  • At Google, key metric: 7-day active users — reflects product state; pick frequency to match product
  • Compare week over week
  • Often expressed as ratios
  • Assume noise; define acceptable variation band

Chapter 7 — Launch

1. Say no to changes During launch prep, frequently say no to new features, bugs, UX tweaks!

2. Launch checklist (iteration backlog) Ensure every follow-up is ordered and described.

3. Blog post State mission, target customer, problem solved — your "lede" in news terms.

Chapter 8 — Winning with teams (startup)

• Program managers glue and lubricate — connect functions, keep org running, smooth friction.
• Resumes highlight products shipped and financial or UX impact.
• Don't rent engineers — build an engineering team

Joining a new team

Don't open with "the product is a mess." Usually understand first before declaring disaster.

Chapter 9 — Winning with technology

Data layer: usually a database — customer records, etc.; SQL-like tools

Business logic: the brain; complex computation lives here.

Response service returns to decision service; decision service chooses what to do for Charlie.

Chapter 10 — Winning with communication

Email

1. Put the most important point first.
2. Make reason sentences clear.
3. Spend 20 seconds on greeting, courtesy, signature so info reaches the right audience.

Precise incremental expression Makes numbers scannable for fast readers. Example: change a number to: increase/decrease XXX (delta), from XXX (start) to XXX (end).

Five meeting types

1. Team meeting: open with metrics review. Know the exact message; deliver clearly, briefly; ≤30 minutes.

2. UX review: prototype is enough. Brief setup: why we're here, who users are, business goals.

3. Engineering review: empower eng lead, gather senior feedback, minimize demo time.

4. Brainstorming: each fishbone = a question; brainstorm answers, then causes; end brainstorm → critical analysis.

Running meetings well

1. Send minutes to stakeholders after.
2. Three purposes: solve problems, gather info, transmit info.
3. ≤15 min presentations; one message; tell a story; one-page summary; demo UX; listen deeply

Presenting well

Summary

1. One key message per deck. Two messages confuse.
2. Once the message is set, present around it.
3. People like stories — stories connect info to real life.

Tips

1. Set the pace. "Imagine…" and "What if…" pull attention, then the story hooks them.
2. Give a concrete example everyone grasps — audience needs a reference frame.
3. Problem stated clearly.
4. Show how the solution improves users' lives.

Chapter 11 — Winning decisions and negotiation

1. Introverts may speak less in groups — listen hard when they do.
2. "Seek first to understand, then to be understood." If you lean one way, say so; let others speak.
3. When they're not a jerk, you find the root. ~90% of conflict is poor communication.
4. To control strong reactions, identify the trigger.

Handling conflict

1. When triggered, cue yourself; return to neutral
2. Objective talk reduces provocation

Chapter 12 — Personal management

Energy management: build a daily schedule.

• Block the first 90 minutes for your hardest task — usually what you avoid and delay most.