Everything from

O'Reilly Media

5 resources from O'Reilly Media we point founders to, and the questions each answers.

📖 Book
✓ Link checked Paid Intermediate

Why we picked it A good spec is not a feature dump, it is the product broken into user-centered slices, and that framing is exactly what Jeff Patton teaches. Story mapping gives you a way to lay out the whole user journey and then decide what is genuinely needed for a first version, which keeps you and your developer aligned on the why behind each feature. Read it before you write the spec, not after, and you will write a better one.

User Story Mapping: Discover the Whole Story, Build the Right Product

From O'Reilly Media by Jeff Patton ~324 pages

  • Map the whole user journey first, then slice out the smallest version that still tells a complete story: this is how you scope an MVP honestly.
  • Stories are a tool for conversation and shared understanding, not just a list of tickets to hand off.
  • Keeping the focus on user needs stops a spec from drifting into a pile of disconnected features.
Open oreilly.com →
Answers How do I write a spec for a developer when I'm not technical?
📖 Book
✓ Link checked Paid Intermediate

Why we picked it This is the book most working UX writers point to when they need a system, not just tips, for the exact text on buttons, errors, labels, and confirmations. Chapter 4 walks through UX text patterns for each of these components, so it directly answers how to phrase a button so people know what will happen and how to turn a vague error into a next step. The 2nd edition is current (2025) and covers writing copy in a world of AI-generated UI text.

Strategic Writing for UX (2nd Edition)

From O'Reilly Media by Torrey Podmajersky

  • Treat every button and error as a pattern with a job to do, then write to that job (buttons name the action, errors name the fix), rather than reaching for generic words like Submit or OK.
  • Keep button labels tight (Podmajersky's guidance is roughly three words max) and use an action verb so the label reads as what happens next.
  • Anchor your voice and word choices to product principles so copy stays consistent across every screen, not written fresh and randomly each time.
Open oreilly.com →
Answers How do I write clear button and error messages so users know exactly what to do next? How do I write microcopy for buttons, empty states, and error messages so the product feels friendly instead of robotic?
📖 Book
✓ Link checked Paid Intermediate

Why we picked it If the article gives you the idea, this book gives you the framework: pick the One Metric That Matters for your kind of business and your current stage, and judge your launch by that instead of a dashboard full of noise. It maps the right metric to each business model (SaaS, marketplace, e-commerce, media, and more) so you are not guessing which number is real. Use it as a starting point to choose the single number your launch is really trying to move.

Lean Analytics: Use Data to Build a Better Startup Faster

From O'Reilly Media by Alistair Croll and Benjamin Yoskovitz 440 pages

  • The One Metric That Matters forces focus: at any given stage you track one number, which stops you from hiding behind a wall of vanity charts.
  • The right metric depends on your business model and your stage (empathy, stickiness, virality, revenue, scale), so a launch that worked for one type looks different from another.
  • It is backed by 30-plus case studies and interviews with 100-plus founders and investors, so the guidance is grounded in what real teams measured, not theory.
Open goodreads.com →
Answers How do I actually measure whether a launch worked, beyond the vanity upvotes and traffic spike?
📖 Book
✓ Link checked Paid Intermediate

Why we picked it The One Metric That Matters chapter is still the canonical treatment of picking a single focusing number, and crucially it ties that choice to your business model and stage, which is the real question hiding behind free plus paid. It maps the right metric to model type (SaaS, marketplace, e-commerce and more) with public benchmarks, so you are not guessing what good looks like. Treat it as the reference you return to when your one metric needs to change as you move stages.

Lean Analytics: Use Data to Build a Better Startup Faster

From O'Reilly Media by Alistair Croll and Benjamin Yoskovitz Around 440 pages

  • The One Metric That Matters means picking a single number to rally around at each stage and ignoring the rest until it is solved, which forces a decision when free and paid both want attention.
  • Your right metric depends on your business model, so a freemium SaaS founder and a marketplace founder should not chase the same North Star.
  • It separates vanity metrics from real ones, useful when free-user counts look impressive but do not predict paid value.
Open amazon.com →
Answers How do I calculate my North Star metric when I have both free and paid users?
📖 Book
✓ Link checked Paid Intermediate

Why we picked it The reason people track the wrong active-user number is usually that they copied a metric from a different kind of business. This book is the canonical fix: it maps which engagement and stickiness metric actually fits each business model (ecommerce, SaaS, mobile app, media, user-generated content, marketplace) so you are not measuring a low-frequency product with a social-app yardstick. Read the chapter for your model, then define your active-user rule to match.

Lean Analytics: Use Data to Build a Better Startup Faster

From O'Reilly Media by Alistair Croll and Benjamin Yoskovitz

  • Different business models call for different core metrics; there is no universal definition of active, so borrow the one that fits your model.
  • The One Metric That Matters idea forces you to pick the single number that reflects real value at your current stage instead of a vanity dashboard.
  • Its stickiness and engagement sections give you a language for defining an honest active-user event rather than counting logins.
Open oreilly.com →
Answers How do I define and track weekly active users in a way that is honest for a low-frequency product?