Easier Products

Formulas#

• Ease of selling = known category × clear outcome × short time to value.
• Build cost ≈ features × environments × integrations. Cut any one of the three and the whole product shrinks.
• Support cost ≈ users × complexity × how different their setups are.
• The one-sentence test: if it takes more than one sentence to say what it does, it'll be harder to sell.

What to make (1–15)#

1. A known category > a new category. Nobody searches for a thing without a name.
2. A better version of an existing habit > a new habit.
3. Used inside work they already do > adds new work.
4. One clear input → one clear output > an open-ended platform.
5. A result you can show in a screenshot > an abstract result.
6. The missing piece in a small, fast-moving market > a rebuild of a big product.
7. Used as often as it's billed > a monthly bill for a yearly task.
8. Digital, with near-zero cost per extra sale > physical or service-heavy.
9. A template, kit, dataset or course as the first product > a full app. It's faster to ship and it tests the audience.
10. Quality you can judge yourself > quality that needs expertise you don't have. In the AI era, judgment is the bottleneck.
11. Something you'd use every day > something only other people need.
12. Helpful but not critical at the start > mission-critical (payments, health, security). Less can go wrong, and less of it is your liability.
13. Async > real-time (chat, video, multiplayer).
14. Mostly reading data > heavy writing with complex sync.
15. Low-sensitivity data > regulated data (health, finance, children).

Scope (16–28)#

16. The smallest version that fully solves one job > a big version that half-solves many.
17. One platform > every platform at launch.
18. Web > native apps with store reviews.
19. One integration done well > ten done halfway.
20. Runs in an environment you control > runs on thousands of unknown setups.
21. Few settings > endless configuration.
22. One kind of user > roles and permissions.
23. Useful to a single user > useful only once many people join. Avoid the cold-start problem.
24. Useful with the data they have on day 1 > needs months of data first.
25. One-sided > a two-sided marketplace with a chicken-and-egg problem.
26. Standard formats (CSV, Markdown, email) > proprietary formats.
27. Built on stable, official APIs > built on scraping.
28. Features you can remove later > features you must promise forever.

Building (29–40)#

29. Skills you already have > learning a new stack while shipping.
30. Reusing your own past code and designs > starting from zero.
31. Buying the commodity parts (login, payments, email sending) > building them.
32. AI writes the grunt code and you judge it > writing every line, or trusting AI blindly.
33. A spec from real customer requests > an imagined roadmap.
34. Shipping weekly > big releases.
35. One app, one codebase > many services.
36. One database > several.
37. Tests on the money paths > testing everything, or nothing.
38. Standard UI patterns > custom interactions.
39. A ready design system > custom design from scratch.
40. Using it yourself every day > testing only before release.

Selling and support (41–50)#

41. Explained in one sentence > needs a demo to understand.
42. A 60-second demo > a 30-minute demo.
43. Value visible before payment > "trust me" products.
44. Self-serve setup > setup calls, unless the price pays for them.
45. An obvious unit to charge by > complicated pricing.
46. Easy to compare with a known product > nothing to compare it to.
47. Output that carries your brand > invisible output.
48. Few ways to break > many failure modes.
49. Mistakes can be undone (undo, versions) > irreversible actions. Recoverable mistakes mean fewer panicked tickets.
50. AI makes it easier to build but not easier to replace > a thin wrapper that the next AI model will absorb.