Think Week

Part III · Choosing a bet · Chapter 10 · 4 min read

Speed vs safety

Which moves pay the first dollar fastest, and which ones last? The two rankings are almost opposite, so you have to chain them.

In this chapter
  1. Four shapes of effort → revenue
  2. Ranking 1: fastest first dollar
  3. Ranking 2: most likely to last
  4. The durability test
  5. What the two rankings show

Four shapes of effort → revenue#

  • Linear: each hour of work brings money soon, but income is capped by your hours.
  • Spike → fade: fast money that shrinks as copycats or the platform catch up.
  • Step: nothing comes in until the product is good enough, then sales start.
  • Snowball: slow for months, then each week's work stacks on the last.

Ranking 1: fastest first dollar#

# Move First dollar Curve Compounds? Why
1 11. Productize a service days linear no, until you automate you sell before you build
2 12. Arbitrage days–weeks spike → fade no the gap closes
3 2. Narrow weeks (if the product exists) linear yes the work is outreach: ConvertKit emailed 100 bloggers, got 5–7 customers, repeated
4 5. Package 2–6 weeks spike → fade no HeadshotPro made $100k in 2 weeks; the window closes
5 4. Cross (X + Y) 1–3 months linear → snowball yes a niche is easy to reach; v1 is small
6 9. Shovels (small) 1–3 months linear some big shovels take 6–24 months
7 3. Move (small) 1–3 months linear yes big versions need capital
8 6. Invert 1–6 months step yes fast if the flip is small (pricing, one feature)
9 10. Boring / small 3–6 months linear, steady yes learn the niche first; then customers rarely leave
10 1. Subtract 3–9 months step yes a whole end-to-end product first
11 7. Rebuild on a shift 6–24 months step → hockey stick yes Gamma founded 2020, took off 2023
12 8. Own the query 6–24 months snowball strongly slow start, then speeds up every other move

Ranking 2: most likely to last#

No move is 100% safe; a product is always a bet. What can be close to safe is the asset you build while making the product: trust, a customer list you own, deep niche knowledge, data a business depends on. So choose moves where every week leaves an asset behind.

# Move Safety What keeps it alive What kills it
1 10. Boring / small high you become the system a business runs on selling a thin tool instead
2 8. Own the query high trust gets scarcer as AI floods content audience that comes via Google or a marketplace
3 6. Invert high the leader can't copy without hurting its revenue the gap was just "they haven't built it yet"
4 4. Cross (X + Y) medium-high combining systems and data the combination is only knowledge, which AI now has
5 3. Move medium-high real local advantages the only difference is language
6 11. Service / outcome medium accountability, relationships customers do it themselves with AI
7 2. Narrow medium trust inside the segment positioning is easy to copy
8 1. Subtract medium owning data, payments or network only advantage is "simpler"; AI makes complex tools easy
9 9. Shovels medium-low (solo) demand grows with AI small shovels are just information
10 7. Rebuild on a shift low (solo) huge ceiling a race against funded teams and the AI labs
11 5. Package low — the next model release
12 12. Arbitrage lowest — the gap always closes

The durability test#

Five questions for any idea. Each "yes" makes it safer:

What the two rankings show#

  1. There's a trade-off no move escapes. Moves at the top of the speed list bring money now, but it's capped by your hours or it fades. The safest moves are slow. None gives both.
  2. So run two engines. A fast one (11, 2 or 5) brings cash and teaches you what people want. A slow one (8 or 4) compounds. It works best when the fast one feeds the slow one: every client job becomes a case study; every conversation shows the next product.
  3. The curve isn't fixed; the order you do things in sets it. Any "step" move becomes gradual if you sell before you build. Ghost pre-sold its launch on Kickstarter. ConvertKit's founder moved customers' lists by hand.
  4. Move 8 is slow once, then makes everything else fast. TypingMind made $22k in its first week because its founder had spent years building an audience.

The chain

Start as a service (11) in a boring niche (10) → turn what repeats into a product → publish what you learn along the way (8).

The service gets the first dollar in days. The niche gives safety. The product becomes what the customer's business runs on. Publishing turns work into trust you own. It's the classic path of Basecamp, Mailchimp and most vertical software.

Sell → deliver by hand → automate

This order turns a step curve into a gradual one.

Takeaways

  • Fast moves are unsafe; safe moves are slow. Chain them, don't choose.
  • The product is a bet. The assets you build while making it don't have to be.
  • Four or five "yes" on the durability test ≈ not a bet.
  • Sell before you build, do it by hand, then automate what repeats.

Sources: ConvertKit batches · HeadshotPro · TypingMind first week