| name | tobi-lutke-business-leadership |
| description | Tobi Lütke's (Shopify founder/CEO) leadership and systems-thinking operating system — rerunnable decision functions over roadmaps, legitimacy banking, subtraction as the founder superpower, Goodhart-proof metrics, compressed change management (the Toby tornado), and maximizing human potential |
| version | 2.0 |
| format | completion-engine |
| workflows | 3 |
| source | claude.ai export 2026-07-01 |
Tobi Lütke — Business Leadership & Systems Thinking
Tobi Lütke built Shopify from a snowboard store into a $60B+ company by running it like a programmer: strategy as a rerunnable function (customer model + capability model + current reality), founder legitimacy as a bank account you deliberately spend, and subtraction — not addition — as the leader's core service. His system treats best practices as "average, risk-adjusted," metrics as instruments rather than goals (Goodhart's law = overfitting), and human potential as the actual product: nobody on the planet is close to their maximum.
Available Workflows
| # | Workflow | Produces |
|---|
| 01 | recompute-priorities | Rerun of the decision function after a reality shift — recomputed priority stack + explicit subtraction list, North Star protected, legitimacy spend planned |
| 02 | audit-orthodoxy | First-principles audit of where best practices and goal-metrics cap performance — Goodhart scan, bureaucracy triage, differentiated bets with subtraction guarantees |
| 03 | compress-the-change | Toby-tornado change package — steel-manned kill/pivot decision, divergence-point analysis of disagreements, refounding recharter, potential-calibrated feedback |
Quick Reference
- Keep the main thing the main thing: protect the North Star, never the current tactic; obstacles change routing, not direction
- Roadmaps are overrated: store decisions as functions with named inputs; when an input changes, recompute — don't defend the stale output
- Yes-tax: saying yes to one thing says no to every other thing you could have done with that time; price every yes that way
- Sediment test: individually good decisions pile into something no one can steel-man — periodically ask "would we start this today?"
- Legitimacy banking: founding-story retellings deposit; hard changes withdraw; spend it on subtraction and direction, never on what a plan could legitimize
- Best practices = don't take risk: use orthodoxy as input, never ceiling; a miss is "the successful discovery of something that didn't work"
- Goodhart = overfitting: any metric that becomes a goal ceases to be a good one; metrics are support instruments, taste ships
- The trilemma: on a suspect project you're incorrect (understand why), correct (unfair to let people continue), or ignoring it (abdication — never valid)
- Refound, don't cancel: people from a killed project are founders of the next version
- Divergence point: disagreement is a map to unstated foundational assumptions; courage is rarer than IQ — reward it visibly
- No speed limit: hold people to their potential, not current-level-plus-epsilon; name the moment they chose path B out of convenience
- Think 100 years forward, then backwards: "what would we want to have done 20 years ago on this?"
Execution Prompts (structure-pure v2)
3 deterministic practitioner prompts — each carries an Output Contract, Output Skeleton, and Quality Gate. When a deliverable matches one, Read it and honor its contract instead of improvising the output shape.
- Tobi Lütke — Orthodoxy Audit —
skills/tobi-lutke-business-leadership/references/prompts-v2/orthodoxy-audit.md
- Tobi Lütke — Priority Recompute —
skills/tobi-lutke-business-leadership/references/prompts-v2/priority-recompute.md
- Tobi Lütke — Toby Tornado Change Package —
skills/tobi-lutke-business-leadership/references/prompts-v2/toby-tornado-change-package.md
