M · Recipe collection
Quick Recipes
Short methods that combine cognitive moves. Try one, then check whether it helps.
Choose a recipe in the practical Guide →Choose a Quick Recipe
- Ban-the-MapWhen your current framing keeps producing the same answers, ban one of its core elements for five minutes and build three replacements before comparing.
- Noun to VerbWhen “I am disorganized” or “they lack commitment” ends the conversation, describe what actually happens. A process gives you places to look and something specific to try.
- Person to HandoffWhen a discussion keeps returning to who dropped the ball, follow the work across the nearest handoff. Make the information, timing, and responsibility visible enough to repair.
- Event to LoopWhen the same problem keeps returning, ask whether the response helps recreate it. Draw the path back to the starting point, then check whether that feedback actually exists.
- Snapshot to MovieWhen today’s picture makes a problem look permanent or mysterious, reconstruct what happened before it. A short sequence can reveal a turning point, a delay, or a changing direction.
- Average to DistributionWhen “it usually works” clashes with someone’s actual experience, look behind the average. Show individual cases, the spread, and any meaningful differences between groups.
- Category to ContinuumWhen “beginner or advanced” or “ready or not ready” does not fit the cases in front of you, place them along a meaningful scale and inspect the boundary.
- Outcome to MechanismWhen a plan says what you hope will improve but not how, connect the proposed action to the result through concrete intermediate steps. Then question the weakest link.
- Boundary SweepIf your carefully studied system keeps surprising you, the missing relationship may sit just outside its border. Briefly include what feeds it, what follows it, a neighboring actor, and an effect it sends elsewhere.
- 10x In / 10x OutIf you cannot locate the useful level of detail, zoom roughly tenfold inward and outward. Compare what a smaller unit and a larger system reveal, keeping complementary views when they explain different parts.
- Reverse the ArrowWhen reasoning forward produces too many branches, begin at a concrete endpoint and work backward. Ask what must have been true just before it, preserving alternative routes and distinguishing necessary conditions from guarantees.
- Change the ObserverWhen everyone explains the situation from the same seat, inspect it from another position. Ask what the recipient, operator, newcomer, or future user can actually see and needs to do.
- Missing ColumnWhen your comparison looks tidy but still cannot support a choice, ask which missing fact could change the ranking. Add that column, then see whether it matters.
- Minimal PairWhen you cannot explain why two cases should be treated differently, make them almost identical. Change one relevant feature and ask whether your judgment should change with it.
- Constraint SandwichWhen “make it better” produces attractive but unusable ideas, put the improvement between what must stay true and the limits you actually face. Then compare feasible changes.
- Table Before ProseYou keep rereading several paragraphs to compare the same few details. Pull the entities into rows and their shared attributes into columns, then keep short prose for the exceptions that the grid cannot faithfully carry.
- Draw Before ExplainYou can name the parts, but explaining how they depend on one another ties your sentences in knots. Sketch nodes and labeled arrows first, then remove any connection you cannot define and use the drawing to guide the explanation.
- Three TranslationsWhen an explanation keeps slipping away, show the same content in three forms. Try the same task with each and compare what becomes easier to understand or do.
- State LedgerEvery interruption sends you back to reconstructing the work from scattered clues. Put the current state, open questions, decisions and next action on one persistent page, so your representation of the work survives the break.
- Manipulable ToyWhen a plausible idea still has vague consequences, make a tiny version you can change. Move one piece or vary one input and watch what your assumptions imply.