Break the frame

Break The Frame

A lateral-strategy skill. It produces angles the user could not reach from inside their current frame, by running a fixed set of inversion operators against their specific situation.

The value of this skill is DISCIPLINE, not cleverness. Good output does not come from "being creative." It comes from taking each operator and forcing it onto the user's concrete assets one at a time, including the operators that hurt. Free-associate and you get generic startup advice; run the operators and you get angles bolted to their reality.

Sibling skills: step-back-as-user (re-enter from the user's side, UX) and step-back-as-maintainer (re-enter as a future code reader). This one re-enters through the strategy / identity / business-model lens, by inverting what the user holds fixed without noticing.

Step 0: load the situation before inverting (gate)

Cold input gives generic output, and a bare noun like "a CLI tool" or "my startup" is cold, not loaded: you are ready only when you can state, in one sentence each, what the artifact does, who uses it, and what the user wants. Before generating angles, confirm you have:

• Concrete assets: what actually exists. The artifact, the engine underneath it, the user's distinctive skill, the data, the audience, the brand.
• The hidden premise nobody questions: the "of course it's a ___" the whole frame rests on (e.g. "of course it's a consumer app", "of course the buyer is the one who uses it", "of course I sell the product, not the byproduct").
• The sacred cow: the one asset the user loves most and would resist removing.
• Goal and constraints: money? career? a feeling? a deadline? solo build?

If the conversation already established these, proceed. If you are cold, ask for them in one batched message and stop there: generate zero angles that turn, and do not soften the gate by tossing in "a few to start." Keep the ask to about four questions, leading with what it does and who uses it (without that, every operator comes out generic). Never invert a blank.

The operator library

Apply at least 5. Aim every operator at the user's most DISTINCTIVE asset (the thing only they have: usually the engine, the data, or the method), not their most VISIBLE one (usually the format or the artifact); an operator aimed at a commodity asset produces a commodity angle. Pick the operators that bite hardest; each one that bites becomes an angle.

1. Axis swap: change the primary dimension. Space→time, ranking→randomness, now→history, global→hyperlocal.
2. Subject flip: turn the lens off "the world / others" and onto the user THEMSELVES. A discovery tool becomes a mirror or identity test (people overpay to see themselves).
3. Customer flip: sell to a different actor. Consumer→producer, demand→supply side, the individual→the institution above them (employer, platform, nation, brand).
4. Recategorize: same artifact, different product category. App→furniture / ambient object, tool→toy, feature→ritual, product→gift.
5. Kill the sacred cow (MANDATORY at least once): remove the most-loved asset. What is left? Is the attachment trapping the judgment? This is usually the uncomfortable angle.
6. Move what is monetized: charge for a different layer. The byproduct not the core, the person not the product (salary / audience / consulting), the distribution, the data exhaust.
7. Constraint-as-feature: take the worst limitation and find the market where it is ideal (messages that auto-delete are useless as records, perfect for candor).
8. Time-horizon flip: one-shot vs ritual/habit, ephemeral vs heirloom. Stickiness usually comes from ritual, not utility.
9. Emotion / status engine: reframe utility into identity, status, nostalgia, belonging. The things people pay for without justifying.
10. Subtract to one: strip everything but the single most delightful moment. Is THAT the whole product?

Output contract

• Open with one framing line: this is to shake the frame, not to be right.
• 5 to 7 numbered angles. Each angle carries:
  • a punchy imperative title (the move),
  • the move made CONCRETE in their situation (one or two sentences),
  • the stake: who has what on the line (money / time / reputation / identity),
  • one line of "why you didn't see it": what in the frame kept it hidden.
• At least one angle is uncomfortable and attacks the sacred cow.
• Close by killing analysis: ask for the raw gut reaction ("which one made you go 'oh?'"), then give YOUR honest bet on which one or two have the most in them, and the one they are most likely avoiding.

One worked angle (illustrative: derive your own from THEIR assets; do not reuse this domain, operator, or move)

Operator: kill the sacred cow, aimed at a solo consultant's pricing.

• Kill the hourly rate.
• Concrete: they bill by the hour and are proud of a transparent rate card. Drop it and sell one fixed-price outcome instead, priced on the value of the result, not the hours it takes.
• Stake: their predictable income and their whole professional identity ("I'm a $200/hr specialist") ride on the hourly model.
• Why you didn't see it: "of course consultants bill for time" was the unspoken definition, so pricing the outcome was never on the table.

Guardrails / failure modes

• Generic = failure. Self-test every angle: swap in a different business and reread the title and move. If it still makes sense, you anchored to the operator, not the asset; rewrite until the user's specific asset is load-bearing in the sentence, or cut it. "Make it a physical object / make it searchable / turn it into a course" fail this test by default.
• All-safe = failure. The sacred-cow angle must assert the beloved asset is CAUSING the stuck state, not merely that it could go. "It's great, but also try…" is a hedge. Test: would the user wince and want to argue back? If they would nod along comfortably, sharpen until it indicts the cow.
• Name-dropping is not an angle. "Be like Spotify Wrapped" counts only if you say what it does to THEIR specific frame and stake.
• Shape over volume. 5 to 7 sharp angles beat a dump of 20.

What this is not

• Not a plan or a recommendation. It opens doors; it does not pick one and build it. Once an angle lands, hand it off to your usual planning, pressure-testing, or prototyping step.
• Not a feasibility study. An angle can be half-impossible and still break the frame usefully.
• Not a pep talk. The point is new sightlines, not reassurance.

When to skip

• The user already has a direction and wants to execute: do not re-open the frame.
• The question is narrow and concrete: a UX moment goes to step-back-as-user, code legibility to step-back-as-maintainer.
• The user wants validation of a choice already made, not alternatives.