Notes: Melanie Perkins on Canva, Column B Thinking, and Crazy Big Goals

Four questions [Adler frame]

Q1 — What is this about? A founder’s account of building Canva from a school project into a $42B company, structured around three interlocking ideas: Column B thinking as a planning methodology (start from the desired end state, not the current situation); Chaos to Clarity as a creative process model (mastery requires passing through an embarrassing first phase); and crazy big goals as a motivational and strategic tool (goals that make you feel inadequate are the right size).

Q2 — How is it argued? Through retrospective narrative. Perkins dates each claim to a specific moment in Canva’s history: the column A/B framework to early-stage planning, the Chaos to Clarity model to Canva’s own sequential product development, the 2-year rewrite to a concrete near-collapse episode. Statistical anchors (100+ rejections, $3.3B revenue, 240M MAU, $50M donated) give scale. The argument is practitioner-first and anti-theoretical — she does not name strategic frameworks or cite prior art, but the Column B approach is structurally isomorphic to backward planning (Covey’s “begin with the end in mind”) and pre-mortem analysis.

Q3 — Is it true? The Column B framework is well-supported in planning research — goal-setting theory and backward design both confirm that starting from a desired end state produces better plans than iterating forward from current state. The claim that inadequacy signals correct goal size is directionally right but not always actionable: inadequacy can also signal a goal that is merely outside current competence rather than genuinely transformative. The 100+ rejection → iteration data claim is credible; the investor feedback loop she describes is a structured practice, not just tolerance for failure. The Chaos to Clarity model is consistent with skill-acquisition research (the competence ladder), though Perkins frames it more as creative courage than skill theory.

Q4 — What of it? For anyone planning a product or company: Column B thinking is immediately applicable — write the desired end state before listing resources and constraints. The pre-mortem variants (“wild success / terrible failure”) are under-used and give licence to name fears that remain implicit in most planning. The Chaos to Clarity frame is useful for managing the psychological toll of early-stage work: the embarrassing first attempt is not a signal to stop but the price of entry to mastery.

Glossary

Column A / Column B — A planning pair. Column A lists where you are now: current situation, resources, constraints. Column B lists where you want to be: the desired end state, independent of current constraints. The planning discipline is to define Column B first, without reference to Column A, and then work backwards to a path.

Chaos to Clarity — Perkins’s model of the creative and execution process. Every new idea begins in chaos: the vision is clear but the execution is opaque. The first attempts are embarrassing because the maker lacks the mastery the vision demands. Clarity emerges through sequential step-by-step work; it cannot be front-loaded. The embarrassment is not a signal to stop — it is the entry fee.

Crazy big goal — A goal large enough to make you feel completely inadequate when you first articulate it. Perkins argues that this feeling is the correct diagnostic: if the goal does not make you feel inadequate, it is too small. The motivating property is that progress toward such a goal is intrinsically meaningful because the distance to close is vast.

Wild success / terrible failure pre-mortem — A planning exercise. Before starting a project, write two futures: (1) it succeeds completely — what does that look like? (2) it fails completely — what went wrong? The exercise forces explicit confrontation with failure modes and makes the success criteria concrete.

AI walk — Perkins’s generative practice: a voice-recorded brain dump while walking, later fed into Canva Docs and structured by AI. The walking separates idea generation from editing; the AI structures without interrupting the generative phase.

Two-step plan — Canva’s long-horizon strategy: (1) build one of the world’s most valuable companies; (2) use that platform to do the most good. The steps are sequential, not simultaneous — the first funds the second.

Column B thinking: the argument

Perkins’s planning method: before listing resources, constraints, or current state, write the end state. Column B is written without reference to Column A. The value is that Column A thinking starts from what exists and projects forward; Column B thinking starts from what should exist and works backwards.

The practical test: if your plan begins with “we have X resources and Y time, therefore we can do Z,” you are in Column A. If your plan begins with “the world we want looks like Z, therefore we need to acquire X resources and Y time,” you are in Column B.

The Canva application: Perkins articulated the end state early — design tools democratised globally, accessible to anyone — without reference to whether Perkins (a student at the time) had any realistic path to that outcome. The Column B frame made the 100+ investor rejections legible as iteration data rather than disconfirming evidence.

Column B thinking is structurally identical to backward design in curriculum planning (Wiggins and McTighe’s Understanding by Design) and to Covey’s “begin with the end in mind.” Perkins does not cite these but the architecture is the same.

Chaos to Clarity: the process model

Perkins’s model of how mastery develops:

Chaos: vision is present, execution is absent. The maker sees clearly what the end state should be but cannot yet produce it.
Embarrassment: first attempts are poor relative to the vision. This gap is not a signal to abandon the project; it is a signal that mastery has not yet been acquired.
Sequential progress: clarity comes through step-by-step execution, not through better planning. You cannot skip the embarrassing phase.
Mastery: the gap between vision and execution closes. The work stops being embarrassing.

The implication for teams: the manager’s job in the Chaos phase is to protect the vision from premature abandonment. Embarrassing early attempts will be offered as evidence that the direction is wrong; the Chaos to Clarity frame re-classifies them as evidence that the direction is right but the execution is not yet mature.

The 2-year rewrite episode tests this model against reality. Canva undertook a full platform rewrite that was expected to take six months. It took two years. The team maintained morale through a rubber duck game board (physical objects representing progress). Perkins held the Column B vision through a period when the embarrassment phase was extended well beyond the original estimate.

Crazy big goals: the diagnostic

The claim: the correct size for a goal is the size that makes you feel completely inadequate. Goals that do not produce inadequacy are too small to be maximally motivating and too small to be meaningfully transformative.

The mechanism: inadequacy-scale goals force a qualitative change in approach. They cannot be achieved by doing what you already know how to do; they require acquiring new capabilities. This is motivationally distinct from achievable goals, which can be pursued via existing competence.

The 2050 vision wall: Perkins maintains a physical artefact in the Canva office that describes the world in 2050 as she wants it to look — basic human needs universally met, design democratised globally. The wall functions as a Column B anchor that survives quarterly planning cycles.

The caveat she does not raise: inadequacy as a diagnostic can conflate “goal that requires transformational capability growth” with “goal that is merely beyond my current competence and will remain so.” The framework does not supply a method for distinguishing between these. The investor rejection practice (structured feedback loops, systematic pitch revision) is the implicit answer, but she frames it separately from the crazy-big-goal claim.

The two-step plan

Canva’s stated long-horizon strategy:

Build one of the world’s most valuable companies — $42B valuation, $3.3B revenue, profitable 8 years, 240M MAU, 100M education users.
Use that platform to do the most good — GiveDirectly ($50M donated, $100M pledged).

The structure: the second step is contingent on the first. You cannot do the most good before you have the platform. The plan is sequential, not simultaneous. This resolves the common tension between “build value” and “do good” by staging them — the first is not in conflict with the second; it is its precondition.

The 100+ rejections as iteration

Perkins spent years raising Canva’s seed round and received more than 100 investor rejections. Her retrospective claim is that each rejection was structured feedback that improved the pitch deck. The pitch deck became a compounding asset.

The practice: after each rejection, identify what the investor objected to, revise the deck to address it, and use the next meeting to test the revision. This converts a demoralising experience into a product-design practice. The investor is the user; the pitch is the product.

The implication: tolerance for rejection is not the skill. The skill is structuring rejection as data and building a system to convert it into improvement. Perkins does not claim to have enjoyed the process; she claims to have made it productive.