Spark Curriculum Architecture

---
name: spark-curriculum-architecture
description: "The user wants a course, book, workshop, paid cohort, or any structured information product. Load whenever you hear \"I want to teach,\" \"I'm writing a book on,\" \"what should be in the curriculum,\" or \"help me outline this.\""
metadata:
  author: wayland
  version: "1.0.0"
  category: "spark"
---

# Curriculum architecture

## When to load this mode

The user wants a course, book, workshop, paid cohort, or any structured information product. Load whenever you hear "I want to teach," "I'm writing a book on," "what should be in the curriculum," or "help me outline this."

## Procedure

Curriculum is built backward. Design endpoint, then evidence, then experience. Four steps.

**1. Write the learner transformation in one sentence.** Format: "By the end, a [named learner in a specific situation] will be able to [observable action] under [conditions] within [timeframe]." The four slots are non-negotiable.

- Named learner: "a solo consultant who has done three free projects" beats "professionals."
- Observable action: "land five paying clients" beats "feel confident about sales."
- Conditions: "using cold outreach and one referral pass" — what they're allowed to do.
- Timeframe: a number with a unit.

If any slot is missing, ask. Once. If the user can't fill it, the course doesn't exist yet.

**2. Derive one to three enduring understandings.** Ideas the learner should still hold a year later — transferable insight that survives forgetting the specifics. Not topics. "How to write a cold email" is a topic. "A cold email is a one-on-one human gesture priced like a mass marketing channel" is an enduring understanding.

Rule of thumb: if the understanding can be stated as a sentence the learner could say a year later, it qualifies. If only as a verb phrase, it's a skill. Want both; understandings come first.

**3. Design the assessment before the content.** For each enduring understanding, write the evidence: what the learner produces, decides, or solves that proves they hold it. Two formats:

- **Performance task** — the learner does the thing in a realistic-enough context. Writes a cold email and sends it. Runs a discovery call. Drafts chapter one. Gold standard.
- **Decision artifact** — a choice under constraint with written reasoning. "Pick one of these three pricing strategies. Defend it in five sentences." Cheaper, still surfaces understanding.

A multiple-choice quiz tests retention of facts. Include retention checks; don't mistake them for the main event.

**4. Map modules to the assessment ladder.** Now sketch modules. Each module moves the learner across one assessment threshold. Two rules:

- **Cognitive level rises across the arc.** Per Bloom's revised taxonomy: remember → understand → apply → analyze → evaluate → create. First apply lesson lands inside the first quarter or the learner disengages.
- **One module, one threshold.** If a module moves two thresholds, it's two modules. If two modules share one threshold, one is filler.

Output: a numbered map with three columns — module title, cognitive level, assessment evidence. The TOC falls out at the end, not the start.

## Decision rules

- **No transformation, no curriculum.** Refuse the brief until the four-slot sentence is written.
- **Assessment before content. Always.** Lessons exist to enable assessments; assessments do not exist to test lessons.
- **Three enduring understandings caps a six-to-eight-week course.** More than that is a multi-course program; route packaging to Forge.
- **A module without a performance task or decision artifact is a chapter, not a lesson.** Books have chapters. Courses have lessons. Don't ship a course made of chapters.
- **First apply-level moment lands in the first 25% of the arc.** Otherwise the arc is front-loaded with theory and the learner will leave.

## Anti-patterns

- **Topic-first outlining.** Listing what you know and slicing it into modules. Produces comprehensive courses nobody finishes.
- **Comprehensive coverage.** Teaching everything you know about a domain. The course teaches the transformation; the rest is a reading list.
- **Assessment as afterthought.** "We'll add a quiz at the end of each module." Now you have lessons with no evidence. Build assessment in step three or it doesn't get built.
- **TOC-first.** Tables of contents anchor on topics; topics anchor on what the author knows; the learner gets a syllabus instead of a path.
- **Stacking remember-level lessons.** Three videos and a worksheet before the first thing the learner does. Drop-off is steepest right there.

## Before / after

**Before:** *"I want to teach my freelance copywriting system. Module one is the foundations of copy, module two is the niche-picking framework, module three is..."*

**After:** *"Transformation: a generalist freelance writer charging under $500 per project lands three retainer clients at $2,000+ per month within ninety days, using a single-channel outbound system. Enduring understanding one: retainer copy is sold on the writer's promise to a business outcome, not on the deliverable. Assessment: the learner sends three outreach messages to named businesses, each tied to one revenue outcome. Module one: pick the outcome (apply). Module two: name the businesses (analyze). Module three: write the message (create). Module four: send and iterate (evaluate)."*