Backward Design, Stage 3: From Outcomes to a Coherent Assessment Plan

By the time you reach Stage 3 of Backward Design, the hard conceptual work is supposed to be behind you. Stage 1 named the desired results — the outcomes, the transfer goals, the enduring understandings the course exists to produce. Stage 2 identified what would count as evidence that students got there. Now Stage 3 asks the seemingly practical question: what's the actual plan? What assessments, in what order, at what weight, so that every outcome is measured once, well, without three assignments accidentally testing the same thing and one outcome slipping through untested?

This is where a lot of otherwise thoughtful course design quietly falls apart. Not because faculty don't know their outcomes — they do — and not because they can't imagine good evidence — they can. It falls apart because a list of good assessments is not the same thing as a coherent plan of assessment. You can have four genuinely excellent assignments that, taken together, over-measure two outcomes, ignore a third, and exhaust your students with redundant work. Each piece is defensible. The whole is incoherent.

Stage 3 is the coverage problem. You have outcomes on one side and a semester's worth of graded work on the other, and the job is to make the second cover the first — completely, efficiently, and without waste. This playbook is about how to build that plan deliberately, instead of letting it accrete one inherited assignment at a time.

Why "I have good assignments" isn't a plan

The most common failure isn't a bad assessment. It's a set of assessments that were each chosen locally, without anyone ever checking them against the outcomes as a set.

Picture a course with six outcomes and the usual complement of graded work: two papers, a midterm, a final, and a semester project. Ask of each assignment, in isolation, "is this a good assessment?" and the answer might be yes every time. Ask instead, "which outcomes does the whole plan actually cover?" and you routinely discover three things at once:

• Redundancy. Both papers and the final are all, at bottom, measuring the same outcome — the writing-and-analysis one — because that's the outcome you know how to assess. Three assessments, one capability, a great deal of student and grading effort spent proving the same point.
• Gaps. An outcome about collaboration, or oral communication, or applying a method in an unfamiliar context, is measured by nothing. It was easy to imagine and hard to assess, so it fell off the plan without anyone deciding it should.
• Weight mismatch. The outcome you'd call the most important — the transfer goal, the thing you actually want students carrying out the door — is worth 15% of the grade, while a recall-heavy midterm is worth 30%. The grade weight tells students what matters, and it's pointing at the wrong things.

None of these are visible when you evaluate assessments one at a time. They're only visible when you look at the plan as a whole and check it against the full set of outcomes. Stage 3 is that check, done on purpose and up front, rather than discovered by accident three years into teaching the course.

Building the plan: cover every outcome once, well

Here's a sequence for turning a pile of outcomes and assessment ideas into a coherent plan. It assumes you've done Stage 1 and Stage 2 — you have outcomes, and you have a sense of what evidence would satisfy each one.

1. Lay outcomes and candidate assessments side by side. Outcomes down one axis, the assessments you're considering across the other. This is the same grid logic as an alignment matrix, used here as a planning instrument rather than an audit — you're building the map, not checking a finished one.
2. Assign each outcome a primary assessment. For every outcome, pick the one assessment that will be its main, deliberate measure — the place you'll point to and say, "this is where I know whether they can do it." Every outcome needs exactly one strong home. If an outcome doesn't have one, you've found a gap before the semester started, which is the only good time to find it.
3. Let secondary coverage be a bonus, not a crutch. An assessment can touch several outcomes lightly, and that's fine — reinforcement is real. But no outcome should rely only on secondary, incidental coverage. "It sort of comes up in the participation grade" is how outcomes go untested.
4. Hunt for redundancy and consolidate. If three assessments all serve as the primary measure of the same outcome, you have room to cut. Retire one, or repurpose it toward an outcome that currently has no strong home. This is usually where the plan gets both leaner and more complete at the same time.
5. Set weights to match stated priorities. Rank your outcomes by how much you actually care about them — the transfer goals near the top. Then check that grade weight roughly tracks that ranking. If your most important outcome carries the least weight, students will optimize for the wrong things, and they'll be right to.
6. Sequence for building difficulty. Order the assessments so earlier ones scaffold later ones. The final measure of your hardest outcome shouldn't be the first graded thing students touch. Coverage is about whether each outcome is measured; sequencing is about giving students a fair path to being measured well.

The test of a finished Stage 3 plan is simple: read down the outcomes and confirm each has exactly one strong home and no critical gaps; read across the assessments and confirm none is redundant or unmoored; then check that the weights point at what you said mattered. When those three things hold, you have a plan, not a pile.

A worked example, briefly

Suppose your consolidation reveals that Paper 1, Paper 2, and the final are all primarily measuring "analyze and evaluate arguments," while "apply the framework to a novel case" and "communicate findings to a non-specialist" have no strong home at all.

The fix isn't more assignments — it's reallocation. Keep one paper as the primary measure of analysis. Rework the second paper's prompt so its primary job becomes application to a novel case — same amount of student work, aimed at an uncovered outcome. And redesign the final so that part of it is a communication-to-non-specialists task rather than a fourth analysis essay. You haven't added load. You've taken three assessments that covered one outcome and spread them across three, and every outcome now has a home. That's the whole move: redistribute coverage before you add volume.

How the tool fits

Everything above can be done on paper, and it's worth doing on paper at least once so you understand the moves. The friction is that the plan lives across several artifacts — your outcomes, your assessments, their weights, their sequence — and keeping them coherent by hand means rebuilding the grid every time one thing changes.

TeachingsByDesign gives Stage 3 a home. The alignment matrix turns from an after-the-fact audit into a live planning surface: you map each outcome to its primary and secondary assessments and immediately see the coverage — which outcomes have a strong home, which are still orphaned, which assessments are doing redundant work. Because it sits inside the backward design workflow alongside your Stage 1 outcomes and Stage 2 evidence, the plan you build here is checked against the results you named earlier, not against a separate document you have to keep in sync.

And once the plan says this outcome is measured by this assessment, the rubric builder makes that assessment concrete — draft a rubric from the outcome itself, structured into criteria and performance levels, so the thing you planned to measure is actually measurable. The plan and the instruments that execute it live in one place, and you export the whole thing clean when it's done. What the tool doesn't do is decide which outcomes matter or how to weight them — that judgment is yours. It makes the coverage visible so the judgment has something to work on.

The bottom line

Stage 3 isn't about having good assessments. It's about having a plan where every outcome is measured once and well, no assessment is redundant, and the grade weights point at what you actually care about. Individually strong assignments routinely add up to an incoherent whole — over-testing what's easy to test, ignoring what's hard, and weighting the transfer goal below the recall quiz. The remedy is to build the plan against the full set of outcomes up front: give each outcome one strong home, consolidate the redundant, cover the gaps, and set weights to match your priorities. Do that, and Stage 3 stops being where the design falls apart and becomes where it finally holds together.

TeachingsByDesign turns your Stage 3 plan into a live coverage map — outcomes to assessments, gaps and redundancy flagged, rubrics drafted from the outcomes themselves, exported clean. Build a plan, not a pile. See how it works.