Every syllabus makes a set of promises. They sit near the top, under a heading like "Learning Outcomes," and they tell the student — and the accreditor, and your future self — what this course will make possible. Students will evaluate competing theoretical frameworks. Students will design an experiment and defend its methodology. Students will communicate technical findings to a non-expert audience. Five to eight of them, usually. Confident, verb-forward, aspirational.
Then there's the other half of the course: the assessments. The midterm, the final, the two papers, the lab practical, the participation grade. These are the things that actually generate the marks in your gradebook. They are what students prepare for, stress about, and remember.
The quiet scandal of course design is that these two halves are rarely checked against each other. The outcomes were written during a redesign or an accreditation cycle. The assessments accreted over years — a midterm inherited from the person who taught the course before you, a final that's easy to grade, a paper prompt you liked in 2019. Nobody ever lays them side by side and asks the two questions that matter: Does every outcome get assessed by something? And does every assessment trace back to an outcome?
When you do lay them side by side, you almost always find a gap. An outcome the course promises but never measures. Or an assessment worth 30% of the grade that maps to nothing you claimed to teach. The tool for finding both is the alignment matrix, and it's the most useful hour of course design most faculty have never spent.
What an alignment matrix actually is
An alignment matrix is a grid. Outcomes run down the rows. Assessments run across the columns. In each cell, you mark whether that assessment provides evidence for that outcome — and, if you want to be rigorous, how strong that evidence is.
That's it. There's no proprietary framework here, no jargon to memorize. It's a spreadsheet you could build in ten minutes. The discipline is in reading it honestly, which is harder than it sounds, because the whole point is to expose work you'd rather believe is already done.
Here's a stripped-down version for a research methods course:
| Midterm | Critique paper | Final project | Participation | |
|---|---|---|---|---|
| O1: Evaluate published methodologies | ✓ | ✓✓ | ||
| O2: Design an original study | ✓✓ | |||
| O3: Interpret statistical output | ✓ | ✓ | ||
| O4: Communicate to non-experts | ✓ (weak) |
A single ✓ means the assessment touches the outcome. A double ✓✓ means it's a primary, deliberate measure — the place you'd point to and say this is where I know whether they can do it. A blank means no evidence at all.
Read the grid, not the syllabus. The syllabus says the course does four things. The grid tells you which of those four it can actually prove.
Reading the two failure modes
There are exactly two things to look for, and they correspond to the two questions above.
Look across every row. A row with no strong mark — or worse, no mark at all — is an orphaned outcome. It's a promise the course makes and never keeps. In the example above, O4 ("communicate to non-experts") is assessed only weakly, buried inside a participation grade that mostly rewards showing up. The course claims it. Nothing really tests it. If a student graduated unable to communicate to non-experts, this course would never have caught it, and neither would the transcript.
Orphaned outcomes are the most common finding, and they're almost never malicious. They happen because program-level outcomes get pasted into a syllabus wholesale, or because an aspirational outcome was added during a redesign and the assessments were never updated to match. The outcome is real. The evidence for it is missing.
Look down every column. A column that doesn't map cleanly to any outcome is an unmoored assessment. This is the final exam that tests recall of material no outcome mentions. It's the assessment that survives from a previous version of the course, or the one that's there because it's easy to grade, not because it measures anything you set out to teach. Unmoored assessments are quieter than orphaned outcomes but arguably worse, because students are spending real effort — and receiving real grades — on work that doesn't serve the stated purpose of the course.
The phrase in the title — the outcome your final exam never tests — is the intersection of these two failures. You have an outcome with no strong assessment (an orphaned row), and simultaneously an exam that's busy testing something else entirely (an unmoored column). Both problems are invisible in the syllabus. Both are obvious in the matrix.
Building one for your own course
You don't need a tool to do this. You need an afternoon and the willingness to be honest. Here's the sequence.
- List your outcomes as they actually appear on the syllabus — not the improved versions in your head. The gap between the two is itself useful information.
- List every graded assessment, including the small ones. Participation, quizzes, and discussion posts all belong here, because they carry grade weight and students treat them as signals about what matters.
- Fill in each cell deliberately. For every assessment, ask: which outcome does this provide evidence for? Not "which outcome is it vaguely related to" — which outcome would a reasonable colleague agree it measures. Use a strong mark only when the assessment is designed to surface that specific capability.
- Scan the rows for orphans. Any outcome without a strong mark is a promise you can't currently keep.
- Scan the columns for the unmoored. Any assessment that doesn't earn at least one clear mark is grade weight spent on something outside your stated purpose.
- Fix the smallest gap first. You rarely need to redesign the course. Often it's one assessment reweighted, one prompt rewritten, or one outcome retired because it was never really this course's job.
A note on honesty, because it's the whole game: the temptation is to mark a cell ✓ because you hope the assessment gets at the outcome. Resist it. A five-page essay does not automatically assess "critical analysis" just because analysis is one thing an essay could contain. Mark evidence, not aspiration. The matrix is only worth building if you let it tell you something you didn't want to hear.
How the tool fits
Everything above works with a spreadsheet. The reason we built the alignment matrix into TeachingsByDesign is that the manual version breaks down at exactly the moment it gets useful — when the course changes.
You reweight an assessment, add an outcome mid-semester, or fold two papers into one, and your hand-built grid is instantly stale. So you stop updating it, and within a term you're back to flying blind. The matrix only pays off if it stays live.
In TeachingsByDesign, you enter your outcomes and your assessments once, map the relationships, and the matrix flags the gaps automatically — it surfaces the orphaned outcomes (an outcome nothing assesses) and the unmoored assessments (an assessment tied to no outcome) without you having to squint down every row and column yourself. When the course changes, the gaps re-flag. And because the matrix lives alongside your curriculum map and your outcomes, the alignment check isn't a separate artifact you maintain by hand — it's a view of the course you already built. When you're done, you export it clean, so the same grid that caught your gap is the one you hand to the accreditation committee.
The tool doesn't decide what's worth teaching or how to close a gap once you've found it. That's your judgment, and it should be. What it removes is the reason the check never gets done: the tedium of maintaining a matrix by hand until it rots.
The bottom line
Your syllabus lists what the course promises. Your gradebook records what it measured. The alignment matrix is the one place those two things get checked against each other — and almost every course, checked honestly, has an outcome nothing assesses and an assessment mapped to nothing. Both are invisible until you build the grid. Outcomes down the rows, assessments across the columns, evidence in the cells; then read across for orphaned promises and down for unmoored work. You can do it in an afternoon with a spreadsheet. The only hard part is keeping it alive as the course changes — and that's the part worth automating.
TeachingsByDesign builds the alignment matrix from your outcomes and assessments, flags the gaps automatically — an outcome nothing tests, an assessment mapped to nothing — and exports it clean for review. Find the gap before your accreditor does. See how it works.