What is an instructional design audit?

An instructional design audit is a structured review of a course against the standards we already know make learning work: measurable objectives, assessments that match them, a sound instructional sequence, real practice, manageable cognitive load, and accessible content. The output is a ranked list of problems with a fix for each one.

It sounds formal. It isn't. It's the review most of us mean to do before a course ships and rarely have time for. Across the courses I have built and reviewed, my own included, the problems I've found are rarely exotic. They're usually the same handful, and a good audit catches them.

What an audit is not

An instructional design audit is easy to confuse with other reviews. They all matter, but they answer different questions.

A course can pass every other review and still fail this one. The content is accurate, every button works, the scanner is happy, and learners still leave unable to do the thing the course was for.

When to run one

• Before launch. The cheapest time to fix a design problem is before anyone takes the course.
• After heavy SME edits. Content changes late in a project quietly break alignment.
• When you inherit a course. Know what you're working with before you promise a timeline.
• Before a refresh. An audit tells you whether you need a patch or a rebuild.
• When AI drafted it. Generated courses look finished. That's exactly why they need a design review. (More on that in how to review an AI-generated course.)

The seven things a good audit checks

Each of these comes from a well-established body of learning science. None of them are new, which is the point: the standards exist, we just rarely have time to check our own work against them.

1. Are the objectives measurable? (Bloom's Taxonomy)

Every objective should describe something a learner can do that you could observe. "Understand phishing" can't be observed. "Identify three warning signs in a suspicious email" can. Check the verb, and check that it sits at the right cognitive level for the job. See Bloom's taxonomy verbs for learning objectives.

2. Do the assessments measure the objectives? (alignment)

This is the most common problem I see. The objective says "apply" and the quiz asks for a definition. Or a question tests something no objective ever promised. Map every question to an objective and every objective to at least one question. The full method is in how to align learning objectives and assessments.

3. Is the sequence sound? (Gagné's Nine Events)

Does the course gain attention, tell learners what they'll be able to do, connect to what they already know, present the content, then give guided practice with feedback? Missing prior-knowledge activation and missing feedback are two common gaps.

4. Do learners practice on real problems? (Merrill's First Principles)

Information alone doesn't build skill. Look for a real task or problem at the center, a demonstration of how it's done, a chance for learners to apply it themselves, and a way to connect it back to their own work.

5. Is the cognitive load manageable? (Mayer's Multimedia Principles)

Walls of text, narration that reads the screen word for word, decorative images that compete with the content, and long unsegmented screens all cost working memory the learner needs for the actual material.

6. Is the content accessible? (WCAG 2.1 AA, at the content level)

At the design stage you're checking what the content says and plans: meaningful alt text and descriptions for complex visuals, captions and transcripts, meaning that isn't carried by color alone, descriptive link text, logical headings, and instructions that don't depend on seeing or hearing. This complements an automated scan of the built course. It doesn't replace one.

7. Can a learner find their way? (usability)

Is it obvious where to start and what to do next? Are instructions given before the task, not after? Can learners tell how far along they are, and recover from a mistake?

How to run an audit by hand

1. Gather everything. Objectives, the content or storyboard, every activity, every assessment item, and descriptions of any media.
2. Check the objectives first. If an objective isn't measurable, everything downstream inherits the problem.
3. Map alignment. Three columns: objectives, practice activities, assessment items. Draw the lines. Anything left over is a finding.
4. Walk the sequence as a learner, screen by screen, against the events of instruction.
5. Check practice and feedback. Does every objective get rehearsed before it's assessed?
6. Scan for load and access issues. Text density, redundant narration, alt text, captions, color.
7. Rank what you found by impact on the learner, not by how easy it is to fix.
8. Write the fix, not just the flag. "Objective 2 isn't measurable" is a note. "Change it to: Classify a record into the correct retention category" is a fix someone can use.

A simple way to rank findings

Doing it faster

A thorough manual audit of a one-hour course takes a few hours, and it's hard to stay consistent across a team. That's why I built Pedagrade. You paste a course or upload the file, and in about a minute you get a scored report against all seven of these lenses, with every finding citing the principle behind it and a rewrite you can use. Then you apply your own judgment to what it found, which is the part of this job that matters most.