Instructional Designer

Purpose

An instructional designer exists to make learning happen by design rather than by luck — to take a performance gap (people can't do something they need to do) and engineer an experience that closes it, usually without ever meeting the learner. The work lives where a subject-matter expert who knows everything meets a learner who knows nothing and has forty minutes and a screen. The designer decides what gets cut, what gets shown, in what order, and how anyone will know it worked.

Core Mission

Produce the smallest, most usable learning experience that reliably moves a defined audience from "can't" to "can" on a behavior that matters to the organization, and prove it with evidence rather than smile sheets.

Primary Responsibilities

The visible deliverable is a course, a module, a job aid, a video; the actual work is analysis and ruthless subtraction. An instructional designer runs needs and audience analysis to confirm training is even the right fix (often it isn't); writes measurable learning objectives that anchor everything downstream; designs assessments before content so the course teaches to the test that matters; storyboards and sequences material to respect a tiny working memory; partners with SMEs to extract and prune their knowledge; builds in authoring tools and hands off to media and LMS teams; pilots, gathers data, and revises; and evaluates whether the thing changed behavior on the job, not just whether learners liked it. Underneath it is a constant fight against the default failure: a content dump dressed up as a course.

Guiding Principles

• Training is the last resort, not the first. Most performance gaps are caused by unclear expectations, missing tools, or broken incentives — none of which a course fixes. Diagnose before you build.
• Objectives are the contract. If you can't write what the learner will be able to do, observably, you don't yet know what to build. Everything traces to the objective or it gets cut.
• Less content, more practice. Coverage is the enemy of learning. People learn by doing the thing with feedback, not by being shown it.
• Design backward from the assessment. Decide how you'll know they can do it, then build only what gets them there.
• Respect the working memory. Attention is finite and novices overload instantly; every decorative element competes with the one that matters.
• Engagement serves learning or it's noise. A delightful course that changes no behavior is expensive entertainment.
• The SME is a source, not the author. Experts have forgotten what it's like not to know; your job is to translate their fluency into a beginner's path.

Mental Models

• ADDIE. Analyze, Design, Develop, Implement, Evaluate — the discipline's spine. Treat it as a checklist of what must happen, not a rigid waterfall; the phases overlap and loop.
• SAM (Successive Approximation Model). Allen's iterative alternative: small prototypes, early review, rapid cycles. Use it when requirements are fuzzy and the client learns what they want by reacting to a draft.
• Bloom's Taxonomy. Remember, understand, apply, analyze, evaluate, create — a ladder of cognitive demand. The verb in your objective sets the level; don't ask a learner to evaluate what they can't yet recall.
• Gagné's Nine Events of Instruction. Gain attention, state the objective, recall prior knowledge, present, guide, elicit performance, feedback, assess, enhance retention. A scaffold for a single lesson's arc.
• Cognitive Load Theory (Sweller). Intrinsic load (inherent difficulty), extraneous (poor design), germane (schema-building effort). Cut extraneous; manage intrinsic by chunking and sequencing.
• Mayer's Multimedia Principles. Coherence (cut seductive details), signaling, redundancy, modality, segmenting — evidence for how words and pictures combine.
• Backward Design (Wiggins & McTighe). Identify results, determine acceptable evidence, then plan the learning. Stops the activities-first drift.
• Kirkpatrick's Four Levels. Reaction, Learning, Behavior, Results. Most shops only measure Level 1; the value lives at Levels 3 and 4.

First Principles

• People learn by retrieving and applying, not by being exposed; exposure feels like learning and isn't.
• If nothing changes in long-term memory or on-the-job behavior, no learning occurred, no matter how the course felt.
• Every element on a screen either supports the objective or steals attention from it.

Questions Experts Constantly Ask

• What can the learner not do today, and where's the evidence it's a skill gap and not a tooling or motivation gap?
• What exactly will they do differently on the job, and how would I observe it?
• Who is the learner — what do they already know, and what's their context of use?
• Would a job aid or checklist beat a course here?
• What can I cut without harming the objective?
• Is this practice with feedback, or just more telling?
• How will we measure whether it transferred, not just whether they liked it?

Decision Frameworks

• Performance-gap analysis (Mager & Pipe). Could they do it if their life depended on it? If yes, it's not a training problem — fix the environment, incentives, or expectations instead.
• ADDIE vs. SAM. Waterfall ADDIE for stable, compliance, high-stakes content with fixed requirements; SAM's rapid prototyping when the client can't articulate the goal until they see something.
• Fidelity decision. Match production fidelity to the cost of getting it wrong on the job. A surgeon's simulation earns high fidelity; a policy refresher earns a slide deck. Don't spend video budget where text transfers fine.
• Build vs. curate vs. buy. If quality content already exists, curate or license it; reserve bespoke build for what's specific and high-stakes.
• Evaluation depth. Decide before launch which Kirkpatrick level you'll measure to — and resource the data collection — or you'll default to a smile sheet.

Workflow

1. Intake and analyze. Interview the requester for the real performance problem, not the requested solution. Confirm training is the fix. Profile the audience.
2. Write objectives. Turn the goal into measurable behavioral objectives (Mager: audience, behavior, condition, degree). These become the spine.
3. Design the assessment first. Decide how each objective is demonstrated; build the practice and the test before the content.
4. Extract from the SME. Pull the expert's knowledge, then prune hard to what the objectives require; capture worked examples and common mistakes.
5. Storyboard. Sequence for load: prerequisites first, chunk, scaffold, fade. Apply Gagné's events and Mayer's principles to each screen.
6. Prototype and review. Get a low-fidelity version in front of the client and a sample learner early, before expensive production.
7. Develop. Build in the authoring tool; integrate media; configure the LMS and tracking.
8. Pilot. Run with a real slice of the audience; watch where they stall; gather data.
9. Evaluate and iterate. Measure to the agreed Kirkpatrick level; revise; report results to the sponsor in their terms — behavior, business outcome.

Common Tradeoffs

• Fidelity vs. cost. High-fidelity simulation transfers best and costs the most; match fidelity to stakes, not budget envy.
• Engagement vs. learning. Gamification and animation raise enjoyment but can add extraneous load; keep only what serves retrieval.
• Just-in-time vs. comprehensive. A short job aid at the moment of need often beats a thorough course nobody remembers; reserve full courses for foundational skills.
• Speed to market vs. instructional quality. Rapid e-learning ships fast and often teaches little; name what you're trading.
• Standardization vs. personalization. One linear path scales cheaply; branching and adaptive paths serve learners better at higher build cost.
• SME's urge to cover everything vs. the learner's capacity. The expert wants completeness; the learner needs sufficiency. You arbitrate.

Rules of Thumb

• If the objective's verb is "understand" or "know," rewrite it — those aren't observable.
• Show, then have them do; never just tell.
• Cut every word and image that doesn't earn its place against the objective.
• If a checklist would solve it, don't build a course.
• Don't narrate on-screen text word for word — that's the redundancy trap.
• Chunk to roughly 5-9 items; working memory can't hold more.
• Pilot with one real novice before you build for a thousand.
• A smile sheet measures catering, not learning.

Failure Modes

• The content dump. Recording the SME's slides and calling it a course; coverage with no practice and no objective.
• Click-next e-learning. Passive page-turners where the only interaction is advancing; high completion, zero transfer.
• Objectives written after the content. Retrofitting goals to justify what was built.
• Measuring only Level 1. Declaring success on satisfaction scores while behavior never changed.
• Seductive details. Adding entertaining but irrelevant stories, images, or music that raise interest and depress learning (Mayer's coherence violation).
• Building training for a non-training problem. A polished course aimed at a gap that was really about broken tools or incentives.

Anti-patterns

• "Spray and pray" — push everyone through the same course and hope it sticks.
• The 47-objective module — so many goals that none is actually taught.
• Read-aloud slides — narration duplicating on-screen text verbatim.
• Quiz-as-afterthought — assessment bolted on at the end, untied to objectives.
• Mandatory annual everything — compliance theater that trains learned helplessness.
• SME-as-author — shipping the expert's mental model untranslated for novices.
• Tool-first design — choosing the flashy authoring feature, then finding content to fill it.

Vocabulary

• Learning objective — an observable, measurable statement of what the learner will do, under what conditions, to what standard.
• Performance gap — the difference between current and required on-the-job behavior.
• Storyboard — the screen-by-screen blueprint of content, media, and interaction.
• Cognitive load — the demand on working memory; intrinsic, extraneous, germane.
• Transfer — applying what was learned to the actual job; the real target.
• Smile sheet — an end-of-course satisfaction survey (Kirkpatrick Level 1).
• SME — subject-matter expert; the knowledge source, not the designer.
• Rapid e-learning — fast, template-driven course production, often low-interaction.
• Microlearning — short, single-objective units delivered at the point of need.
• Knowledge check — a low-stakes formative question used to prompt retrieval.

Tools

• Authoring tools — Articulate Storyline/Rise, Adobe Captivate, Lectora.
• The LMS — Moodle, Cornerstone, Docebo, SCORM/xAPI for delivery and tracking.
• Storyboarding — document templates, Twine for branching scenarios.
• Media production — Camtasia, Vyond, Audacity, Canva.
• Analysis frameworks — the objective template (Mager), the ADDIE/SAM map.
• Evaluation instruments — Kirkpatrick survey banks, observation rubrics, LMS analytics.

Collaboration

Instructional designers rarely build alone. They extract from subject-matter experts (who own truth but not teachability), hand off to graphic designers, video producers, and developers (who own craft and production), coordinate with LMS administrators (who own delivery and data), and answer to a sponsor (who owns the performance problem and the budget). The recurring friction lives at the SME handoff — the expert wants everything in, the designer fights for the learner — and at the sponsor handoff, where "make a training" hides an unexamined performance problem. The best designers manage SMEs like a journalist manages a source: extract, verify, prune, and protect the audience from the curse of knowledge.

Ethics

Instructional designers shape what people believe they know how to do, often in high-stakes settings — safety, compliance, medicine, finance. Duties: don't certify competence the assessment didn't actually verify; don't build training as cover for an organization's failure to fix a real problem (the "we trained them" liability shield); respect learners' time by cutting everything that doesn't serve them; design for accessibility (WCAG, captions, screen-reader paths) so no learner is excluded; cite and license sources honestly; and tell the sponsor the truth when the answer is "this isn't a training problem," even when they're paying you to build training. The gray zone — compliance courses that exist to manage legal risk rather than change behavior — deserves to be named rather than quietly produced.

Scenarios

"We need a course on the new expense system." The requester wants a 60-minute module by month-end. The novice starts storyboarding. The expert runs a five-minute analysis: who's failing at expenses, and why? The system isn't confusing — people just can't find the approval button, which moved. This isn't a skill gap; it's a UI discoverability gap with a two-minute fix: a tooltip and a one-page job aid showing the new location. The designer recommends killing the course and ships the job aid. The hardest part is convincing a sponsor who already decided they wanted training.

A compliance module with a 100% pass rate and no behavior change. Annual data-security training shows perfect completion and high satisfaction, yet phishing-click rates haven't moved. Level 1 looks great; Level 3 is broken. The expert sees the design flaw: the course tells learners about phishing but never makes them do the discrimination task. The redesign replaces the lecture with a scenario simulation — learners triage real-looking inbox messages, get immediate feedback, and must hit a threshold on novel examples. The assessment now mirrors the job, and evaluation shifts to a tracked metric: simulated-phishing click rate at 90 days. Learning is measured where it matters, not where it's easy.

Fuzzy request, impatient SME. A startup wants onboarding training but can't say what "good" looks like; the SME is a brilliant engineer with no time who explains everything at once. The expert abandons waterfall ADDIE for SAM: builds a rough one-screen prototype of the first task in a day, puts it in front of the SME and one new hire, and lets their reactions surface the real objectives. Three rapid cycles later the scope is clear, the SME has corrected the worst misconceptions by reacting rather than authoring, and nothing expensive was built before the target was known.

Related Occupations

An instructional designer shares the learning focus of educators but works without a live room, designing the experience rather than delivering it. Teachers teach the designed thing in real time and read a class the designer never meets. UX designers share the obsession with reducing cognitive load and designing for a user's context — an e-learning screen is a UX problem. UX researchers supply the audience-analysis muscle the designer relies on. Technical writers solve the adjacent problem of documenting how something works, often the better answer when the gap is reference, not skill.

References

• Design for How People Learn — Julie Dirksen
• Analyzing Performance Problems — Mager & Pipe
• Multimedia Learning — Richard Mayer
• The Conditions of Learning — Robert Gagné
• Evaluating Training Programs — Donald Kirkpatrick
• Leaving ADDIE for SAM — Michael Allen