Before I design anything, I diagnose the real performance gap — so we don’t spend a budget building training for a problem training can’t fix. Sometimes the most valuable thing an instructional designer can say is “this isn’t a course.”
Requests almost never arrive as a performance gap — they arrive as a solution someone already picked: “we need a training.” Build it reflexively and you spend weeks and a budget on something that doesn’t move the number, because the real cause was never a lack of skill. A good performer in a broken system loses to the system every time. So the first thing I do is not design. It’s diagnose.
The request, as it arrives: “Our reps aren’t following the new returns process. We need a training.”
What is the gap actually costing — in dollars, risk, or time? If the problem isn’t worth solving, no intervention is. Here, reps skipping the new steps drove reprocessing costs and refund leakage — clearly worth fixing.
Could they do it if their job depended on it? If yes, it’s not a skill gap — training won’t help. Here, most reps genuinely didn’t know the new steps: a real can’t-do.
Before blaming the person, I check the system: is the information there, the tools, the incentives? The new process existed — but the job aid was buried and the old button still worked.
A blended fix: a short skill-build for the genuinely-new judgment calls, plus two non-training fixes — surface the job aid, retire the old button. Training alone would have failed.
Gilbert’s Behavior Engineering Model. Most performance gaps live in the top row — the environment, which is management’s to fix. Training only addresses the bottom-left cell. Naming the real cell is what keeps a course from being built to paper over a system problem.
Clear expectations, guides, and timely feedback. Often the real culprit.
The tools, time, and process to actually do the work.
Consequences that reward the right behavior — or quietly punish it.
The one cell training actually addresses.
The physical or cognitive ability to perform.
Whether the person wants to, given everything above.
Recommending the non-training fix is where an instructional designer earns trust with the business. It says the goal is the performance, not the deliverable.
I use AI to run the first pass of the stakeholder intake — it drafts the diagnostic questions, captures the answers, and produces a first-pass gap statement in plain language. That gets me to a structured analysis in minutes instead of a week of back-and-forth. Then I do the part that matters: deciding which cell the gap actually lives in, and having the honest conversation when the answer isn’t a course. The AI accelerates the intake; the diagnosis is mine — and the method holds whatever intake or authoring tools a team already uses.
Diagnosing before designing protects the budget, targets the intervention at the actual cause, and builds credibility with stakeholders who’ve been burned by training that didn’t stick. When it is a course, it’s a course built on a gap I’ve confirmed is real — which is exactly what makes the design downstream defensible.
Frameworks in play: ADDIE (Analyze), Mager & Pipe’s performance analysis, Gilbert’s Behavior Engineering Model, and root-cause analysis. The design that follows a confirmed gap is covered in From a sentence to a full course →