The Need-to-Know principle:
less content, more learning

There is a particular kind of pride that subject matter experts take in the completeness of their knowledge — and it is a quality that, when channelled into learning design, causes more damage than almost anything else. I have sat in more discovery workshops than I can count where a brilliant technical expert, asked what they want learners to know, answers with a version of the same thing: "Everything. They need to understand all of it."

And I understand the impulse. When you have spent years developing deep expertise, the idea of teaching only part of it feels like a betrayal of the subject. But here is what that impulse produces when it drives curriculum design: a forty-five-minute eLearning module that a learner cannot get through without losing the will to continue. A programme that covers everything and teaches nothing. A completion certificate that proves endurance, not competence.

Every learner arrives with a limited cognitive budget

Cognitive load theory — first articulated by psychologist John Sweller — tells us that working memory has a finite capacity. We can hold roughly four chunks of information in active processing at any one time. When a learning experience exceeds that capacity, something has to give. And what gives is understanding. The learner may continue clicking, but they stop absorbing.

This is not a failure of the learner. It is a failure of the design. When we present too much information, we are not giving learners more — we are giving them less. We are diluting the signal with noise until nothing gets through clearly.

The Need-to-Know principle is the antidote. It starts with a deceptively simple question: what is the minimum a learner needs to know in order to perform their role competently and safely? Not everything that would be useful. Not everything the SME finds fascinating. The minimum that makes the difference between capable and not capable.

The airport test — a practical filter

Here is a heuristic I use when auditing content: the airport test. Imagine your learner is sitting in an airport, thirty minutes before boarding, with their phone in their hand. They encounter a situation in their work and need to recall what the training told them. What do they remember? What can they act on?

The answer is almost never a comprehensive policy document or a forty-point checklist. It is one or two anchoring concepts — a mental model, a decision rule, a vivid scenario — that give them enough to act correctly in the moment. The training that lasts is the training that was designed with that moment in mind, not with a coverage requirement.

This is why role-specific design is almost always superior to generic training. A medical device technician does not need to know everything about regulatory compliance — they need to know what their role requires them to do when an inspector arrives. A customer service representative does not need a complete history of data privacy legislation — they need to know the three things they must never do with a customer's personal data, and why. Precision of scope is a form of respect for the learner's time and cognitive capacity.

How to apply the principle in practice

The first step is to start with performance, not content. Before opening a single source document, ask: what does good performance look like in this role? What decisions does this person make, and what information do they need to make them well? This reframes the design question from "what should I cover?" to "what does this person need to be able to do?"

The second step is to conduct what I call a ruthless content audit. Take everything the SME has given you and apply a single test to each item: if the learner does not know this, will their performance suffer? If the answer is no — if the information is context, background, history, or enrichment rather than essential knowledge — it does not belong in the core programme. It might belong in a supplementary resource, a job aid, or a reference document. But it should not compete for working memory with the content that actually matters.

The third step is to design for retrieval, not coverage. The goal is not to expose learners to information — it is to ensure they can access it when they need it. Retrieval practice, spaced repetition, and application activities all serve this goal far better than additional content ever can. A learner who has practised recalling three key concepts ten times will outperform a learner who has been exposed to thirty concepts once.

The hardest part — the conversation with the SME

None of this is easy, because it requires saying no — or at least "not here" — to people who know more about the subject than you do. That conversation requires confidence in your role. You are not the expert on the content. You are the expert on how people learn, and on what learning design serves the performance goal. Those are different domains, and they carry equal weight in the design process.

The best SME conversations I have had started with a shared agreement on one thing: the purpose of this programme is not to teach everything we know. It is to close a specific performance gap. Once that is agreed, the question of what to include becomes a technical discussion rather than a territorial one.