/ Articles

Most training content is built from the source document outward. A procedure gets turned into slides, the slides get a quiz at the end, and the course ships. Training content development done that way covers everything in the document, at equal weight, for people who may already know most of it.
The alternative starts at the other end. Find out what your people do not know yet, build content for that, cut what they already have, and then check whether the new material moved the number. For an EHS manager, a nurse educator or a chief of training, that order decides where the next round of training hours goes.
Table of Contents
A large review of the training literature in Psychological Science in the Public Interest lays out the order. The journal's summary of the review says organizations "should take steps before training to create an appropriate learning climate and conduct a Training Needs Analysis." It defines that analysis as "a diagnosis of what needs to be trained, to whom, and in what organizational system" (Salas et al., 2012).
That diagnosis is its own job, and a training needs assessment covers how to run one: the levels of analysis, the questions to ask and a template for the results. This article picks up where the diagnosis ends. You have a list of knowledge areas, each with a measured result for the people who need it. Now you have to decide what content to build.
The content brief is the gap list. The procedure manual tells you what is true. The gap list tells you which of those truths your people cannot currently produce when asked.
Gap data is only useful if it changes what you build. A simple way to use it is to sort every knowledge area into one of four content decisions. The table below is our own working summary, not a published model.
| Measured result | What it usually means | Content decision |
|---|---|---|
| Most people miss it on the first try | A real knowledge gap | Build new or rewritten content |
| One team or role misses it, others do not | A local gap | Build for that group only |
| Most people answer it correctly | Knowledge already in place | Cut or shorten; keep a light check |
| People know it but still skip the step on the job | Probably not a knowledge problem | No new content; look at tools, staffing or the procedure |
The last row matters more than it looks. Some failures are not training problems at all, and new content will not fix a missing tool or an unclear procedure.
Expect the gaps to show up in places you assumed were covered. In a study of 37 F-16 and C-17 pilots, described in a conference paper prepared for I/ITSEC 2026 (Smith & Randolph, Paper No. 26299), fully qualified pilots averaged 66.9% accuracy the first time they answered each operational knowledge question. Qualification and current knowledge are different things. Watch also for items only one veteran can answer, the expertise that leaves when experienced people go.
Cutting content is the step most teams skip, because repeating the full course feels safer. But repeating material people already know has a cost beyond the hours.
Instructional research has a name for one version of that cost: the expertise reversal effect, described in a 2003 review in Educational Psychologist. Its core finding: "Instructional techniques that are highly effective with inexperienced learners can lose their effectiveness and even have negative consequences when used with more experienced learners" (Kalyuga et al., 2003). The authors add that "strong evidence has emerged that the effectiveness of these techniques depends very much on levels of learner expertise."
That work looked at instructional techniques in learning experiments, not at annual refreshers. Applying it to a refresher is our inference. A step-by-step walkthrough built for new hires may do little for a crew with ten years on the line, and it takes time from the step they actually miss.
In practice, cutting looks like this:
Once you know what to build, the format matters. A job-critical fact is only useful if someone can pull it from memory on shift, without a binder in hand.
A review of learning techniques for students in Psychological Science in the Public Interest rated two approaches highest. Taking practice tests and spreading study out over time were rated high utility "because they benefited students of many different ages and ability levels" (Dunlosky et al., 2013). Commonly used techniques such as rereading and underlining "were found to be of surprisingly low utility."
Those findings come from studies of students, so treat the following as our suggestions for workplace content built on them:
Two terms come up below. Spaced repetition means reviewing material at intervals over time. Retrieval practice means answering from memory instead of rereading. Spaced retrieval is the two combined.
Packaging these items into a full microlearning course (one objective per item, sequenced and spaced) is a separate design job with its own decisions. The gap data still sets which items go in.
Content development does not end at launch. The journal's summary of the same Salas et al. (2012) review says that after training, employers "should evaluate the training methods to see if any improvements could be made." The Kirkpatrick Model has four levels of evaluation: Reaction, Learning, Behavior (which Blank Slate's write-up calls Impact) and Results. For content decisions, Level 2 (Learning) is the one to watch first.
A practical check uses the same questions that found the gap:
A post-course quiz with a high pass rate is a weak signal. It measures recall minutes after exposure, not what people can produce on shift weeks later.
The content you build will not hold on its own. Without reinforcement, memory fades, so the gap you closed in March can reopen by summer. That is where content development hands off to reinforcement, and where Blank Slate fits.
Blank Slate is a cognitive-readiness platform built on AI-optimized spaced retrieval. It works on top of the training content you already have. Employees answer short questions in brief sessions (about 60 seconds), 3-4 times a week, adding up to a few minutes a week.
Each answer is a data point, and the platform's True Knowledge Level™ shows what a person or team knows now.
The evidence starts with the pilots above, whose first-attempt accuracy rose from 66.9% to 81.4% over the 16-month study, a 21.6% relative improvement (Smith & Randolph, I/ITSEC 2026). In a year-long case study of 70 data-center security guards presented at MODSIM World 2025, guards who started below the team median went from about 56% to over 90% accuracy (Smith, MODSIM World 2025). They used three short sessions a week, six questions each.
Two studies with students speak to the cost side. In a randomized controlled trial, Blank Slate's spaced retrieval matched daily review's retention in about one-third of the study time (McHugh et al., 2021). In a small proof-of-concept study, students learned one topic through Blank Slate alone, with no lecture. On the final test, they did as well on it as on a lectured topic (Gordon, Hughes & Smith, 2025).
Blank Slate does not deliver required training hours, a refresher course or hands-on practice. It reinforces and measures retention between them, and it coexists with your LMS: the LMS records what people completed, and Blank Slate shows what they still know (see the difference). The how it works page shows the setup, and our piece on interactive training covers why engaging content still needs reinforcement afterward.
If you do not yet have gap data to build from, a Knowledge Gap Audit is a free, short assessment taken in the browser that returns a one-page readiness snapshot.
Training content development is the work of turning what people need to know into material they can learn and recall: courses, modules, job aids and check questions. Starting from measured knowledge gaps keeps the effort on what people are missing, instead of re-covering what they already know.
Measure first. Topics most people answer correctly on the first try are candidates to shorten into a few check questions, while required topics stay. Keep measuring the shortened topics; if accuracy drops, bring the content back.
Re-ask the questions that found the gap, both right after delivery and again weeks later. If first-attempt accuracy on those items rises and holds, the content worked. Items people still miss need rewriting or a closer look at whether the problem is knowledge at all.
Usually, yes. Procedures, protocols and existing courses already hold the facts; the work is choosing which parts to target and turning them into questions people answer from memory. Blank Slate works on top of the training content an organization already has.
Before you build the next course, find out what your people don't know yet. A Knowledge Gap Audit is free: a short assessment in the browser that returns a one-page readiness snapshot, so new content goes where the gaps are.