Turning a document into lessons without rewriting it

Part of a single argument: the engagement gap is a workload problem.

Everyone agrees a fifty-page policy should not be handed to a new starter as training. There is less agreement on what to do instead, and the usual answer, rewrite it as a course, is where the project stalls.

The conversion is more tractable than it looks, provided you separate the mechanical parts from the editorial ones and stop treating it as one undifferentiated act of writing.

Split before you write

The first move is not writing. It is cutting the source into pieces that each contain one idea.

This is mostly mechanical. Sections, headings and paragraph groupings in the original usually map onto candidate lessons, and where they do not, it is generally because the source combined two things for filing reasons rather than for comprehension reasons.

The output is not a course yet. It is a set of pieces, and having them in front of you removes the blank page from every subsequent step.

Order for the learner, not for the rules

Source documents are ordered by the logic of the subject. Courses need to be ordered by the state of the learner, and those are rarely the same sequence.

The practical version of this: put first the thing that lets someone do their job on Monday, even if it is logically a special case that the document covers in section nine. Put the general principle after it, when there is something concrete for it to attach to.

This reordering is the single largest improvement in most conversions, and it costs nothing but the decision.

Cut what protects rather than teaches

Parts of most source documents exist to establish that something was communicated. That has real legal value and no learning value.

Keep it in the document of record. Keep it out of the lessons, and have people acknowledge the document separately. Mixing the two is why compliance training reads the way it does, and learners can tell when they are being processed rather than taught.

Add the thing the document cannot have

A document states what is true. A lesson has to deal with what learners believe before they read it, which is usually something slightly wrong.

This is the part that cannot be extracted from the source, because it was never in the source. It comes from whoever has taught the material and watched people get it wrong in the same place every time. One sentence naming the common misconception is often worth more than a page of correct explanation.

If your conversion process has no step where someone with teaching experience adds this, you have produced a reformatted document rather than a course.

Check for understanding, not for reading

Every lesson needs at least one question whose answer would differ between someone who understood and someone who skimmed.

The test for a good check is uncomfortable: if the pass rate is above ninety-five per cent, it is decoration. Somewhere between seventy and eighty-five means it is doing work. Below sixty and the lesson, not the learner, is usually the problem.

Most checks written during a conversion fail this test, because they are written by someone who has just read the source and is unconsciously testing recall of phrasing.

What the whole loop looks like

Split the source. Reorder for the learner. Cut the protective material. Add the misconceptions. Write checks that discriminate.

Two of those five are mechanical, one is a judgement call that takes minutes, and two require someone who has taught the subject. That is a much smaller ask than rewrite it as a course, and it is a description that a team can actually staff.