Case study · Curriculum and course development
Hundreds of thousands of people do this job. There was no textbook, no accreditation precedent and no competitor syllabus to borrow from. Here is how a decade of practitioner experience became ten modules a college could adopt, staff and stand behind.
The subject was a job that hundreds of thousands of people do and for which no formal training existed anywhere. There was no textbook, no accreditation precedent, no competitor syllabus to borrow from, and no agreed body of knowledge. There was, on the other hand, eleven years and nine thousand hours of direct experience doing the work.
That combination — deep practitioner knowledge, zero existing structure — is the hardest kind of curriculum to build and the most valuable when it works.
The sequence, and where everyone gets it wrong
Backward design is standard instructional practice and routinely ignored. The discipline is refusing to write a single slide until the outcomes and the assessment exist.
Every module has to answer one question: what can the student do at the end of this hour that they could not do at the start? If the answer is “know more about it,” the module is not finished. That single rule removed more material than it added.
The course was built with heavy AI assistance and it would not have been finished otherwise. But the division of labor matters, and overstating it is how bad courses get made.
The honest split
AI compressed the production work by a large factor. It did not and could not supply the practitioner judgment, which is the only reason the course is worth taking.
An early draft contained illustrative stories that read well and were not verifiable. They were cut. A workforce course that carries an institution’s name cannot contain a single anecdote that will not survive a student asking where it came from. Every claim in the final version is either the author’s own experience or externally sourced and linked.
A good course and an approvable course are not the same document. The second one is mostly paperwork, and it is the part that stalls projects.
| Deliverable | Who reads it | What it has to survive |
|---|---|---|
| Learning outcomes, per module and overall | Academic affairs | Mapping to the institution’s own outcome framework |
| Assessment structure and weighting | Academic affairs | “How do you know they learned it?” |
| Student handbook | Students and administration | Grading, academic integrity, appeals, expectations |
| Disclaimers on earnings and outcomes | Legal and administration | Anything that could read as an income guarantee |
| Instructor materials | Whoever teaches it next | Someone other than the author running it cold |
| Delivery and platform plan | Continuing education | Asynchronous online at their standard |
Instructor materials. A course that only its author can teach is not a course, it is a performance — and an institution will not adopt it, because adoption means owning something it cannot staff.
This is the same discipline as everything else on this site. Define the outcome, work out how you would measure it, then build toward that — and refuse to produce content that has nothing to be measured against.
| The usual approach | What was done here |
|---|---|
| Start writing content | Start with what the student can do afterward |
| Use AI to produce volume | Use AI for production, a person for judgment, and say which is which |
| Good stories | Only verifiable ones |
| A syllabus | A package an institution can actually adopt and staff |
Onboarding, certification, internal training and documentation are all the same problem. Most corporate training fails for the same reason most courses do: it was written as content first, with nobody able to say afterward whether anyone learned anything.
Marketing Analytics Consultants
We build the outcomes, the assessment and the package an institution or a team can actually adopt — and we are honest about which parts a machine drafted and which parts required someone who has done the job.