The internal DevOps course at a mid-sized software company had launched to applause. Within two months, the weekly session was half empty. The training coordinator, who had built the programme in spare hours, opened the attendance sheet on a Monday and decided that this was the kind of problem worth treating properly. What follows is an invented case study, assembled from patterns common in corporate training, about how a team moved from “people are not interested” to a cause it could repair by asking why five times.
The first answer was comfortable and wrong
In the post-session review, the first explanation offered was motivation. Engineers were busy, the thinking went, and a voluntary course would always lose to deadlines. The coordinator wrote it down, as the process required, and then asked why attendance had been healthy for the first three weeks and fallen sharply in the fourth.
Motivation does not change that abruptly. Something else had shifted. That mismatch between the explanation and the timeline is the signal a careful reviewer looks for: when an answer cannot account for the shape of the data, it is probably a symptom.
Why one: who stopped coming?
The sheet showed that absences clustered in two squads, the platform team and the payments team, while others kept turning up. So the second question became narrower: why those two squads? The coordinator talked to a lead from each. Both said the session began at the same time as their daily handover between the morning and afternoon on-call engineers.
Why two: why did the session overlap with handover?
The slot had been chosen in week one, when the platform and payments squads shared a rotation with a different schedule. In week four, an unrelated change moved their handover half an hour earlier. Nobody told the training team, because nobody thought of the training team as affected by on-call timing.
Why three: why did nobody tell the training team?
The company had a channel for operational schedule changes, but the course was not listed as a subscriber. It was an informal programme without a formal owner in the engineering organisation, so it had no seat in the places where changes were announced.
Why four: why was the course outside those channels?
When the course was approved, the approval email described its content and budget. It did not describe its dependencies: that it needed a time window free of on-call duties for the people it was meant to serve. The template for approving internal programmes had no field for dependencies.
Why five: why does the template lack that field?
The template had been copied from a purchasing form designed for software licences, which depend on nothing but money. No one had asked what a programme involving people’s time needed that a licence did not.
Five answers in, the story had travelled from “engineers lack motivation” to “our approval template for people-centred programmes was borrowed from a form for purchases”. That finding was a relief to everyone, because it blamed nobody and could be fixed in a week.
What the team actually changed
- The coordinator subscribed the course to the schedule-change channel, so future shifts would be visible.
- The approval template gained a “depends on” field with examples such as rotas, shared rooms and tool access.
- The session moved to a time that cleared both squads’ handovers, and a short repeat recording went to anyone who still could not attend.
- The review document gained a final line: “What would have told us earlier?” In practice it proved the most valuable prompt on the page.
Attendance recovered over the next month, though not to the first-week peak, which the coordinator wisely treated as novelty rather than a target. More importantly, when a third squad changed its working pattern the following quarter, the channel subscription flagged it before the session suffered.
How the review sessions were run
The coordinator kept the format deliberately light. Four people met for thirty minutes with a shared document on a screen. One person asked the questions, another typed the answers verbatim, and a third was assigned to challenge any answer that sounded like an opinion rather than an observation. When the group first said “the squads are too busy”, the challenger asked what evidence supported it, and the room went quiet until someone pulled up the handover schedule. That small interruption changed the direction of the whole review.
Afterwards the typed chain was shared with the two affected squad leads before it went anywhere else. Seeing their own constraint described accurately, with no hint of fault, made them far more willing to help choose the new time. Their suggestions on timing were better than anything the training team would have guessed, which is a recurring bonus of sharing a chain early: the people closest to the problem usually hold the missing detail.
Why the story is useful beyond the course
It shows the pattern most chains share. The visible symptom, empty seats, sat at the end of a long path whose earliest link was a document. By following the path the team changed the document and, with it, the odds of the same problem recurring in an entirely different programme.
It also illustrates a quiet advantage of the method: it produces fixes proportionate to the cause. A motivation problem would have called for incentives, reminders or management pressure. A template problem called for a single field. Spending effort on the wrong diagnosis is expensive, and the cost is usually hidden until much later.
A glance at how other sites organise material
While redesigning the course portal, the coordinator looked at unrelated sites for navigation ideas. One was a directory-style orthopaedic site with a keyword and country search plus separate areas for articles and community posts, which showed how a visitor can reach a destination from several directions. No claims were taken from it; the lesson was only that search and sections can sit side by side.
Two internal pieces gave the coordinator further context: an essay on the outlook for DevOps both in Japan and around the world, and a practical guide to managing cultural change alongside technical change. Both reinforced a point the case study makes by example: tools rarely fail first; arrangements around them do.
Takeaways for your own programme
If your training, mentoring or study group has seen attendance fall, resist the first comfortable story. Check whether the drop is abrupt or gradual, ask which groups are affected, and then ask why, repeatedly, until you reach a document, a channel or a default you can edit. Record the chain, make the change, and set a date to see whether the seats fill again. The method costs an hour, and it often saves a programme.









