Part 3 of the six-part series From Course Factory to Capability Function.
Previously in this series: Before You Automate L&D, Decide What Should Disappear
For a step-by-step diagnostic, see How to diagnose a performance problem before commissioning training.
When a business unit has a performance problem, the request that reaches L&D usually sounds like this: “We need training on X.” Most teams accept the framing and start building. That is where the waste begins.
Training solves one kind of problem: people cannot do something they need to do, and the reason is a gap in skill or knowledge. Many performance problems are not that kind. They are process problems, incentive problems, tool problems, or clarity problems. No course fixes those. A team that cannot follow a broken process will not follow it better after a workshop. A team paid to do the wrong thing will keep doing the wrong thing, politely, after training.
The discipline is to diagnose before you prescribe. For the next request that lands on your desk, answer four questions in writing before you agree to build anything.
First, what does good performance look like, in observable terms? Not “better communication.” Something you could watch: “Account managers confirm the scope in writing before work starts.”
Second, where is the gap? Who is not performing at that standard, and where exactly does their work break down? If you cannot name the gap precisely, you do not have a training need. You have a vague complaint.
Third, what causes the gap? This is the question most teams skip. Ask whether the cause is skill. Could a capable, motivated person do this correctly with the current process, tools, incentives, and information? If yes, you may have a skill problem. If no, you have a different problem, and training is the wrong tool. Fix the process, the incentive, the tool, or the expectation first.
Fourth, what evidence would show the problem is solved? Name it before you start. If the only evidence available is completion rates, you are measuring your own activity, not the business result.
Run a request through those four questions and three outcomes are possible. Sometimes you confirm a genuine skill gap, and you build training with a clear target. Sometimes you find a non-training cause, and you send the request back with a recommendation: fix the process, change the incentive, clarify the expectation. Sometimes you find that nobody can describe good performance at all, which means the organization is asking L&D to solve a problem it has not defined. That is useful information too.
Most L&D teams will find that a large share of requests fail at question three. That is not a reason to feel threatened. It is the most valuable thing L&D can do: stop the organization from spending money on training that cannot work, and redirect the effort to the cause that will.
A capability function earns its place here. Anyone can take an order for a course. It takes a diagnostician to tell a senior leader that training is the wrong answer, and to show them what the right answer is. That conversation is uncomfortable. It is also the job.
Questions this article answers
When is training the wrong response to a performance gap?
When the constraint lies in an unclear expectation, a broken process, poor tools or incentives, training will not remove it. Test the cause before commissioning a course. The AI training triage illustrates the diagnostic question.
What should L&D ask before accepting a course request?
Define observable good performance, find the actual gap, examine its cause, and agree what evidence would show improvement. If those answers are missing, start with diagnosis. See the Solve the Right Problem Diagnostic.
Next step
If this problem is not a training problem, start with diagnosis before you fund a solution. Use Who Should Do the Work? to decide who or what should do each part of the work, or use the consultation form to check whether your next investment is aimed at the right problem.
