Most succession planning training teaches the framework and skips the part that actually breaks. Managers leave able to describe a 9-box grid and still unable to place one of their own people defensibly, because the hard skill was never the model. It was writing down evidence, separating potential from polish, and holding a rating when someone senior pushes back.
This is a guide to training the people who make the calls. It assumes you already have a process; what you need is for the managers inside it to be competent at the four or five things the process quietly depends on.
What succession planning training actually needs to cover
There is a standard curriculum that gets taught, and there is the curriculum that changes outcomes. The standard one covers definitions, the nine-box, readiness windows, and the annual calendar. It is not wrong, it is just insufficient, because none of it is what managers get stuck on.
What they get stuck on is narrower and more uncomfortable. How do you tell the difference between a strong performer and a person ready for a bigger job? What counts as evidence for potential when the only thing you have observed is competence in the current role? What do you say when a peer disagrees with your placement in front of the leadership team? How do you have a development conversation with someone you have just rated as not ready?
Training that covers the model but not those four questions produces managers who can participate in the meeting without improving the decisions made in it.
Train the raters, not the audience
The most common design error is training everyone equally. Succession planning touches the whole organization, so the training gets broadcast to the whole organization, and it becomes an awareness session.
The people whose skill actually determines quality are the ones who assign ratings: the managers of the people being assessed, plus whoever facilitates calibration. That group is usually small enough to train properly, in a room, with real practice. Everyone else needs something much lighter, closer to a communication than a course.
Splitting the audience this way costs less and works better. A ninety minute working session with fifteen managers who will each rate five people beats a company wide webinar by a distance that is not close.
The four things a manager has to be able to do
Scope the training around capabilities rather than topics. A manager is trained when they can do these four things without help.
Write an evidence statement. Not "strong leader" but "ran the recall response over three weeks, kept the team shipping, escalated the two decisions that genuinely needed escalating." The test is whether someone who was not there can tell what happened. This is the single most useful thing to practice, and most managers have never been asked to do it in writing.
Separate performance from potential. Performance is about the current role. Potential is about capacity for a bigger one. They correlate loosely and get conflated constantly, which is how the best individual contributor ends up rated as the strongest successor for a job requiring skills nobody has ever seen them use.
Name a gap without softening it. A useful assessment says what is missing. Managers routinely avoid this because it feels like a judgment on the person, so ratings drift upward and every candidate is two years out forever. Practicing the sentence out loud helps more than explaining why it matters.
Hold or change a rating under challenge, on the evidence. Both outcomes are fine. What is not fine is changing a rating because someone senior disagreed, or holding one out of stubbornness. The skill is knowing which is happening.
Run it as calibration practice, not a course
The format that works is a rehearsal. Take three or four anonymized profiles, have every manager place them independently and in writing, then compare the placements in the room.
The disagreements are the lesson. When two managers place the same profile in different boxes, the exercise is not to determine who was right but to surface the different standards they were applying, then agree which standard the organization is going to use. That conversation is the actual content, and it cannot be delivered as slides.
Run this once before the first real review, then again as a fifteen minute warm up at the start of each calibration meeting. Standards drift between cycles, and rereading the criteria out loud before placing anyone is boring and highly effective.
How to tell whether the training worked
Satisfaction scores from the session will tell you nothing useful. Three signals from the next review cycle will.
Look at whether evidence statements are specific enough that a reader who was not present can follow them. Look at how many placements changed during calibration, and whether they changed because of evidence or because of who was in the room. Look at whether any candidate had a readiness window advance without a completed development action behind it, which is the clearest sign that a rating moved on sentiment.
If those three look healthy, the training took. If ratings are still adjectives and readiness still drifts on its own, more training on the framework will not fix it, because the framework was never the problem.
Where this fits
Training is one input into a succession process, and it decays like everything else. Managers change roles, standards drift, and the discipline that felt obvious in the session is gone three cycles later. Whatever you teach, the parts that survive are the ones the process requires in writing: an evidence field that cannot be skipped, a reason attached to every rating change, a development action with an owner and a date.
That is the honest limit of training. It raises the ceiling on quality; it does not hold the floor. The floor is held by what the system makes people write down.