Free · the guide

Completion isn't capability.

Your LMS can tell you who finished the training. It cannot tell you whether they can do what the training was supposed to teach.

This guide explains why those two things became confused, what completion actually tells you, what it cannot tell you, and what is worth checking before you go looking for outcomes.

Why completion won

Not because anyone believed it measured learning. Because it was the only thing the system could count.

Completion won because it was standardized, portable and easy to report. Capability wasn't.

An LMS is a delivery and record-keeping system, and it became very good at answering operational questions. Who enrolled? Who finished? How long did they spend? What score did the package report?

Those are useful questions. They are not the same question as: can this person now do the thing?

Activity is observable. Capability has to be demonstrated.

What completion hides

That the assessment may not distinguish anything. In each of the runs we have published against courses we did not author, a learner could obtain a valid completion record without the capability being demonstrated: in one because every answer key sat in the page source, in the others because a pass mark somebody set deliberately was not what the package actually reported. The runs are published in full.

That the content may never have required the capability. A course can cover every topic on the syllabus at the level of explain while the job requires decide. Completion cannot see the difference. Neither can a satisfaction survey.

That some learners may not have been able to use it. A fixed-width course on a phone, a keyboard trap, an assessment that tests reading speed rather than knowledge. The LMS can tell you whether someone progressed through the course. It cannot tell you the experience was effectively unusable for them.

So what does measure capability?

Demonstration.

• If the capability is identify, ask the learner to identify.
• If it is decide, give them a decision.
• If it is perform, have them perform.
• If it is diagnose, give them something to diagnose.
• If it is handle a difficult conversation, put them in one.

And where the capability matters beyond the training, follow it into the work: observation, quality data, error rates, performance evidence, supervisor assessment, or whatever real signal actually corresponds to the capability.

Completion tells you the training happened. Assessment tells you what the learner can demonstrate. Performance evidence tells you whether it transferred.

Three levels of evidence

Your LMS is very good at Level 1. Most training barely reaches Level 2. Level 3 requires evidence from the work itself.

Learning Preflight™ sits before all three. It asks whether the training is structurally capable of producing meaningful evidence in the first place.

Six things worth checking before you measure outcomes

1. Objective alignment. Does every stated objective have instruction you can point to?
2. Capability coverage at the required level. Where the objective is apply, does the course go beyond explaining?
3. Practice adequacy. Does the learner do the thing, with feedback that explains why?
4. Assessment alignment. Does the completion record depend on demonstrating the capability?
5. Prerequisites. Does the course teach what it assumes?
6. Whether the record can be obtained without learning. The one that overrides the rest.

All six are questions about the course rather than about the learner, which is why they can be answered before a single real learner sees it. That is the whole basis of a Preflight, and it is why the verdict is "cleared for human pilot" rather than "proven effective."

What this argument does not say