Did it feel relevant and usable?
Look for usefulness, clarity, confidence, trust, and perceived relevance.
Useful when adoption and learner trust matter.Field Guide
Before you measure learning, decide what evidence would matter.
Use this guide to turn evaluation into a design tool. Map what should change, where that change would show up, and what evidence would be credible before the learning launches.
Move beyond completion and smile sheets.
Define stronger evidence before launch.
Connect learning signals to behavior and outcomes.
What this helps you do
Turn evaluation from a reporting ritual into a design decision.
Evaluation is often treated as something that happens after the course launches. This guide flips that order. Before building, decide what evidence would show that the learning is relevant, understood, applied, and connected to real work.
✓
Clarify success earlier
Define what should change before the project turns into content production.
✓
Choose better evidence
Separate useful signals from data that only proves people clicked through.
✓
Design better practice
Use expected evidence to shape scenarios, feedback, reflection, and reinforcement.
✓
Have sharper stakeholder conversations
Move from “Did people like it?” to “What would convince us this helped?”
What to inspect
The four evidence layers
Kirkpatrick is useful when it becomes a design lens, not a checkbox. Each layer helps you ask what kind of evidence would be meaningful for the problem you are trying to solve.
Can they recall, decide, or practice?
Look for retrieval, decision quality, confidence calibration, and practice performance.
Useful when capability needs to be built.Does it transfer into work?
Look for workflow use, manager observation, support requests, quality signals, and behavior change.
Useful when performance depends on application.Did anything meaningful improve?
Look for risk reduction, consistency, speed, customer impact, quality, or operational outcomes.
Useful when learning supports business results.Field guide tool
Run a quick evidence mapper.
Rate how strong your current evidence plan is across the four layers. The point is not to collect everything. The point is to choose evidence that matches the learning problem.
R
Reaction
Do you know whether learners find the experience relevant, useful, clear, and worth trusting?
L
Learning
Do you have evidence that people can recall, explain, choose, practice, or apply the important ideas?
B
Behavior
Do you have a way to see whether the learning transfers into the actual workflow?
O
Outcomes
Do you know what operational, quality, consistency, risk, or customer signals should improve?
Weak evidence signals
Signs evaluation is being treated like reporting.
01
Success is mostly defined by completion, attendance, or satisfaction.
02
The evaluation plan is created after the learning is already built.
03
Stakeholders want business impact but only collect learner reactions.
04
No one knows where behavior change would actually show up in the workflow.
Design questions
Ask better questions before choosing metrics.
01
What should people do differently after this experience?
02
What evidence would show they can do it?
03
Where would that evidence appear in the real workflow?
04
What can we reasonably measure without pretending certainty?
Request a Field Guide
What should the Lab build next?
Have a learning design problem, audit idea, checklist, framework, or messy stakeholder conversation that deserves a practical guide? Send it in. The best Field Guides usually start with a real design problem.
Build better evidence plans
Evaluation should shape the learning, not chase it after launch.
The strongest evaluation plans clarify what should change, what evidence would matter, and how the learning system will support transfer before the first asset is built.