Parker Lee · Instructional Design

Kirkpatrick L1–L4 Planning Canvas

Fill this in with your sponsor at kickoff. If L3 and L4 stay blank, that's a signal — not a problem to solve after launch.

L1
Reaction
Did learners find it relevant, engaging, and well-delivered?

What we measure

Perceived relevance, satisfaction, and whether learners would recommend it. Not "did they like it" — "did they find it useful."

  • Overall satisfaction (1–5)
  • Content relevance to their job
  • Pacing and length
  • Clarity of objectives

How we capture it

3–5 question survey immediately at course end. Completion rate drops fast after 5 questions — keep it short.

  • LMS end-of-course survey
  • Reaction survey embedded in course
  • 1–2 open qualitative questions
Low effort to collect

Who owns the signal

L&D designer reviews each cohort. Results feed into design improvements, not stakeholder reports.

  • L&D: reviews scores + qualitative
  • Informs next iteration, not business case

Honest ceiling

L1 measures perception, not learning. High score ≠ retention. Low score is useful — it flags confusion or credibility problems worth fixing.
L2
Learning
Did knowledge, skill, or confidence actually change?

What we measure

Knowledge acquisition, skill demonstration, and confidence shift from pre to post training.

  • Knowledge check scores (pre vs. post)
  • Scenario performance in branching
  • Scored simulation accuracy
  • Self-reported confidence shift

How we capture it

In-course assessments and pre/post checks. If you're not adjusting design based on failure rates, you're not using L2 data.

  • Knowledge checks (LMS tracks)
  • Branching scenario scoring
  • Pre-course baseline survey
  • xAPI statements for interaction data
Moderate effort

Who owns the signal

L&D tracks failure rates by question and scenario. Manager sees aggregate completion status.

  • L&D: question-level failure analysis
  • Manager: pass/fail completion status
  • L&D adjusts design based on results

Honest ceiling

L2 measures what they knew in the training context — not what they'll do on the job. High L2 with low L3 is common and usually a design problem, not a learner problem.
L3
Behavior
Are learners applying it on the job, 30–90 days out?

What we measure

Observable behavior change in the work environment. Not what they say they're doing — what managers observe or systems record.

  • Manager observation reports
  • Behavioral checklist at 30/60/90 days
  • System/tool usage metrics
  • Quality audit results post-training

How we capture it

Requires manager pre-commitment before launch. If managers aren't in the plan before rollout, L3 data usually doesn't happen.

  • Manager observation form
  • Pre/post quality audit
  • System data pull (if tool usage is the behavior)
  • Follow-up survey to manager + learner
High effort to collect

Who owns the signal

Shared between L&D and the direct manager. L&D designs the instrument; manager collects and reports.

  • Manager: observation + reporting
  • L&D: instrument design + synthesis
  • Sponsor: sees aggregate trend

Honest ceiling

Even real behavior change is hard to attribute to training alone. Acknowledge other factors — manager coaching, process changes — in any L3 readout.
L4
Results
Did business outcomes move — and can we defend the causal link?

What we measure

The business metric this training was supposed to move. Defined at kickoff with the sponsor — not retrofitted after launch.

  • Compliance / audit pass rates
  • Quality scores or error rates
  • Sales conversion metrics
  • Customer satisfaction scores
  • Time-to-productivity for new hires

How we capture it

This data usually exists in the business already. The work is connecting L&D's training cohort to the relevant metric window.

  • Business reporting / BI tools
  • Pre/post with control group
  • Time-series analysis vs. baseline
  • Partner with data/analytics team early
Very high — plan at kickoff

Who owns the signal

Sponsor owns the business metric. L&D's job is to connect it to the intervention and be honest about what can be attributed.

  • Sponsor: business metric tracking
  • L&D: connects cohort to data
  • Analytics team: statistical support

Honest ceiling

L4 attribution is almost never clean. Market conditions, leadership, and process changes affect the same metrics. Overpromising damages L&D credibility more than admitting uncertainty.