The Four Levels of the Kirkpatrick Model: How Training ROI is Measured

Chief Learning Officers and human resources directors decide which corporate training programs receive funding, which are expanded, and which are cut. They make these allocations during annual budget planning, relying on evaluation data to justify the spend. To do this, they use the Kirkpatrick Model, a four-level framework that measures whether an educational intervention actually changed employee behavior and delivered a financial return.[3]

Developed by educator Donald Kirkpatrick in 1959, the model breaks training evaluation into four sequential tiers: Reaction, Learning, Behavior, and Results. Moving up the tiers increases both the strategic value of the data and the operational cost to collect it. The Association for Talent Development notes that each level builds upon the information gathered from the previous one, providing a comprehensive approach to assessing training impact.[1][4]

Level 1, Reaction, costs almost nothing to deploy. It relies on post-training surveys—often called "smile sheets"—to measure learner satisfaction immediately after a session concludes. While ubiquitous, it provides zero evidence of business impact, measuring only how participants felt about the experience and the instructor.[4][5]

Level 2, Learning, introduces formal assessments. It tests whether the knowledge was actually acquired through pre- and post-training exams. A 100 percent completion rate on a compliance module satisfies Level 2, proving that the employee understands the new regulation, even if they never apply it in their daily workflow.[1][4]

Level 3, Behavior, requires managers to observe employees on the job, typically 30 to 90 days post-training. This level measures transfer—whether the classroom knowledge changed daily habits. According to Kirkpatrick Partners, "The Kirkpatrick Model is the world's most widely used framework for measuring the impact of learning," but Level 3 is where most organizations stall because it requires cross-departmental coordination and manager bandwidth.[5]

Level 4, Results, ties the training directly to business metrics. This is the return on investment calculation: did the $50,000 sales training program yield a 15 percent increase in closed deals? Did the safety module reduce workplace accidents by 20 percent over six months? Clinical environments, such as cricothyroidotomy simulation training, use Level 4 to measure direct reductions in patient mortality rates.[2][3]

The trade-off is stark. While an estimated 80 percent of organizations stop at Level 1 due to its low cost, only Levels 3 and 4 provide the behavioral and business-impact data that executives actually use to justify budgets. This creates a persistent gap between what learning departments measure and what the C-suite requires to authorize funding.[3][6]

In the 2010s, Jim Kirkpatrick and Wendy Kayser Kirkpatrick updated the framework into the New World Kirkpatrick Model. This revision emphasizes planning backward: organizations must define the Level 4 business result first, identify the Level 3 behaviors needed to achieve it, and only then design the Level 2 learning experience to support those behaviors.[5]