← Back to Insights October 14, 2024

What L&D Gets Wrong About Manager Training ROI

By Lauren Humphrey · 8 min read

Learning and development teams measure what's easy to measure. Completion rates are easy to measure. Satisfaction surveys are easy to administer. Kirkpatrick Level 1 ("did participants like the training?") requires 5 minutes of post-session survey collection. That's why most L&D ROI reporting stops at Level 1 - not because teams don't know the limitation, but because the alternative requires more organizational infrastructure and a longer time horizon to show results.

The problem is that Level 1 metrics predict nothing about business outcomes. A manager who rated a training program 4.5 out of 5 and did nothing differently afterward generated zero ROI. A manager who rated the same program 3 out of 5 but changed three specific behaviors generated real ROI. The satisfaction score and the business impact are simply not correlated in the research.

Kirkpatrick Level 3 Is Where ROI Actually Lives

The Kirkpatrick model maps training evaluation across four levels: reaction (did they like it), learning (did they absorb the content), behavior (did they change what they do), and results (did the organization benefit). The ROI of manager training lives at Level 3 - behavior change - and becomes visible in business metrics at Level 4. Almost no corporate training measurement reaches Level 3. A small minority gets to Level 4.

Level 3 evaluation for manager training requires observing or surveying behavior change after the training. Specifically: are managers who completed the program running their 1:1s differently? Are they giving more specific feedback? Are their direct reports reporting higher engagement or development quality? This requires measuring actual behavior, not self-reported recall of training content.

The standard vehicle for Level 3 measurement in manager training is a 360 survey sent to each manager's direct reports at 60 and 120 days after program completion. Four to six questions covering the specific skills trained. Baseline the same survey before the program starts. The delta is your Level 3 evidence. If manager behavior changed in the expected direction, the program did what it claimed. If it didn't, the program failed regardless of what satisfaction scores showed.

The Engagement Score as a Leading Indicator

Team engagement scores are one of the most actionable Level 4 metrics for manager development investment. Gallup's meta-analysis of engagement data found that manager behavior is the primary determinant of team engagement - and that a 10-point increase in team engagement score correlates with a 22% reduction in turnover, a 21% increase in productivity, and a 10% improvement in customer scores.

If an organization runs quarterly engagement pulse surveys by team, manager development investment can be measured directly: did teams whose managers completed a development program show engagement score improvement compared to teams whose managers did not? That's a controlled comparison with real business metric output. It's not perfect causal attribution - engagement scores have multiple drivers - but it's meaningful evidence of direction and magnitude.

The companies in our enterprise partnerships that run this measurement consistently report engagement score increases of 14 to 19 points across manager cohort participants in the 6 months following program completion. That figure, applied to the Gallup correlations, produces a retention and productivity impact that exceeds the program cost by a factor the finance team can understand.

Measuring Voluntary Turnover on Managed Teams

Voluntary turnover is a direct cost with a calculable price: recruiting fees, onboarding time, productivity ramp-up, and the informal knowledge that walks out the door. Most estimates put the total cost of a voluntary departure at 1.5 to 2x the departed employee's annual salary. For a company with 100 managers each leading a team of 8, a 5% reduction in voluntary turnover on those teams translates to a specific dollar figure that finance will recognize.

Tracking voluntary turnover by managed team and correlating it with manager development participation is a Level 4 measurement that requires only data HR already has. The analysis is not complicated. The willingness to run it and report the results to leadership is what separates L&D teams that get sustained budget from those that fight for it annually.

The Problem With Attributing Too Much

A legitimate critique of L&D ROI measurement is the attribution problem. Team performance is driven by many variables. Attributing an engagement score improvement entirely to a manager training program is methodologically questionable if you haven't controlled for other changes - new product launches, leadership changes, office environment changes, compensation adjustments.

The appropriate response to this critique is not to abandon measurement but to be precise about what you're claiming. "Teams whose managers participated in the program showed X-point engagement improvement compared to baseline. Teams whose managers did not participate showed Y-point improvement over the same period. The delta is Z." That's not perfect causal attribution, but it's evidence that the program is likely contributing to an outcome and it's honest about the methodological limit.

L&D teams that overclaim ROI lose credibility with finance partners quickly. L&D teams that show rigorous methodology with honest limitations build the credibility to sustain budget through economic cycles when training is often the first line item cut.

Building the Measurement Infrastructure Before the Program

The common mistake in L&D measurement is designing the program first and the measurement second. When measurement is designed after the fact, it can only capture whatever data happened to be collected. The metrics may not align with the program's stated objectives. The baseline comparison is often missing because no one thought to capture it before the program started.

The correct sequence: define the business outcomes you're trying to influence first. Then design the training program backward from those outcomes. Then build the measurement infrastructure that will capture whether those outcomes changed. The measurement design takes perhaps 10% of the effort of the program design. It is worth every bit of that effort if it means the program can demonstrate its impact rather than hoping someone believes the satisfaction scores.

As we discuss in our piece on cohort learning vs. solo training, behavior change duration varies significantly by program format. Building that distinction into your ROI measurement model - comparing 6-month outcomes by program type - is how you optimize the mix of formats over time.