A $60,000 Leadership Program That Moved Nothing

The reaction scores are telling you people enjoyed the class, not that anything changed on the floor.

• Level 1 data cannot answer the question being asked. A 4.5/5 measures whether the room was pleasant. It says nothing about whether supervisors now do onboarding check-ins or give corrective feedback.
• The exit themes point to a different root cause than 'leadership skill.' 'No one checked on me' and 'I didn't know what good looked like' describe a missing onboarding and performance-standards process, not a deficit of leadership theory. A generic program addressed the wrong problem.
• Structural blockers are the usual reason training doesn't transfer. Supervisors who are 'too busy' almost always have spans of control, staffing models, or scorecards that actively punish time spent coaching. Training a supervisor to coach inside a system that measures only handle time produces nothing.
• Gather before buying: (1) last 12 months of exit interviews coded by theme; (2) span of control and supervisor-to-agent ratios by team; (3) a two-week sample of what supervisors actually spend time on; (4) 90-day retention broken out by team - if it varies widely, the problem is supervisor-specific, and a targeted intervention beats a company-wide course; (5) a short pulse survey of new hires at day 7, 30, and 60.

Start with the smallest intervention that addresses the diagnosed cause, and measure at the right levels.

• Likely intervention, given the evidence: a structured 30/60/90 onboarding with a mandatory 15-minute check-in at day 3, week 2, and week 6; a written definition of what 'good' looks like for the first 90 days; and supervisor coaching on those specific conversations. Target the six to eight teams with the worst 90-day attrition first as a pilot.
• Evaluation, mapped to levels:
  • Level 1 - reaction (keep it, but stop treating it as outcome evidence).
  • Level 2 - learning: can supervisors demonstrate the check-in conversation? Short observed role-play, not a quiz.
  • Level 3 - behavior: the key measure. Are check-ins actually happening? Pull from the HRIS or a one-line log. If they aren't happening, the program failed regardless of how people felt.
  • Level 4 - results: 90-day turnover, and first-contact resolution as a lagging indicator.
• Set the baseline now and compare pilot teams against matched control teams over two quarters. If 90-day turnover doesn't move, you have learned the cause is something else - schedule, pay, or staffing - which is a much cheaper lesson than $45,000.
• On the advanced tier: decline for now. Present the pilot design and a decision point, not a purchase order.

Frame it as protecting the development budget, not cutting it.

• The strongest argument is the one the executives already care about: $60,000 was spent and 90-day turnover remains 41%. Spending $45,000 more on the same design is asking them to fund a repeat.
• Reframe the ask: the same $45,000, redeployed - a pilot of structured onboarding plus targeted supervisor coaching, with a control group and a two-quarter read. If it works, you have a business case for scaling that is far stronger than satisfaction scores.
• Bring the data, not the opinion: coded exit themes, retention by team, and the Level 3 measure you'll use. Executives respond to a measurable plan and defensibly to a course-catalog request.
• Preserve the relationship: acknowledge the program was well received and note that it may still be the right answer for the manager population once onboarding and supervision are fixed - but sequence matters. Fix the first 90 days first; leadership development is more valuable for people who stay past them.
• Offer a specific date to report back, and commit to the pilot results regardless of direction. That is what makes HR credible in the next budget cycle.