How to Measure Training ROI
Training ROI becomes credible when you capture full cost, pick one business metric and compare against a control group.
By FOGS Consultants L&D Practice · Corporate learning advisory team
Published · Updated · 9 min read
Capture the full cost, not just the trainer fee
Fully loaded cost includes trainer fees, travel, venue, materials, platform cost and — usually the largest line — participant time away from work. Reporting ROI on trainer fees alone destroys credibility the moment finance recalculates it.
Pick one business metric per programme
Choose a metric the programme can plausibly move within the measurement window: rework rate, safety incidents, first-call resolution, sales conversion, attrition in the trained team, or statutory penalty incidents.
Use the four levels, but weight the last two
Reaction and learning are easy to collect and weak evidence. Behaviour (level 3) and results (level 4) are what justify budget. Schedule a 60 to 90 day manager observation as a formal step, not an afterthought.
- Level 1 — Reaction: post-session feedback
- Level 2 — Learning: pre and post assessment delta
- Level 3 — Behaviour: manager observation at 60 to 90 days
- Level 4 — Results: the chosen business metric
Compare against an untrained group
Where cohorts are phased, the not-yet-trained group is a natural control. Comparing trained versus untrained over the same period removes most of the 'the market improved anyway' objection.
Report it simply
ROI equals net benefit divided by fully loaded cost, expressed as a percentage, with the assumptions listed. One page, with the metric, the window, the control and the assumptions, is more persuasive than a dashboard nobody trusts.
Frequently asked questions
What is a good training ROI?
It varies by programme type. Compliance training is usually justified by avoided penalty and risk rather than a percentage return, while sales and productivity programmes are commonly evaluated against a 3 to 6 month payback.
How long after training should we measure?
Measure learning immediately, behaviour at 60 to 90 days, and business results over one to two quarters depending on the metric's natural cycle.
Can ROI be measured without a control group?
Yes, but confidence drops. Use trend comparison against the same team's prior period and document confounding factors explicitly.