Return on Learning · The guide
How to measure Learning ROI
Learning ROI is how you prove a training or development programme returned more than it cost. This guide covers the formula, a worked example, the two models worth knowing, the metrics that count — and the one step that decides whether your number is honest.
Definition
What is Learning ROI?
Learning ROI (return on investment) is the monetary value a learning programme produces, measured against what it cost to run. Borrowed from finance, the logic is simple: spend money on training, get value back, compare the two. Put plainly — for every dollar you invest in developing people, how many dollars of value come back?
The appeal is obvious. A single percentage is easy to put in a board pack and easy to compare across programmes. The risk is just as obvious: that clean number hides a pile of assumptions. Most of the work in measuring Learning ROI is making those assumptions visible and defensible, not doing the arithmetic.
The formula
The Learning ROI formula, with a worked example.
Learning ROI % = ((benefit − cost) ÷ cost) × 100
Take a leadership programme for 50 managers. Costs are $800 per person plus $5,000 in design and platform — a total of $45,000. You estimate each manager creates $2,500 of annual value through better decisions and lower team attrition — $125,000 across the group.
If you credited the programme with all of that, ROI would look like 178%. But the managers also got a new director, a reorganisation and a better market. Attribute, say, 60% of the gain to the programme and the honest figure is $75,000 of benefit, a net $30,000, and an ROI of 67%. Same programme, two very different stories — and the difference is one honest step.
The models
Two frameworks worth knowing.
Kirkpatrick — four levels
Reaction (did they value it?), learning (did they learn?), behaviour (are they applying it?), and results (did the organisation change?). It is the common language of training evaluation — but most organisations stop at Level 1, the satisfaction “smile sheet”, which predicts almost nothing about impact.
Phillips — the fifth level
The Phillips ROI methodology adds a fifth level on top of Kirkpatrick: convert results into money, and — crucially — isolate the share of the gain the programme actually caused. That isolation step is what separates a defensible figure from a flattering one.
A fuller comparison of Kirkpatrick and Phillips is coming as part of this guide cluster.
What to measure
The metrics that count.
ROI is only as good as the benefit you feed it. These are the tangible measures most often tied back to learning — pick the few your programme is genuinely meant to move, and decide how you will evidence them before it runs.
Productivity
Output per person, cycle time, quality / error rates.
Retention
Lower regretted attrition and the rehire/onboarding cost it avoids.
Time to productivity
How fast new joiners or newly-trained staff reach full output.
Compliance
Certifications achieved; incidents, breaches or claims avoided.
Skills gaps closed
Roles you can now fill internally instead of hiring out.
Revenue
Conversion, deal size or upsell where training touches the number.
Some of the most important effects resist a tidy number. Track them honestly as evidence rather than forcing a false figure:
- Manager and peer confidence in the trained group
- Engagement and discretionary effort
- Psychological safety and willingness to raise issues
- Collaboration across teams and functions
- Employer brand and internal mobility
Why it’s hard
The three things that break a Learning ROI number.
Isolation. Rarely is a programme the only thing that changed. Without a way to estimate its specific contribution — a comparison group, a trend line, or an explicit expert estimate — any ROI figure is a guess wearing a decimal point.
Cost honesty. The visible invoice is the small part. Participants’ time away from the job is usually the largest cost and the most often ignored. Leave it out and ROI is inflated from the start.
Time lag. Behaviour change shows up weeks or months after the course, not on the day. Measure too early and you capture enthusiasm; measure the results that matter and you have to wait, and keep measuring in waves.
The honest alternative
A stack, not a single number.
A single ROI percentage is where most measurement stops — and where it gets least honest. The alternative is not to abandon ROI, but to hold it inside a fuller picture: effectiveness measured as a stack — reaction, knowledge, transfer, results — expressed as a single Learning Outcome Score calibrated to the kind of learning, then set against spend as a cost-per-outcome you can compare across every programme.
We call this Return on Learning: designing the measurement before the programme, around the change it is meant to create — a theory of change — so the evidence is built in, not bolted on. It is how you get a number you can defend, not just one you can quote.
Questions
Learning ROI, answered.
What is ROI in learning?
Return on investment in learning is the monetary value a training or development programme returns, set against what it cost. It is expressed as a percentage — ((benefit − cost) ÷ cost) × 100 — and answers the budget-holder’s question: did this produce more value than we spent on it?
What is the ROI model of training?
The two established models are Kirkpatrick’s four levels (reaction, learning, behaviour, results) and the Phillips ROI methodology, which adds a fifth level that converts results into a financial return and isolates the share of the gain attributable to the programme. Most organisations measure only Kirkpatrick Level 1 — the “smile sheet”.
How do you measure ROI in L&D?
Define the outcome before the programme runs, cost it fully (direct and indirect), put a monetary value on the gain, discount that gain by how much the programme genuinely caused, and apply the ROI formula. The hard part is not the arithmetic — it is honestly isolating the programme’s effect from everything else that changed.
What is the 70-20-10 rule in L&D?
The 70-20-10 model holds that people learn roughly 70% from doing the job, 20% from others, and 10% from formal training. It matters for ROI because most behaviour change happens on the job, after the course — so measurement has to look at transfer and results over time, not just what happened in the room.
What does a good Learning ROI look like?
There is no universal benchmark, and a high headline figure can be the least trustworthy. A modest, well-isolated return you can defend in front of finance is worth more than a 400% number that credits the programme with effects it did not cause. Calibrated confidence beats the marketing gain of overclaiming.
Request a demo
Tell us what you run and we’ll show you SessionData on it.
No spam. No sales calls unless you want one. Just a conversation when the time is right.