To prove leadership development is changing business performance, a CHRO needs a trace from capability demand to operating evidence, and in my view nothing less survives a finance partner's questions.
I would never ask a CHRO to promise that every leadership programme will produce a clean financial return. I would ask for a credible chain between the investment and the work that changed afterwards. That chain starts with the capability the business needed, follows the behaviour leaders practised and ends with evidence from the operating context where the change mattered. It is the practical answer to leadership development ROI for CHROs: the return figure is the last link, and it is only as good as the links before it.
The common failure is jumping from attendance to performance. A programme runs, leaders report confidence and the people team points to a business result. The missing middle is where credibility is won or lost. Which decisions changed? What did teams experience? What operating condition made the change possible? What else changed at the same time?
How can a CHRO prove leadership development works?
A CHRO proves leadership development works by showing a traceable chain from the business demand, through live practice and observed change in behaviour, to operating evidence, collected at the start, during, at the end and six to nine months after the work. The report has to say what the development contributed and what else was moving at the same time. A precise return that nobody can audit persuades a finance partner less than an honest chain.
Return on investment sits at the end of that chain. SHRM's 2025 guidance gives the arithmetic: take the total tangible benefits, subtract the total programme cost, divide by the total programme cost and multiply by 100. The hard part is the first number, and it is only credible if the links before it exist. Two companion pieces cover the neighbouring questions. I explain why most organisations cannot produce the number at all in why most companies cannot measure leadership development ROI, and how to track capability across the whole organisation in how to measure leadership capability across an organisation. This article is the chain a CHRO can put in front of a chief executive and a finance partner.
Build the Chain Before You Try to Prove Leadership Development Works
Start with the business demand. Perhaps the organisation needs managers to make decisions closer to customers, executives to integrate across units or leaders to build successors. Define the capability in observable terms, then identify the work where that capability should appear. This gives the development programme a job beyond delivering content.
| Evidence layer | Question | Example evidence |
|---|---|---|
| Business demand | What capability does the strategy require? | Faster cross-functional decisions or stronger succession depth |
| Practice | Where will leaders use the capability? | A market decision, talent review or operating redesign |
| Behaviour | What will people see differently? | Clearer ownership, better challenge or more useful feedback |
| Operating change | What becomes easier or more reliable? | A hand-off, decision rhythm or manager routine |
| Performance context | What result moved, and what else influenced it? | Customer, delivery, risk or people outcomes with context |
What should a CHRO collect, and when?
Published guidance converges on timing. Harvard Business Impact's 2024 guide says data should be collected before, during and after a programme, and that a programme should deliver impact in three areas: the learner's experience, on-the-job performance and the business. Jaason Geerts's 2024 framework, published in the journal Behavioral Sciences and drawing on four literature reviews, advises gathering several kinds of data from several sources at the outset, throughout, at the conclusion and six to nine months afterwards. A leadership consultant quoted in SHRM's 2025 article goes further and advises waiting at least nine months after training before returning to participants to see whether they applied it.
The reason for the late checkpoint is practical. Geerts notes that outcomes can take time, so a measurement restricted to the end of a programme risks a false negative, and that confidence gained in the room can decline once people try to apply it at work. Geerts also notes that entertaining sessions earn top ratings straight afterwards without that always translating into application. A satisfaction score at the door and an operating result nine months later are measuring different things, and a CHRO needs both.
ATD's 2025 article by Patti and Jack Phillips of the ROI Institute, which teaches the method it describes, sets out five levels of evidence: reaction, learning, application, impact and return on investment. It insists that the effects of the programme be isolated from other influences, and it says that without application there is no impact. It also says impact is the measure top executives want most, with return on investment second. I read those levels as a ladder. A CHRO who reports only the bottom rung has reported attendance.
Which metric, source and owner belong to each link?
The table below turns the four links into a working plan. The metrics are illustrations of the kind of evidence to look for, and every organisation should replace them with measures it already trusts. The owners follow a principle I hold firmly: the business owns the demand and the operating result, leaders own the practice, and the CHRO owns the method and the integrity of the chain.
| Link | Illustrative metric | Where the data lives | Owner | When to collect |
|---|---|---|---|---|
| Demand | The work problem in plain words, and the decisions that should change | Strategy brief and the executive team's own statement of the problem | Chief executive or business unit head | Before the programme starts |
| Practice | Share of participants working on a live assignment with a named sponsor and a review point | Programme records and sponsor notes | Programme owner with the sponsor | During the programme |
| Transfer | Changes reported by direct reports and peers on three to five repeated questions, plus a decision the leader approached differently | Short pulse questions; a multi-rater review at the start repeated after six to twelve months | The leader's manager, using the CHRO's method | At about three months and again at six to nine months |
| Outcome | The operating signal tied to the original demand, such as decision cycle time, retention in the leader's team or readiness of successors, with a log of other changes | Operating records and people data | Business owner | Six to nine months after, then at twelve |
| Return (optional) | Benefits minus cost, divided by cost, where the method and finance partner support it | Finance-validated benefit and cost figures | Finance partner with the CHRO | At twelve months, not earlier |
Notice what is missing. There is no row for satisfaction, because satisfaction belongs inside the practice link as a design signal. McKinsey's own advice on measuring results points the same way. It suggests assessing behavioural change through a multi-rater review at the start of a programme followed by another after six to twelve months, tracking what happens to participants' careers afterwards, and monitoring business impact where training is tied to real projects.
The four links CHROs should measure
- Demand: The strategic or operating challenge that makes the capability important.
- Practice: The live work where leaders apply the capability and receive feedback.
- Transfer: The observable change in decisions, relationships, routines or behaviour.
- Outcome: The business or people result, interpreted with timing and context rather than claimed as simple causation.
Demand
Write the demand as a work problem rather than a competency label. The organisation needs leaders to make pricing decisions without waiting for the CEO. Managers need to coach performance before a formal process begins. Business units need a shared way to resolve trade-offs. A demand statement tells people why the development matters now.
If the demand cannot be stated clearly, pause before building the programme. Leadership development should not be used to make an unclear strategy sound more actionable.
Practice
Choose practice that resembles the work. A case discussion can build language, but a live decision creates evidence. Give leaders a real assignment, a clear boundary and a review point. Let them practise with colleagues who will experience the consequences, as well as peers in a learning room.
Practice needs support. If a leader is expected to delegate but every decision is pulled back to the centre, the programme is testing an impossible behaviour. Record the design conditions that help or block transfer so the people function can improve the system as well as the learner.
Transfer
Transfer is what happens after the workshop. Ask direct reports what the leader now makes easier, peers what has changed across the seam and the leader what decision they approached differently. Use a small number of repeatable questions so the evidence can be compared without pretending that leadership is a single score.
Published guidance sets the standard I would hold the transfer evidence to. Collect it before, during and after the work, report what was applied, where it was supported and what remained blocked, and read a low transfer result as a prompt to ask whether it is a programme problem, a role problem or an operating-model problem. The link to leadership development ROI matters here: if the data was never collected at the start, nothing later can repair that.
Outcome
Outcomes matter, but they need honest interpretation. A customer result, delivery improvement or retention change can have several causes. Record the timing, the comparison point, other material changes and the confidence of the explanation. It is more credible to show contribution and limits than to claim a clean line of causation the evidence cannot support.
- Choose the business demand: Name the strategic or operating challenge the capability should address.
- Define observable practice: Describe the decisions, routines and relationships where leaders will apply the capability.
- Set a transfer review: Collect structured observations from leaders, teams and peers after the work begins.
- Track operating evidence: Review the relevant decision, customer, delivery, risk or people signals with context.
- Report contribution honestly: Show what changed, what supported it, what else influenced it and what should happen next.
Use a portfolio of evidence
- The decision or operating problem the programme was designed to address
- The live assignments where leaders practised the capability
- Observed changes from teams, peers and stakeholders
- Operating signals that moved during the review period
- Context that may explain or limit the result
- A next decision about scaling, adapting or stopping the intervention
The portfolio gives the board a more useful answer than a satisfaction score. It shows whether the development investment reached the work and whether the work created evidence of change. It also gives the people team a way to improve the programme without defending every design choice.
Include the experience of people who were not enrolled. A development programme can improve an executive's confidence while increasing workload for managers who must absorb the new behaviour. Direct reports may see better clarity, or they may see a new set of demands with no support. The evidence chain should include those effects because leadership capability is experienced by the system around the learner.
Look for negative evidence as well. A capability may be used in one setting and avoided in another. A team may report more challenge in meetings while decisions still return to the same senior person. A customer result may improve while employee trust falls. These contradictions are not inconvenient footnotes. They tell you where the intervention or the operating design needs attention.
Separate contribution from attribution
Contribution asks how development helped make an outcome more possible. Attribution asks whether development caused the outcome by itself. In complex leadership work, contribution is usually the more defensible question. Record the other changes that occurred, the decisions leaders made and the evidence that connects their practice to the result.
This language protects credibility with finance and the board. It does not weaken the case for development. It shows that the people function understands how performance is produced and is prepared to make a decision with the evidence available rather than hide uncertainty behind a precise percentage.
Use a simple confidence statement. High confidence may be appropriate when the same change appears across comparable teams and the operating context is stable. Medium confidence may fit when the pattern is strong but several material changes overlap. Low confidence can still justify another test when the capability is strategically important and the learning is clear.
You will meet headline returns, and it helps to know how much weight they carry. A 2023 survey of 752 professionals carried out by the training platform New Level Work reports an average return of $7 for every $1 spent on leadership development, within a range of $3 to $11, and SHRM's 2025 guidance repeats the average. The figures are built from survey responses and come from a vendor in the field, so I treat them as evidence of what the market believes and not as a benchmark your finance partner should accept. The more useful finding in the same report is that 56% of the companies surveyed had calculated, or were calculating, the return on their programmes. The question your finance partner is about to ask is a common one.
Design comparison carefully
CHROs often want a control group or a before-and-after comparison. Those methods can help, but leadership development rarely happens in a stable environment. Teams change, strategies move and leaders are selected into different contexts. Be explicit about what the comparison can show and what it cannot.
Use matched evidence where possible. Compare similar decisions, teams or time windows. Look for repeated patterns rather than a single favourable story. If a result differs across units, ask what the context reveals. Variation can identify where the capability transfers and where the operating system needs repair.
Make the report useful to operators
A report should help a leader make a decision. Use plain language and show the chain on one page: demand, practice, transfer, outcome and next action. Include one example that makes the change concrete and one limitation that keeps the conclusion honest.
Avoid dashboards that turn every signal into a green light. A good report can say that behaviour changed but the operating result is not yet visible, or that the result moved but transfer evidence is mixed. Those are decisions about timing and design, not failures of storytelling.
Give operators a next action. If the capability transferred but decisions remain slow, repair the authority boundary. If teams see better feedback but managers lack time, change the workload or rhythm. If results move in one unit only, study the local condition before scaling the intervention. Evidence earns its place when it changes what the organisation does next.
The CHRO's role in the evidence chain
The CHRO owns the integrity of the chain. That means asking whether the capability is connected to strategy, whether leaders had a chance to practise it and whether the business context has been described accurately. It also means resisting pressure to convert uncertainty into a number that looks final.
Partner with business leaders rather than asking the people function to carry proof alone. The business should own the demand and the operating evidence. Leaders should own the practice and transfer. The CHRO should bring the method, the challenge and the decision about what to learn next.
Bring the chief executive and finance partner into the design early. Agree what decision the evidence will support, what level of confidence is useful and which signals are already available, so the final report is not a surprise request for approval and the business is a participant in the learning and not an audience for a people metric. Keep the method simple enough to repeat, because a smaller chain used consistently will teach the organisation more than a complex evaluation that appears once at the end.
Also set a boundary around what the programme cannot fix. Development cannot compensate for a broken incentive, an impossible workload or a leadership role with no authority. Naming those conditions protects participants and gives the organisation a clear choice: change the system, change the expectation or change the intervention.
Repeat the review after the work has settled. Early evidence shows practice and transfer, while later evidence shows whether the new behaviour survived pressure, leadership changes and competing priorities. The final report should end with a decision and not a celebration: continue the practice, change the design, collect more evidence or stop. That is how measurement becomes part of leadership governance, and it is how the CHRO can tell whether a programme built capability or only created a temporary burst of attention around it.
A ninety-day evidence cycle
In the first month, agree the demand and define the live practice. In the second, collect structured transfer evidence from people close to the work. In the third, review operating signals, context and limits with the business owner. Then decide whether to scale, adapt, pause or stop. The cycle creates a habit of evidence without pretending that every outcome appears immediately.
The goal is not to prove that leadership development controls the business. The goal is to show that the organisation is building the capability its strategy requires, that leaders are applying it and that the work is producing evidence worth acting on.
The one-page chain to request at your next programme review
At the next programme review, ask for a single page with five headings, namely demand, practice, transfer, outcome and next action. Require one concrete example and one stated limitation. If the page cannot say what the business needed, where leaders practised it and what the people around them noticed, the programme has attendance data rather than evidence. Ask the team to give a confidence level of high, medium or low and to explain why.
Attendance proves the programme ran. Transfer proves it reached the work. Only a chain from demand to outcome proves it mattered, and a chain that states its own limits is the one a chief executive will believe.
If you want to build that chain with a real baseline, the free Leadership Capability Diagnostic and the leadership assessments give you a repeatable starting measure, CapabilityAI helps map where leadership capability is concentrated in too few people, and working with me is the route if you want the method designed with your executive team. For building capability so that it can be measured at scale, see building leadership capability at scale and leadership succession planning.
Three pieces help with the follow-up. Evidence-based leadership development covers moving from intuition to science-backed models, what causes leadership development programmes to fail names the structural reasons behind weak transfer, and the team collaboration and effectiveness framework is useful when transfer depends on how a team works together.