Leadership development that works: a data-driven guide for HR

Organisations spend heavily on leadership programmes, yet few can say what those programmes actually changed. Even HR Norge, in the framing of its own leadership development conference, puts it bluntly: much is invested in heavy and complex leadership development programmes, but it is demanding to know what actually creates value.

That is the honest starting point for this guide. Leadership development (lederutvikling, as it is known in Norway) too often means sending every manager on the same course, with no diagnosis beforehand and no measurement afterwards. This article flips the model: start with the data your teams already generate, develop each leader against real signals, and measure the effect where it matters, in the team.

What is leadership development, and why does so much of it fail?

Leadership development is the structured work of improving how managers lead: how they build trust, give feedback, set direction and respond to what their teams tell them. Done well, it is continuous and tied to evidence from the leader's own team. Done badly, it is an annual course catalogue detached from daily leadership.

The most common failure modes are easy to recognise:

• One programme for everyone. A leader struggling with psychological safety gets the same content as one struggling with prioritisation. Neither gets what they need.
• No baseline. Without a measurement of how each team experiences its leader before the programme starts, there is nothing to compare against afterwards.
• No follow-up. Insight from a two-day workshop decays quickly if nothing in the leader's weekly routine reinforces it.
• Success measured by attendance. Completion rates and satisfaction scores tell you people showed up and enjoyed themselves, not that anything changed for their teams.

The alternative is not more content. It is better targeting, and a feedback loop that closes.

The data-driven model for leadership development

Leadership development becomes measurable the moment you treat the team's experience as the outcome variable. A leader's development need is visible in how their team answers questions about trust, feedback, clarity and wellbeing, continuously, not once a year. That gives HR a simple five-step model.

1. Establish a baseline per team. Use continuous engagement or temperature data, pulse surveys, eNPS, category scores such as leadership, participation and meaningfulness, to see how each team experiences its manager today. If you only run an annual survey, your baseline is twelve months stale before the programme even begins.
2. Diagnose, then segment. Group leaders by signal, not by seniority. Some teams will show low scores on feedback frequency; others on clarity of direction; others on psychological safety. Each pattern points to a different development need.
3. Match development to the signal. A leader whose team reports weak feedback culture needs practice in continuous feedback conversations, not a generic strategy module. A leader with low trust scores needs to work on the behaviours that build trust-based leadership, we outline five of them in our guide to trust-based leadership.
4. Build the new behaviour into weekly routine. The development happens between sessions: acting on this week's team feedback, holding better one-to-ones, closing the loop on what employees raised. The programme should give leaders something to do every week, not something to remember from a slide deck.
5. Measure the effect in the team, not the classroom. Return to the same signals after 90 days. Did feedback scores move? Did eNPS shift? Did participation improve? That is the evidence line HR can put in front of the executive team.

Notice what changes in HR's role here. Instead of buying and administering programmes, HR becomes the function that connects development investment to team outcomes, the same shift we describe in how HR can develop better managers.

Why targeting beats volume

Generic programmes assume all leaders need roughly the same things. But once you look at team-level data inside a single organisation, it is reasonable to expect wide variation: some teams experience clear direction and strong trust in their manager, others report the opposite, even under the same programme catalogue and the same HR policies. That variation is exactly what a one-size-fits-all programme cannot address, and exactly where targeted development creates value.

Targeting also solves the motivation problem. Leaders are notoriously sceptical of development initiatives they perceive as HR box-ticking. Show a leader their own team's data, anonymised, trended, specific, and the conversation changes from "HR wants me on a course" to "my team is telling me something, and here is help to act on it". The data creates the demand for the development.

Leadership development should start where leadership actually happens: in the signals a team sends its manager every week.

Connecting leadership development to retention and performance

The business case for leadership development runs through the team. How a manager leads ripples outward into engagement, wellbeing and whether people stay, a dynamic we explore in the ripple effect of great leadership. That means the return on leadership development shows up in metrics HR already tracks:

• Retention. Track voluntary turnover and leadership-category scores per team before and after development. If teams whose leaders went through targeted development retain people better than comparable teams, you have a defensible effect estimate. Our guide to understanding and reducing employee turnover covers how to structure that analysis.
• Engagement trend. A single post-programme survey is noise; a 90-day trend in temperature data is signal. Look for sustained movement, not a spike the week after the workshop.
• eNPS. Employee Net Promoter Score is a blunt but board-friendly measure. If targeted leadership development moves eNPS in the affected teams while control teams stay flat, the investment argument writes itself.
• Sickness absence and early-warning signals. Persistent stress or workload signals in a team are often leadership problems in disguise. Development that changes how a leader distributes work and follows up wellbeing should show up here over time.

What to measure per leader

Keep the leader-level scorecard short. Three to five team-level indicators, for example leadership score, feedback frequency, clarity, eNPS, measured continuously, with a 90-day review cycle. More metrics dilute focus; fewer make the trend fragile.

Common objections, and honest answers

"Isn't this surveillance of managers?" No, provided the data stays at team level, is anonymised, and is framed as support rather than scoring. The leader sees how their team experiences their leadership; HR sees where to direct help. The moment the data becomes a punishment tool, both response rates and trust collapse.

"Our leaders don't have time for continuous development." They do not have time for development that is separate from the job. That is precisely the argument for building it into existing routines: the weekly review of team feedback, the one-to-one, the team meeting. Ten focused minutes a week on real signals beats two offsite days a year.

"We can't prove causality." True, team dynamics have many drivers. But a before/after trend, compared against teams that did not receive the intervention, is a far stronger evidence base than the satisfaction survey most programmes rely on. Perfect causality is not the bar; better evidence than today is.

What to do on Monday morning

You do not need a new programme to start. You need a baseline and a first segment.

1. Pull the leadership-related data you already have, engagement scores by team, eNPS, exit-interview themes. If you have continuous temperature data, trend it per team over the last two quarters.
2. Identify the spread. Rank teams by leadership-category scores. The gap between your top and bottom quartile is your development map.
3. Pick one signal and one segment. For example: the five teams with the weakest feedback scores. Design a short, practical intervention for those leaders only, training plus a weekly routine for acting on team feedback.
4. Set the 90-day checkpoint now. Decide before you start which three metrics will define success, and book the review meeting.
5. Tell leaders why they were selected. Show them their team's data and offer the development as support. Selection based on real signals, communicated respectfully, is what separates targeted leadership development from remedial training.

Leadership development stops being an act of faith the moment you connect it to the team's own voice. Start with the signals, develop against them, and measure where the leadership actually lands, in the everyday experience of the people being led.