Leadership Development Programs: How to Design One That Changes Behavior

Leadership development programs change how managers behave at work more often than their reputation suggests, but the design decides whether they do. When Christina Lacerenza and colleagues at Rice University pooled 335 samples of employees in 2017, on-the-job behavior was better after leadership training on average, and the gains were larger where programs used a needs analysis, feedback, practice and spaced sessions.¹

To design a program that changes behavior, work through seven steps in order:

1. Start from a business problem
2. Choose two to four behaviors to change
3. Spread the sessions out and fill them with practice and feedback
4. Use 360 feedback as a starting point, with goals attached
5. Follow the classroom with coaching or a peer group
6. Bring the participants’ own managers in
7. Measure behavior change

Prefer to watch? The video covers the same ground as this article.

VideoLeadership Development Programs: How to Design One That WorksWiserHours

What is a leadership development program?

A leadership development program is a planned set of activities, usually a mix of sessions, practice at work, feedback and follow-up, meant to change how people lead others. It can serve first-time team leads or senior executives, and it can be built in-house or bought from a provider.

A one-day workshop is an event; a program runs over weeks or months and connects what people learn to the work they go back to. Most design choices below serve that connection, which researchers call transfer of training: whether a skill learned in a session is used, and kept, on the job.²

If what you are planning fits into a single day, call it a workshop and expect a workshop’s results; the rest of this guide, like the other guides on training and developing staff, assumes you want behavior at work to shift.

What makes leadership development programs effective

Leadership development programs work best when they are designed around transfer to the job, according to the broadest summary of the research we found, Lacerenza and colleagues’ 2017 Journal of Applied Psychology meta-analysis. It found gains at every level evaluated and linked specific design features to stronger effects.¹

The study

Moderate evidence

Lacerenza and colleagues: which design features went with bigger gains

Christina Lacerenza and colleagues pooled evaluations of leadership training and found gains in participants’ reactions, learning, behavior on the job and organizational results. Effects were larger in programs that started with a needs analysis, used feedback, combined several delivery methods (especially practice), spread training across sessions, and were delivered face to face on site rather than self-administered. Content, attendance policy and duration also made a difference.¹

The authors call these features associated with more effective programs, because a meta-analysismeta-analysis: A study that combines the results of earlier studies on the same question into one overall estimate. Pooling makes the estimate more precise, but it cannot repair the studies it pools: a meta-analysis of surveys is still survey evidence.Full entry in the glossary like this one compares separate studies rather than testing one design against another. Its abstract leaves the direction of the attendance-policy and duration effects unstated, so this guide leaves them aside.¹

Why should transfer need deliberate design? Leadership is what Brian Blume and colleagues call an open skill: there is no single right way to run a hard conversation, so people must adapt what they learned to each situation. In their 2010 meta-analysis pooling 89 studies, motivation and a supportive work environment were more strongly related to transfer for open skills such as leadership than for closed skills such as using software.³

Consider a newly promoted team lead fresh from a session on giving specific feedback. Back at work, the first chance comes in a busy week, the team member reacts badly, and the lead’s manager never asks how it went. Nothing in the workshop was wrong, but the design stopped at the classroom door.

How to design a leadership development program, step by step

The seven steps run from the problem that justifies a program to the evidence that it worked. The early steps rest most directly on the 2017 meta-analysis; the later ones draw on research about feedback, coaching and transfer.

1. Before: start from a business problem and narrow it to a few behaviors people can practice and others can see
2. During: spread the sessions out, fill them with practice and feedback, and give people real work to try between them
3. After: keep it going with coaching or a peer group and the participants’ own managers, and check behavior with the people who see it

1. Start from a business problem

A needs analysis asks what the organization needs leaders to do differently, and why, before anyone picks content. Programs that began with one were linked to stronger results in the 2017 meta-analysis.¹

A program bought from a catalog teaches whatever the catalog covers, which may be good material about problems you do not have. A needs analysis starts from evidence you already hold: exit interviews, engagement survey comments, missed targets, safety incidents. Suppose a company finds that most people leaving in their first year come from teams whose leads were promoted in the past six months. That points to a program for new team leads on a few specific skills, not a general course on leadership, and it ties the program to the same goal as other strategies for keeping employees.

Put the problem into a single sentence before designing anything. If no one can, the program has no agreed purpose yet.

2. Choose two to four behaviors to change

Name the handful of on-the-job actions that would ease the problem, in words specific enough that a colleague could see them happen: “holds a 30-minute one-on-one with each direct report every two weeks”, not “communicates well”.

Specific behaviors give the program something to practice and something to measure. A 2005 meta-analysis of 117 studies of behavior modeling training, a method in which people watch a skill demonstrated and then rehearse it, found that skill gains were largest when the key learning points were written out as rules.⁴ Transfer was also greatest when trainees saw examples of the skill done badly as well as well.⁴

3. Spread the sessions out and fill them with practice and feedback

Spaced sessions, practice and feedback were each linked to more effective leadership training in the 2017 meta-analysis, which makes them the design choices with the best support behind them.¹

Spacing has a simple logic. Each gap gives people a chance to try the skill at work and bring back what happened. Memory research agrees: a 2006 review of 317 experiments on distributed practicedistributed practice: Spreading study or practice of the same material across several sessions with gaps between them, rather than cramming it into one. Also called spacing, it tends to give longer-lasting learning than the same time spent in a single block.Full entry in the glossary found that spaced study led to markedly better recall than massed study, and the best gap grew with how long the learning had to last.⁵ Those were memory tests, mostly of words and facts, so treat them as a reason the pattern makes sense rather than proof for leadership.

Practice works best with real material: the 2005 behavior modeling meta-analysis found transfer was greatest when trainees practiced with scenarios they generated themselves.⁴

4. Use 360 feedback as a starting point, with goals attached

A 360 survey, in which a leader’s manager, peers and direct reports rate the same behaviors, is useful for showing where a participant starts. On its own it changes little: James Smither and colleagues’ 2005 meta-analysis of 24 longitudinal studies showed ratings generally rising only a little over time.⁶

The same review describes when improvement is more likely: when the feedback shows that change is needed, when the person reacts positively and believes change is feasible, and when they set goals and act on them.⁶ Feedback can also backfire. In Avraham Kluger and Angelo DeNisi’s 1996 meta-analysis of feedback interventions, performance improved on average, yet over a third of the interventions left it worse, and effects weakened as feedback drew attention to the self rather than the task.⁷

So run the 360 on the chosen behaviors, not on a long list of personal traits. Have each participant pick one or two behaviors to work on, and plan the first steps with a coach, facilitator or manager in the same week.

5. Follow the classroom with coaching or a peer group

Follow-up keeps practice going after the sessions end. Coaching has the clearest evidence: a 2016 meta-analysis of 17 workplace studies from seven countries by Rebecca Jones and colleagues linked coaching to positive effects on skills, attitudes and individual results, mostly among managers and senior managers.⁸

Two details from that review are useful for design. Effects were larger with internal coaches than external ones, and smaller when coaching was combined with multisource feedback, which the authors suggest may pull attention toward the feedback itself.⁸ They also found no clear difference between face-to-face and blended coaching, or by number of sessions. With only 17 studies, the authors caution against generalizing far.⁸

Peer cohorts, small groups of participants who meet between sessions to try the behaviors and compare notes, are cheaper and common. We found no meta-analysis that tests them on their own, so count them as a sensible way to deliver practice and feedback, not a proven ingredient. A peer group for everyone plus coaching for those in the hardest roles is a reasonable compromise.

6. Bring the participants’ own managers in

Participants’ own managers decide whether new behavior is noticed, asked about and rewarded. In the 2005 behavior modeling meta-analysis, transfer was greatest when trainees’ superiors were also trained and when the work environment attached rewards and sanctions to the new behavior.⁴

A 2011 review by Rebecca Grossman and Eduardo Salas named the work-environment factors with the most consistent links to transfer: a climate that favors using new skills, support at work, the opportunity to perform the skill, and follow-up after training.⁹

The cheapest version takes an hour: brief each participant’s manager on the chosen behaviors, ask them to hold a short conversation after each session about where the participant will try the skill, and give them two or three questions to ask a month later.

7. Measure behavior change

Measure the chosen behaviors before the program and again a few months after it ends, using ratings from people who see the participant work. Satisfaction is the wrong target, because George Alliger and colleagues’ 1997 meta-analysis found that whether trainees found a course useful was more closely related to later performance on the job than whether they enjoyed it.¹⁰

Who rates matters too. Blume’s 2010 meta-analysis found that transfer relationships looked stronger when the outcome was measured by the same source in the same setting, which inflated them.³ In practice, ask direct reports or peers rather than relying on participants’ own ratings, and use the same questions before and after. The four levels of training evaluation in the Kirkpatrick model give a vocabulary for planning the rest. Where you can, start one group a few months later than another and compare the two; that is the nearest most organizations get to a fair test.

Before you launch a leadership program

• The business problem is written in one sentence
• Two to four behaviors are named in observable terms
• Sessions are spaced weeks apart, with practice and feedback in each
• 360 or baseline ratings cover the chosen behaviors only
• Coaching or a peer group is booked for the months after
• Participants' managers are briefed and know their follow-up questions
• Behavior will be rated by others before and a few months after

How long should a leadership development program be?

There is no tested ideal length, but the evidence favors a program spread over weeks or months rather than packed into a few days. The 2017 meta-analysis linked spaced sessions to more effective leadership training and reported that duration also mattered, though its abstract does not say how.¹

In the behavior modeling meta-analysis, skill development was greatest when training time was longest, and effects on skills and job behavior held steady or grew over time even as factual knowledge faded.⁴ How many coaching sessions people had did not clearly change the size of the effects in the coaching meta-analysis.⁸ Taken together, a practical shape is three to six months: a handful of spaced sessions, practice at work in between, follow-up afterward, and a behavior check at the end.

Myth

An intensive week away is the most efficient way to develop leaders.

Fact

Spaced sessions were linked to more effective leadership training in a 2017 meta-analysis; one long block leaves no time to practice at work between sessions.

Which design choices rest on solid ground

Needs analysis, spaced practice with feedback, and manager involvement have the firmest support; peer cohorts have none of their own yet. Most rows below come from meta-analyses that pool studies of mixed design, so they show which features travel with better results, not that each one causes them.

Two limits apply throughout. The key 2017 meta-analysis was read here only as an abstract, so this guide reports its findings without the effect sizes or study details behind them. And published evaluations tend to come from programs someone thought worth studying, so the average program may do less well. If budget forces a choice, protect the practice between sessions and the manager briefing before adding content. Whether leadership programs pay off overall is a broader question, saved for another guide.

The bottom line

The weeks between sessions are where new leadership behavior is tried or dropped, so design them as carefully as the sessions. Start from a business problem and two to four visible behaviors, and book coaching or a peer group and brief participants’ managers before the first session. Then judge the program by what colleagues see months later.