The individual development plan: why it stays on paper, and what makes it work

An individual development plan stays on paper because it describes an intention rather than a change in the work. It starts working when it contains four things: the observable behaviour you want different, the real work where it gets practised, the cadence of the conversation with the manager, and the evidence someone else will see.

Once a year, in March or in November depending on the company calendar, someone in HR sends out a file. It has three columns: development goal, actions, deadline. Managers fill it in during the last week, employees sign, the file is saved to a shared folder.

Eleven months later the same file opens again. In almost every case the "actions" column holds a course that either happened or did not, and the "goal" column holds a phrase nobody could verify: "improve communication skills", "develop strategic thinking", "show more initiative".

The conclusion everyone draws is that people are not motivated enough to develop. It is a comfortable conclusion and almost always the wrong one. The problem is not motivation. It is that the document never asked for anything verifiable, and a document that asks for nothing verifiable can be neither honoured nor broken.

TL;DR

• An individual development plan usually fails on wording, not on motivation. "Improve communication" can be neither verified nor contradicted, so it commits nobody to anything.
• The 70:20:10 model that quietly underpins most development plans has no empirical base. The numbers come from 191 executives recalling what had shaped them, and Kajewski and Madsen's 2012 review concluded that the data supporting the ratio is missing.
• The development contract: four lines that replace the form. The observable behaviour, the anchor work where it gets practised, the cadence with the manager, and the evidence someone else will see.
• Specific and difficult goals produce higher performance than the instruction to "do your best", provided there is feedback on progress and commitment to the goal (Locke and Latham, American Psychologist, 2002). A development plan with no progress marker loses exactly that mechanism.
• The employee owns the ambition, the manager owns the conditions, and HR owns the system. The common confusion is that HR owns the outcome, which is impossible: HR cannot change what happens in a team's week.
• The anchor work: if development is not practised inside a project already on the calendar, it is not practised at all. Do not add development time on top of a full week; change how one piece of the existing work gets done.

Why does an individual development plan stay on paper?

Because it is built around the HR calendar rather than around the work. It gets filled in during a reporting week, states intentions nobody can verify, and reopens a year later. Between those two moments there is no mechanism that brings it back into anyone's attention.

Three things break at once, and none of them is about the employee's character.

The goal is not verifiable. "Be more proactive" describes an impression, not a behaviour. Two people reading that sentence eight months later will not agree on whether it happened. A goal you cannot contradict is not a goal, it is a wish.

The plan never touches the work. Almost every action written into the form is an event separate from daily activity: a course, a book, a certification. All of them assume spare time in a calendar that has none. The first busy week removes them, and nobody notices, because none of the team's deliverables depended on them.

Nobody owns it between the two dates. HR owns the form, the employee owns the intention, the manager owns the week. Since the week is the only one of the three that actually produces behaviour, the plan disappears precisely where it was supposed to live.

This is the same architectural failure we described when we wrote about why training programmes fail. The content is usually good. The context that should carry it is missing, and without that context every kind of content looks identical: like spend with no effect.

What does the research on 70:20:10 actually say?

That it is a summary, not a finding. The numbers come from 191 executives recalling the events that had shaped them, published in 1988, and the 70/20/10 phrasing first appears in a 1996 career planner. Later reviews found no empirical data supporting the ratio itself.

The chain is worth following, because it explains why so many development forms look the same. Andrew Jefferson and Roy Pollock traced the model's origin and show that the source cited most often is McCall, Lombardo and Morrison's "Lessons of Experience" (1988), in which 191 successful executives were asked to identify at least three key career events and what they learned from each. The phrase 70:20:10 appears nowhere in that book. It appears in Lombardo and Eichinger's "Career Architect Development Planner" (1996), as a rough summary of those conclusions (Jefferson and Pollock, "70:20:10: Where Is the Evidence?", ATD, 2014).

When Kajewski and Madsen went looking for the empirical basis of the ratio in 2012, they concluded that the supporting data is absent. The original sample was American executives recalling in hindsight where they had learned, which is a legitimate source of hypotheses and a very poor source of proportions.

The useful part of the original study gets quoted far too rarely. The coursework that mattered to those executives had two things in common: it dealt with an issue that was relevant to them, and it arrived at a good time. Not volume, fit. The practical difference is large: a ratio tells you how much to budget, a fit condition tells you when to schedule.

This is where most individual development plans break. They copy the ratio (one course, one mentor, one project) without copying the condition, and the condition was the only part the data ever supported.

What does an individual development plan that works contain?

Four lines, written with the direct manager: the observable behaviour you want different, the real work where it gets practised, the cadence of the conversation that carries it, and the evidence someone else will see. One page. The rest of the form is documentation, not development.

We call it the development contract, because it has the structure of one: a clear obligation, a context of execution, a review rhythm, and proof.

1. The observable behaviour. Not "better communication", but "runs the weekly status meeting without escalating decisions that sit inside their own scope". The test is simple: if two colleagues reading the sentence three months later could disagree about whether it happened, the sentence is not finished. This is where the goal research earns its place. Edwin Locke and Gary Latham synthesised thirty-five years of studies and show that specific, difficult goals produce higher performance than easy goals or the instruction to "do your best", provided there is feedback on progress, commitment to the goal, and the ability and knowledge required for the task (Locke and Latham, American Psychologist, 2002). Those conditions are not footnotes. They are half the result.

2. The anchor work. A project, a client, a recurring meeting that already exists in this person's calendar and in which the new behaviour has to be practised. Without an anchor, development is extra work, and extra work always loses to mandatory work. With an anchor, development costs no additional time; it changes how one piece of existing work gets done.

3. The cadence with the manager. A name and a frequency: twenty minutes every two weeks for one quarter, about the anchor. Not a career conversation, not a review. A conversation about how Tuesday's meeting went.

4. The evidence. What someone else will see changed, and by when. "By the end of November, the team makes its prioritisation calls in the Monday meeting instead of emailing them to me afterwards." Evidence is observable by a third party, otherwise we are back to impressions.

The failure pattern repeats with predictable regularity, and the description below is a composite of it rather than any single client. A strong specialist is given the goal "develop leadership skills" and a two-day course in Q3. They attend, they enjoy it, they come back with seven pages of notes. Nobody allocated them a context in which to lead anything. At the annual review they record "course completed", and those seven pages remain the only trace. The form says the plan was executed. The work says nothing changed. Both statements are true, which is the part that should worry us.

Who owns the development plan: the employee, the manager, or HR?

The employee owns the ambition, the manager owns the conditions, HR owns the system. The common confusion is that HR owns the outcome, which is not possible: HR cannot change what this person works on next week. A plan the manager does not hold in their calendar has, in practice, no owner.

Gallup reports that 70% of the variance between the most engaged and the persistently disengaged teams is explained by the manager, based on meta-analytics of 100 million employee interviews in its global database (Clifton, "Gallup Finds a Silver Bullet: Coach Me Once Per Week", Gallup, 2021, updated 2026). The data is international rather than local, but the mechanism has no nationality: whoever controls the week controls the development.

The practical recommendation from the same source is one short, meaningful conversation per week with each person, about progress on goals. Not a system. A habit. And that habit is the only thing that turns a page signed in February into a different behaviour in October.

For HR this means not less work but a different kind of work: building managers' capacity to hold the conversation instead of collecting forms. We have described what that capability looks like in the Learning domain of Capability OS, and for managers new to the role the logic is the same as in the role mandate: write it down early, briefly, and concretely, instead of hoping it gets inferred.

What to do on Monday

Do not redesign the form. Take five individual development plans signed this year and apply a single test.

For each one, ask: can I name the piece of work in this person's calendar where the goal gets practised next week? Not the course. Not the intention. The piece of work.

If the answer is no for most of them, you have your explanation for why nothing is happening, and it has nothing to do with the people. Rewrite those five plans as four lines each, together with their managers. It takes roughly an hour in total and it replaces a year of reporting.

If you want to see where development leaks at the level of the system rather than the individual, a capability diagnostic shows the real constraint before you invest in a course catalogue. And if you work in HR or L&D and recognise your own March spreadsheet in this, what we ask of HR teams sets out the alternative briefly.

So we will leave you with one question. Of the development plans signed in your company this year, how many would survive if we deleted the "actions" column and asked instead for the name of the project where the goal gets practised?

FAQ

What is an individual development plan?

It is the document in which an employee and their manager agree what capability gets built over a given period, and how. In its common form it holds a goal, a list of actions, and a deadline. In the form that works, it holds an observable behaviour, the real work where it gets practised, the cadence of the conversation with the manager, and the evidence someone else will see.

How do you write a development goal that can be verified?

Write it as a behaviour a colleague could observe without asking you. "Improve communication" fails the test. "Sends the decision summary within 24 hours of each project meeting" passes it. Locke and Latham show that specific, difficult goals produce higher performance than vague ones, provided there is feedback on progress.

Why do development plans built on 70:20:10 not work?

Because the ratio has no empirical base. It comes from a 1988 study in which 191 executives recalled in hindsight where they had learned, and Kajewski and Madsen's 2012 review found the data supporting the proportion to be missing. The conclusion that survived is not the proportion but the condition: training matters when it addresses a relevant problem and arrives at the right moment.

How often should an individual development plan be reviewed?

Every two weeks for a quarter, in twenty-minute conversations with the direct manager, about the project where the goal gets practised. An annual review is not a review, it is an archiving step. Gallup recommends one short, meaningful conversation per week with each person, about progress on goals.

Who is responsible for the development plan: the employee, the manager, or HR?

The employee is responsible for the ambition and the effort, the manager for the conditions in which it can be practised, HR for the system that makes the conversation possible. Responsibility cannot be moved wholesale to HR, because HR does not decide what this person works on next week. A plan the manager does not hold in their calendar has no real owner.

Sources

• Jefferson, A. and Pollock, R. (2014). 70:20:10: Where Is the Evidence? ATD Blog. Traces the 70:20:10 model to McCall, Lombardo and Morrison, "Lessons of Experience" (1988), a study of 191 executives recalling formative career events, and to Lombardo and Eichinger's "Career Architect Development Planner" (1996). Quotes Kajewski and Madsen (2012, DeakinPrime) on the absence of empirical data. US and international executive sample, retrospective self-report.
• Clifton, J. (2021, updated 2026). Gallup Finds a Silver Bullet: Coach Me Once Per Week. Gallup. The 70% of variance between the most engaged and persistently disengaged teams explained by the manager, based on meta-analytics of 100 million employee interviews. Gallup global database.
• Locke, E. A. and Latham, G. P. (2002). Building a Practically Useful Theory of Goal Setting and Task Motivation: A 35-Year Odyssey. American Psychologist, 57(9), 705-717. Specific, difficult goals produce higher performance than easy goals or "do your best", conditional on feedback about progress, goal commitment, and the ability and knowledge required. Largely experimental research, North American and European.
• Azimut program data: over 600 programs delivered, over 20,000 participants, 9.27/10 average trainer score (since 2008).

By Szilárd Kacsó