Every leadership program ends the same way. The room is warm, the feedback forms are strong, and eleven people write down a commitment they mean. Ninety days later the organization is running exactly as it did before.
This is not a small or local failure. Writing in Harvard Business Review in 2016, Michael Beer and his co-authors called it the great training robbery. Companies spent $160 billion in the United States and close to $356 billion globally on training and education in 2015 alone, without a matching return.
The practical problem for anyone buying this work is that the program itself is usually fine. The content is sound, the facilitator is good, the participants engage. What fails is everything that happens after they walk out, and almost none of it is under the provider’s control.
The decay is measurable
There is a number attached to this. In research reported by Saks and Belcourt, training professionals across 150 organizations estimated that 62 percent of employees applied what they learned immediately. At six months it was 44 percent. At one year, 34 percent.
Read that as a curve rather than a verdict. Two thirds of people do try. The loss happens between the trying and the sixth month, which means the intervention point is not the program and not the selection of participants. It is the period nobody budgets for.
It also means the standard fix, more content, is aimed at the wrong end of the problem. Adding a second module improves the 62 percent slightly. It does nothing to the slope.
The classroom is not where it fails
A leader leaves a workshop having decided to delegate more, to run shorter meetings, and to give feedback closer to the event. All three are reasonable. All three are then tested against a calendar they did not design and a boss who rewards responsiveness. And against colleagues who expect the person who walked in on Monday.
Beer and his co-authors put it plainly: people revert to their old ways because the context around them sets the stage for success or failure. If the system does not change, it will not sustain individual behavior change. It will set people up to fail.
That reframes what you are buying. A leadership program is not a knowledge transfer. It is an attempt to change behavior inside a system that is currently producing the behavior you dislike, and which will keep producing it unless something else moves.
Six barriers that beat any syllabus
The same HBR work names six organizational barriers that show up together often enough that the authors call them silent killers. They are worth reading as a checklist before commissioning anything.
- Unclear direction on strategy and values, which produces conflicting priorities.
- A senior team that does not work as a team, and has not committed to a new direction or to changing its own behavior.
- A top down or hands off style from the leader, which prevents honest conversation about problems.
- Poor coordination across businesses, functions or regions.
- Inadequate leadership time and attention given to talent.
- Employees who are afraid to tell the senior team what is actually getting in the way.
Note what is not on that list. Not the quality of the training. Not the learning platform. Not whether the cohort was engaged. Every item is something the sponsor owns and the provider cannot fix.
Four conditions that make learning hold
Programs that survive contact with the business tend to have four things in common, and they are design decisions rather than content decisions.
A real problem attached. Participants work on something their organization actually needs solved, with a named owner and a date. Practice on a case study is a rehearsal. Practice on live work is the work.
The boss in the loop. Each participant’s manager knows what the person is trying to change, and is asked about it afterwards. This single mechanism does more than any amount of post-program content.
Spacing, not a single event. Sessions separated by weeks of application, with a structured return, beat an intensive week. The gap is where the learning is tested.
One removed obstacle. Before the program runs, the sponsor identifies and removes one structural thing that makes the new behavior costly: a meeting, an approval threshold, a metric. One is enough to prove the organization is serious.
Sequence the system before the syllabus
The uncomfortable implication of the research is about order. If unclear strategy and a divided senior team are the binding constraints, teaching two hundred managers to coach their people will not help. It may make things worse, by raising expectations the system cannot meet.
The sequence that works is to fix the direction and the senior team first, then use development to build the capability the new direction requires. That is slower, and it is the version that produces a different organization rather than a well liked program. Since this order also changes what you spend, it is worth reading alongside what a leadership development program actually costs.
Measure something the finance team already tracks
Most programs are evaluated on reaction and on self-reported confidence, both measured on the last day, when everyone is well disposed. Those numbers are real and they predict nothing.
Pick two or three measures the business already collects and would notice moving. Time to fill internal roles. Regretted attrition in the participating population. Cycle time on a decision the program targeted. Agree them before the first session, with the person who owns the number.
The point is not proof. Attribution at this level is genuinely hard and anyone promising a clean causal claim is selling. The point is that a number somebody already cares about keeps the work attached to the business after the enthusiasm fades.
Four things worth fixing, and three not worth arguing about
- Worth it: the sponsor’s own behavior. If the senior team is not visibly doing the thing the program teaches, participants read the program as optional. This is the largest single factor and the least often addressed.
- Worth it: conflicting priorities. When the strategy is ambiguous, every new behavior competes with an old one that is still being rewarded. Leaders are not resisting. They are choosing correctly given the signals.
- Worth it: the calendar. New behavior needs somewhere to live. If nothing was removed, nothing was added.
- Worth it: fear of naming obstacles. If people cannot tell the senior team what is blocking them, the design will keep solving the wrong problem with better materials.
- Not worth it: generational difference. It is the most discussed and least predictive explanation for why a program failed.
- Not worth it: the delivery format. Virtual against in person is a logistics and cost question. Coaching research has repeatedly failed to find a meaningful difference in outcome between formats.
- Not worth it: participant enthusiasm. End of program scores are high almost everywhere, including in programs that change nothing. A low score tells you something. A high one does not.
The step that gets skipped when the program ends
It is the manager conversation at week four. Not a survey, not a platform nudge, not a refresher module. A scheduled fifteen minutes in which each participant’s boss asks what they are trying to do differently, what has been hard, and what needs to change around them.
It gets skipped because it is nobody’s deliverable. The provider is finished, the learning team is running the next cohort, and the manager was never briefed. Build it into the contract, name who runs it, and put it in diaries before the program starts. If your provider treats that as out of scope, you have learned something useful about how to choose a leadership development partner.
If your last program was well received and changed nothing, the diagnosis usually lives in the system rather than the syllabus. Our leadership development work starts there, with what the organization currently rewards.
Frequently Asked Questions (FAQs)
Is it true that most training is forgotten within a week?
That claim is usually a loose reading of nineteenth century memory research and it does not describe workplace learning well. The better evidence looks at application rather than recall, and shows a decline from roughly 62 percent applying training immediately to 34 percent a year later. The problem is sustained use in context, not memory.
Does follow-up coaching make training stick?
It helps, and it is not sufficient on its own. Coaching gives a leader somewhere to take the friction they hit when they try something new, which is exactly where most attempts die. It cannot overcome conflicting priorities or a senior team that models the opposite behavior.
How long should we wait before judging a program?
Look at behavior at ninety days and at business measures at six to twelve months. Anything assessed on the last day is measuring mood. The six month point matters most, because that is where the application curve falls fastest.
Should we train the senior team first?
Usually yes, and not as a courtesy. Four of the six documented barriers sit with the senior team, so starting lower in the organization means asking people to adopt behaviors the system above them still penalizes. Starting at the top also tests whether the appetite for change is real before you spend at scale.