Six months into the new L&D job, you launch the manager coaching program that leadership asked for. Forty-five managers go through it. Feedback scores are strong. People say they learned something.
Three months later, nothing has changed. Same managers, same avoided conversations, same patterns showing up in skip-levels. The program landed by every measure that gets reported. It moved nothing.
If you've been in L&D for more than two years, you've watched this happen. A program runs well, ends well, and changes nothing on the ground.
Most of the cultural decisions had already been made before HR posted your req. That's the part of the job nobody puts in the request upfront.
TL;DR
- Most companies hire L&D after the cultural patterns have already hardened. Founders set how feedback flows, who gets promoted, and what "good" looks like long before HR posts the req
- The manager coaching program you've been asked to build is usually a confession that the system doesn't reward the behavior the program is supposed to teach
- Steve Blank called this "organizational debt." Like technical debt, you can't fix it with a workshop. Bob Sutton's work on Uber shows it takes structural change to pay it down
- Your real job is diagnosing the system, not designing programs. The first move is figuring out which problems training can actually fix and which ones it can't
- Pushing back on training requests takes political capital you may not have yet. There's a quieter version of the move you can run on day one
You're inheriting a culture, not building one
Edgar Schein spent decades studying how culture forms inside organizations. The patterns get set in the founding stage by the founders themselves, and they harden fast.
Schein called the mechanism "anxiety avoidance." A group runs into a problem. Someone handles it a certain way. It works, or appears to. The next time the same problem shows up, the group reaches for the same move. Over enough cycles, the move becomes the way things are done around here. Nobody decided it. Nobody wrote it down. It became invisible because it stopped getting questioned.
By the time a company is large enough to need a dedicated L&D function, those patterns have calcified into the operating system.
- How feedback gets given
- Who gets promoted
- What people learn to hide
- What "good" looks like in performance reviews
None of those are training questions, and all of them were answered before you got the call.
This is the mental shift the job requires. You're arriving at an existing system that already decided how feedback flows, how people get promoted, and what behaviors get rewarded. The work is figuring out what you inherited before you start trying to change it.
Most L&D requests don't account for this. They get written as if the company is a clean container waiting for the right curriculum. Build a coaching program. Roll out a feedback framework. Launch a manager development series. The implicit assumption is that the company needs these things added to it. The company isn't a clean container, though. It has a working operating system, and that operating system has opinions about whatever you try to install on top of it.
That's why the first ninety days of the job matter more than the next nine months. You either spend them mapping what's already there, or you spend the rest of the year wondering why your programs keep washing out.
The training request is usually a confession in disguise
Watch what happens when leadership asks for a training program.
The CEO has been hearing complaints that managers don't give enough feedback. Or that hard conversations get avoided. Or that promotions feel arbitrary. Something is wrong, they want it fixed, and they reach for the lever they know how to pull. "We need to train our managers on coaching."
The brief request on your desk. It looks like a training request. It reads like one. The person who wrote it believes it is one.
It isn't.
Read it again. What is the CEO admitting? They're saying their managers don't do something the company says it values. Why don't they? Probably because nothing in the manager's day rewards them for doing it:
- The performance review doesn't ask. Their evaluations track numbers hit, not whether their reports felt heard last quarter. So the time spent on coaching shows up as time taken away from the things that get measured.
- The calendar doesn't allow it. Their days are packed with operational check-ins, status meetings, and escalations. There's no slot left for the slow work of developing someone.
- The bonus doesn't reward it. Variable comp ties to what closed this quarter, not the manager move that paid off two quarters later. Coaching feels expensive in the moment and invisible at payout.
The system is producing exactly what it was designed to produce, and the training request is a confession that the design has a gap.
Companies spend roughly $356 billion a year on training globally, and most of it doesn't change behavior because organizational design fixes must come first, according to Harvard Business Review's research. When the systems and incentives stay the same, the trained behavior runs into the unchanged system and loses every time.
Every request gets read twice. Once, as a training request, the way leadership wrote it. Once, as a confession, the way you have to read it. What is the org admitting it can't fix structurally? That is the actual question on the table.
You inherited a debt. Training won't pay it down
Steve Blank coined the term organizational debt. The idea was borrowed from technical debt, the engineering concept in which teams cut corners to ship faster and end up with code that costs more to maintain than it would have cost to write the first time properly. Blank applied the same logic to people's decisions:
- The compromises companies make in the early days
- The hires they should have made differently
- The conflicts they avoided
- The values they wrote on the wall, but didn't enforce
Those are debts. They accumulate and become harder to repay over time.
Bob Sutton at Stanford took the concept further. Sutton and his co-author Huggy Rao wrote a long case study on Uber as the canonical example of cultural debt at scale. Uber's "always be hustling" culture wasn't an accident. It was a choice that drove fast growth and accumulated debt simultaneously. When the bill came due, paying it down required Travis Kalanick being pushed out, leadership being restructured, the values being rewritten, and Frances Frei being brought in to lead the cultural rehabilitation.
Frei launched something called Uber University. It was a real, well-funded learning program. She left in under a year. The debt wasn't repaid by training. It was repaid by the structural changes that occurred around the training.
Culture is downstream of structure. When a company realigns its incentives, decision rights, and promotion criteria, its culture follows. When it doesn't, no amount of training will pull culture in a new direction, because the levers people are actually responding to didn't move.
Doesn't sustained training investment shift culture over the years? Sort of. Training can amplify a structural change after the structure has shifted. It can reinforce new behaviors that the system has started rewarding. What it can't do is cause the shift on its own. Pulling the training lever and waiting for culture to follow is the wrong physics. Cultural debt is principal. Training is an interest payment, at best. At worst, it's a theater that lets leadership feel productive while the principal keeps growing.
Diagnose first, design second
The discipline you're being asked to do already exists. It doesn't usually live in L&D. Organizational development practitioners have been conducting formal diagnoses for 50 years, and the frameworks are well established. Weisbord's Six-Box Model. McKinsey's 7S. Cameron and Quinn's Competing Values Framework. The one most useful for the L&D job is Burke-Litwin.
Burke-Litwin splits an organization into two layers. The transformational layer holds the slow-moving stuff: mission, leadership, culture. The transactional layer holds the day-to-day: systems, practices, climate. Training, almost by definition, operates in the transactional layer. Cultural problems live in the transformational one. When you try to influence the first layer by acting on the second, the system absorbs the change without moving.
That split is the most useful thing you can hold in your head when walking into a new L&D role. Every request that hits your desk lives in one layer or the other. Requests about skills, processes, and tools belong in the transactional layer, where training can actually help. Requests about feedback culture, accountability, and "the way we work together" live in the transformational layer, where training is the wrong tool.
Michael Watkins, author of The First 90 Days, recommends a set of diagnostic questions for any new leader stepping into a role.
- What behaviors is leadership asking for that the system doesn't reward?
- Where do managers get praised for the opposite of the behavior we're being asked to teach?
- What was the founder's original anxiety pattern, and how is it still showing up in how decisions get made?
- When was the last time someone got promoted for the behavior we're being asked to train?
The L&D team at Noom, which built the function from scratch for a company of 3,000-plus employees, described the shift this way. Their conversation moved from "what can we build to teach learners about this skill?" to "how can we partner with leaders and employees to grow in the identified areas?" That's the diagnostic without the jargon. The work changed before the deliverables did.
Saying it out loud is the harder part
Diagnosis is the inside-your-head version of the move. At some point, you have to share what you found with leadership. That conversation is where credibility gets built, and where it gets risked.
A first L&D hire at a 600-person company, reporting two layers down through HR Operations, has a different kind of conversation than a Senior Director of L&D with five years of trust at a 5,000-person company reporting to the CHRO. The advice that works for the second person can get the first person fired. Both versions of the move need to exist.
Both versions are doing the same job. One has permission to say it out loud. The other has to do it through the work itself. Either way, the reframe is the same. You are an organizational diagnostician who happens to design learning when learning is the right tool, and the order of operations is diagnosis first, design second.
That sentence is the entire job in a single line. Some L&D pros will read it and feel relief. Others will read it and feel the floor shift. Both reactions are right. The reframe asks you to put down the tool you were trained to use first and pick up another.
The work will tell you what the request didn't
The cultural debt you inherited isn't going anywhere. It was there before you got the offer letter, and it'll be there after you accept the next brief. The only thing you control is the first move you make.
You can spend the next year building programs that bounce off it. Or you can start by reading what you inherited and getting honest with leadership about what training can and can't fix. None of that was in the job description. The work will say it anyway.
Frequently asked questions
Q: What is cultural debt in an L&D context?
A: Cultural debt is the people and culture version of technical debt. It's the accumulated cost of compromises early-stage companies make to grow fast, like avoiding hard conversations, hiring the wrong managers, or writing values that nobody enforces. By the time L&D arrives, those compromises have hardened into patterns that training alone can't fix.
Q: How is cultural debt different from a skills gap?
A: A skills gap is something training can close. Cultural debt is something training runs into. If your managers know how to give feedback but the system rewards them for not doing it, that's debt, not a gap. The test is whether the behavior would persist if the training were perfect. If yes, the problem isn't curricular.
Q: Why do most manager training programs fail to change behavior?
A: They run into the unchanged system. HBR research found that most of the $356 billion spent globally on training doesn't shift behavior because the organizational design fixes have to come first. When incentives, decision rights, and promotion criteria stay the same, trained behavior loses to the system that hasn't changed.
Q: What should L&D actually do in the first ninety days at a new company?
A: Diagnose before designing. Read every training brief twice, once as a request and once as a confession. Map what was decided about feedback, promotion, and accountability before you arrived. Borrow from formal OD frameworks like Burke-Litwin to separate transformational issues (where training can't help) from transactional ones (where it can).
Q: What's the difference between a needs assessment and organizational diagnosis?
A: A needs assessment asks what skills people lack. Organizational diagnosis asks why those skills aren't showing up in behavior even when people have them. Diagnosis is a fifty-year-old OD discipline with formal frameworks. Needs assessment is the watered-down cousin that L&D inherited and stripped of its rigor. Reclaiming the discipline gives L&D pros a vocabulary for the work they should already be doing.
Q: How do you push back on a training brief without losing credibility?
A: It depends on how much capital you have. If you're early in a role, deliver the program leadership asked for, but pair it with one structural recommendation in the post-program report. If you have executive trust, name the cultural debt directly and offer to lead the diagnostic work. Both are the same move. One has permission to say it out loud.
