Most L&D triage runs one test — can’t versus won’t. Here is the second test it misses, the four situations the two produce, and what to sell in each.
Every experienced L&D person has the same reflex when a leadership training request lands. You ask some version of: could they do this already if it really mattered?
It is a good test and it has been the profession’s main tool for decades. Robert Mager built a career on it. The Learning Guild’s obituary of him quotes the version he was known for — “If you put a gun to their head, could they do it?” — and I should say plainly that I am quoting the obituary here, not his book, because I have not had the book open in front of me while writing this (Learning Guild).
The test sorts one thing, cleanly: whether the constraint is inside the person or around them. If they could do it under pressure, the capability exists, and building capability will not change anything.
Here is the problem. That test tells you whether it is a training problem. It does not tell you whether it is a solvable problem. And those are different questions, with different answers, different interventions and different invoices.
Which means the honest triage has two tests, not one.
(If you have not yet had the conversation that produces the evidence for either test, the discovery-call piece covers that part. This piece assumes you already have the answers and are deciding what to do with them.)
The familiar one. Ask whether anyone in the same population does this well under the same conditions. If yes, capability exists somewhere in the room, and the difference between them and everybody else is not knowledge.
Then ask what happens to a person who does the thing you are being asked to train. If deciding gets you overruled, if raising a risk gets you a reputation, if delegating means you carry the failure — the behaviour is being priced correctly by the people not doing it. They are not undertrained. They are accurate.
That is test one. Nothing new so far; it is well covered.
This is the test almost nobody runs on a training request, and it changes the answer more often than the first one does.
Ronald Heifetz and Donald Laurie put the distinction in the form of a question a leader has to answer about their own situation: “Would expert advice and technical adjustments within basic routines suffice, or would people throughout the company have to learn different ways of doing business, develop new competencies, and begin to work collectively?” Adaptive work, they write, “is required when our deeply held beliefs are challenged, when the values that made us successful become less relevant, and when legitimate yet competing perspectives emerge” (Harvard Business Review, December 2001, reprint r0111k — full reprint).
Translated into the language of a brief: a technical problem is one where somebody, somewhere, already knows the answer and it just has to be applied here. An adaptive problem is one where nobody knows the answer yet, and finding it requires somebody to give something up.
Note what that is not. Adaptive does not mean hard, or cultural, or soft. Rewriting an incentive scheme is hard and expensive and entirely technical — the mechanics are known, somebody has done it before, and it can be implemented by authority. Asking two regional MDs to accept a shared P&L when their bonuses currently depend on separate ones is adaptive, no matter how small the change looks on paper, because it requires them to lose something.
The test I actually use: name the specific person who would have to accept a loss for this to work, and name what they lose. If you can do that in one sentence and the loss is real, it is adaptive. If you cannot name anyone, it is technical.
Heifetz and Laurie also name the failure mode that follows from getting this wrong, and it is the single most useful sentence in the article for anyone selling programmes. Leaders, they write, keep “providing leadership in the form of solutions” — which “restores equilibrium in the short term but ultimately leads to complacency and habits of work avoidance that shield people from responsibility, pain, and the need to change.”
A training programme commissioned in response to an adaptive challenge is that sentence, in product form. It is a solution provided by authority. It restores equilibrium — everyone is doing something, there is a plan, the calendar has cohorts in it — and it shields the people who would have to change from having to. That is why these programmes evaluate well and produce nothing. They are working exactly as commissioned.
Cross the two tests and you get four cells. They are not four flavours of the same job. They have different buyers, different units of intervention, different deliverables and different prices.
The two tests, crossed
One test gives you two answers. Two tests give you four — and four different things to invoice. Cell 1 is where you want the answer to be, because it is the biggest engagement. Be suspicious of how quickly you got there.
The people genuinely cannot do it, and the way to teach it is known.
The rarest of the four in leadership work, and the only one where the request as stated is correct. It is common in the sub-parts of a leadership brief: running a structured one-to-one, writing a development objective, reading a P&L.
Unit: the individual. Deliverable: the programme they asked for. What you do next: pick the frame the symptom actually points to — a separate decision with its own failure modes — and build.
The trap: cell 1 is where you want the answer to be, because it is the biggest invoice. Notice that, and be suspicious of how quickly you got there.
They can do it. Something around them makes it a bad idea. And the fix is a thing somebody has authority to change: decision rights, an approval threshold, a metric, a manager’s review behaviour, a missing job aid, a broken handover.
Unit: the system, usually one or two named mechanisms. Deliverable: a performance analysis that names the mechanism and the evidence for it, a design for the change, and — often — a small enablement piece so the people affected know what changed and why. Not a six-month cohort programme.
This is the cell most often mis-sold as cell 1, because the client asked for training and training is easier to buy than a change to someone’s decision rights.
The skill is genuinely absent and using it costs somebody something. A manager who has never given hard feedback, in an organisation that has always prized harmony. The skill is teachable. The practice of it is a loss for someone.
This cell is the most commonly botched, because it looks exactly like cell 1 in a scoping call. Train it as cell 1 and you produce forty managers who can now do a thing they will not do, and a Level 2 result you will be quite proud of.
Unit: the individual for the skill, the group for the adaptive half. Deliverable: a deliberately smaller skills core, plus structured work with the group on what the new behaviour costs and who absorbs it. Sequence matters: the adaptive half does not follow the skills half as a nice-to-have at the end. It runs alongside, or the skills half has nowhere to land.
They can do it, something stops them, and the something is a trade-off the organisation has not decided to make. Two regions with separate P&Ls. A leadership team that wants candour and promotes on loyalty. A sponsor who wants his managers to decide and has not decided to live with their decisions.
No training touches this. Heifetz and Laurie are explicit about where the answer lives: “Solutions to adaptive challenges reside not in the executive suite but in the collective intelligence of employees at all levels, who need to use one another as resources”, and “the locus of responsibility for problem solving when a company faces an adaptive challenge must shift to its people.”
Unit: the group holding the competing perspectives — which usually includes the person who commissioned you. Deliverable: a diagnostic engagement and one or two facilitated sessions in which the trade-off is named and somebody decides. Small, short, and frequently the most valuable work you will do for that client.
Every article on this subject ends by telling you to say “it isn’t a training problem.” None of them mentions that saying so converts a six-figure programme into a conversation, and that you have a mortgage.
So here is the part that gets left out.
Cells 2, 3 and 4 are not smaller versions of cell 1. They are different products, and they are priceable. A performance analysis with a named mechanism and an evidence trail is a deliverable. A designed change to decision rights is a deliverable. Two facilitated sessions with a leadership team, with a written record of the trade-off they agreed to, is a deliverable — and it is the one a CEO will remember a year later.
Three things make this survivable in practice:
1. Never arrive at the smaller recommendation without the alternative already attached. “Training won’t fix this” leaves the client with a problem and no plan, and it reads as you talking yourself out of work. “I’d do about half of what you’re describing, plus one other thing, and the second one is what makes the first one hold” is a proposal. Same diagnosis, entirely different meeting.
2. Price the diagnosis separately, and early. If the analysis is bundled free into a programme, you cannot afford for the answer to be “no programme.” Charging for the diagnosis is not a billing preference — it is the thing that makes an honest answer economically possible. Clients accept it more readily than most consultants expect, because it is how every other professional service they buy already works.
3. Take the small piece of work. Cell 4 engagements are short and they are how you become the person who gets called before the brief is written next time — which is the only structural fix for receiving pre-diagnosed briefs.
One page. It goes into the proposal, not into your own files. It exists so that in month nine, when someone asks why the programme did not fix the thing it was never designed to fix, the answer is a document rather than a memory.
1. What was requested. In the client’s words, verbatim. Not your reframing of it.
2. The stated cause. What the client believes is producing the behaviour.
3. Test one — capability or consequence. Who does this well under the same conditions, and the client’s own explanation of why. If nobody does, say so; that is evidence too.
4. Test two — technical or adaptive. The named person who has to accept a loss, and what the loss is. “None identified” is a valid and meaningful answer.
5. What I looked for that would have proved me wrong. The single most important line. Name the evidence that would have supported the client’s original diagnosis, and whether you found it.
6. The cell, and therefore the intervention. One sentence.
7. What we are not doing, and why. The scope you declined, in writing, with the reason.
8. The stated assumption. The barrier nobody in this engagement can move, and how it will be reported.
Line 5 is the one that makes this an honest instrument rather than a rationalisation. Anybody can build a case for the diagnosis they arrived at in the first ten minutes. Writing down what would have falsified it — before you go looking — is the difference between a diagnosis and a preference.
Line 7 is the one that saves the engagement. Declined scope that was never written down becomes, nine months later, scope you failed to deliver.
The fashionable version of this argument says the environment is almost always the real cause, and training is almost always the wrong answer. It is a satisfying thing to say and it is not what the strongest evidence supports.
Blume, Ford, Baldwin and Huang’s meta-analytic review of training transfer (Journal of Management, 36(4), 2010, pp. 1065–1105) reports mean correlations with transfer of .27 for transfer climate (k = 8, N = 870) and .21 for support (k = 12, N = 1,075) — with the support figure splitting into .31 for supervisor support and .14 for peer support, though the paper does not give separate k values alongside that split in the section I read. So the environment matters, measurably, and supervisor behaviour matters considerably more than peer behaviour, which is worth knowing when you are deciding whose behaviour to design for.
But the largest mean correlation among the predictors in that same set is .37 for cognitive ability (k = 10, N = 1,452) — a trainee characteristic, not an environmental one. And the authors state the position plainly on page 1092: “there is no clear superiority of individual variables over situational variables, or the reverse.” (full text)
Which is the whole argument for running a triage instead of holding a position. “It’s never a training problem” is as lazy as “let’s run a leadership programme.” It just sounds more sophisticated at a conference. The evidence supports diagnosis, not a house view — and if you have a standing answer before you have the evidence, you are the thing you think you are protecting the client from.
I will also say, because it belongs here: I think most leadership training requests I have received were misdiagnosed at the point they reached me. That is my experience across a working life in this job, not a finding. I have no study to put behind it and I am not going to dress it up as one.
The two tests are not independent, and pretending they are will mislead you. A consequence problem often exists because somebody would lose something by removing it — that is a cell 4 in disguise wearing cell 2’s clothes. If the “known technical fix” has been available for three years and nobody has applied it, stop calling it technical. Something is holding it in place, and that something is the actual brief.
The adaptive test can be used as an excuse. “It’s adaptive” is available to any consultant who does not want to build the thing. It is only a real finding if you can name the person and the loss. If you cannot fill in that sentence, you have a hunch, and you should say so.
Four cells is a simplification and a real brief usually contains three of them. A first-time-manager programme is typically a small genuine capability gap (cell 1), a large consequence problem around escalation (cell 2), and one adaptive question about what the sponsor is willing to tolerate (cell 4). The point of the triage is not to pick one cell. It is to size each part honestly, so the design puts the budget where the constraint actually is — and so that the part of it that is genuinely training is not asked to carry the parts that are not.
Try it on a live brief. The Diagnostic Conversation Coach is free, needs no signup, and runs both tests against a gap you describe in your own words — including the disconfirming evidence you should be looking for and what each answer commits you to. If it does not sharpen the way you run a discovery call, nothing else here will interest you.