The solution request

Nobody calls the fire department to report a shortage of water.

“We have a water problem”

When a building is burning, nobody telephones the fire department to report a shortage of water. The problem is the fire. Whether water is the right answer depends on what is burning, and in a chemical plant it is emphatically not.

The second example is closer to home and slightly funnier. “We have a pesticide problem.” No: there are unwanted insects. Pesticide is one candidate answer. A screen on the window is another, and it costs less, works faster and does not have to be repeated every spring.

Both examples are Jim Fuller’s, from the book he wrote with Jeanne Farrington on the move from training to performance improvement (Fuller & Farrington, 1999). They have been in this profession’s teaching for a quarter of a century and they still land, which is itself a finding about how little the habit has shifted.

Requests inside organizations arrive in exactly this form. Not “sales per visit have fallen eleven percent since March”, but “we need customer-experience training”. Not “three of the last nine handoffs went wrong”, but “the team needs to communicate better”. The solution is in the sentence and the problem is not. Whoever accepts the sentence as written has committed to a fix before anybody established what it was fixing.

Dana and James Robinson named this conversation and the role that has to hold it. Their point was that a request of this form is one of only two ways work reaches a practitioner, the other being a business result that is already moving, and that the whole craft consists of getting from the first to the second (Robinson & Robinson, 1995). Number 02 places them in the lineage.

This is not carelessness, and treating it as carelessness is the fastest way to lose the conversation. It is how competent people behave under time pressure. A manager who has watched a problem for six weeks has formed a theory, and the theory arrives dressed as a request because that is the only form the organization reliably accepts. Ask for an analysis and you get a raised eyebrow. Ask for a course and you get a budget code.

Everyone diagnoses in their own vocabulary

Roughly a tenth of an iceberg sits above the water. Managers and advisors alike work from what they can see there, and what they can see is shaped by what they already know how to do.

The pattern is uncomfortable because it is so regular. A coach walks into a struggling team and perceives a coaching problem. An interim manager perceives a structural one. A process consultant sees a process. A learning specialist sees a skills gap. Each is sincere, each is experienced, and each has diagnosed the case in the vocabulary of their own toolkit.

Harless put the mechanism plainly forty years ago, and included his own profession in it: what gets rewarded is the last event in the chain, which is choosing a solution, so the role predicts the remedy, and the consultant’s most probable solution is whatever the consultant is selling at the time (Harless, 1985, p. 4).

The practical correction is a rule about evidence rather than a rule about skepticism. Perception is a hypothesis until it has been tested against something the organization already counts. Below the waterline sit the numbers (output, rework, cycle time, complaints, absence, retention) and they do not care what anybody sells.

Demanding the result, choosing against it

There is a specific trap underneath almost every request of this kind, and it has a name in this practice: the performance paradox. A manager demands a better result and chooses an intervention that cannot reach the cause of the shortfall, and then expects the result anyway.

Stated that plainly it sounds like carelessness, and it is not. It is what happens when four ordinary pressures meet in one person, and each of the four is reasonable on its own.

What the paradox costs is worth spelling out, because it is not only the money. The time and the fees are spent. The manager who commissioned it becomes the person who commissioned it. Skepticism about training hardens on the floor, and the next proposal, which may be entirely sound, meets a room that has been here before. A sense of déjà vu sets in that makes the good case harder to win than the bad one was.

Why the requester cannot do this for you

The obvious objection is that the manager knows their own operation far better than any advisor ever will. That is true, and it does not produce the conclusion it seems to.

There is a durable finding in psychology that people have limited access to the processes behind their own judgements, and that when asked to explain them they report plausible reasons rather than the operative ones (Nisbett & Wilson, 1977). Number 07 uses the same finding for a different purpose: it is why an exemplary performer cannot simply tell you what makes them exemplary. Here it cuts in the other direction. Somebody who has watched a problem daily for six weeks has an account of its cause that feels like observation and is partly construction.

Put the two together and the position is clear. The person best placed to notice the problem is not thereby well placed to diagnose it, and the closer they are, the more confident and the more coherent the account will be. Confidence is a property of the narrator, not evidence about the cause, and the four pressures above are all pushing the same account towards the same remedy.

One further point about who tests. This is the reason a request should never be tested only by the function that would deliver the solution. Not because that function is untrustworthy, but because the same limitation applies to it, in the same way, about its own reasoning, and because it is subject to the second of the four pressures more than anybody.

No number, no case

Before any solution can move a business result, three elements have to be on the table: a gap stated as the distance between current and desired performance, evidence that the gap matters to the organization, and a cause established below the symptom. Most requests satisfy none of the three, and still sound entirely reasonable.

Read the middle column again. Every entry is plausible and every entry is untested. “People feel undervalued” may be true and still be a consequence rather than a cause. Absence and turnover in the fourth row are not causes of workload; they are what high workload leaves behind.

What a number actually does

The absence of a number is not a formality. It changes what kind of object the request is, and therefore what can ever be said about it afterwards.

“Repeat orders are at 62 percent and the business needs 75” is a case. You can tell whether it is worth solving, you can tell where the shortfall sits, and afterwards you can tell whether you solved it. “They need to be more customer-focused” is a wish, and a wish cannot be evaluated, which is precisely why programs built on wishes are never quite finished and never quite fail.

Where no number exists, the emptiness is itself the first finding: the same finding number 05 reaches from the other door. A result nobody counts is a result nobody is accountable for, and that is worth reporting before any discussion of remedies. It is also, frequently, the cheapest thing this whole method ever produces.

Four questions that test a request

Twenty minutes, in the first conversation, before any commitment. The order matters: each question is answerable, none of them is hostile, and together they establish whether there is a case at all.

Note what is not in the list. There is no question about whether the requested solution is a good one, and no question that requires the requester to defend their reasoning. The four questions are about the work, not about the request, which is why they can be asked by somebody junior, of somebody senior, without the conversation turning.

Three endings, and only one of them is “no”

The fear that stops this method being used is that testing a request means refusing it. In practice refusal is the rarest of the three outcomes, and the least useful one to lead with.

Which is the argument to make internally, and it is not the one usually made. Testing requests is not a way of saying no more often. It is a way of making the yes defensible, and of ensuring that when the answer really is a program, it is aimed at something, sized against something, and provable afterwards against the standard in number 11.

The onboarding that was never the problem

A logistics operator, 900 warehouse staff, heavy seasonal hiring. The request: an extended onboarding program, because new starters were making too many picking errors in their first months. Delivered at the rates the operator already paid, that is a substantial program.

The request was accurate about the symptom, and the requester, a site manager with eleven years there, was right about almost everything. Four questions, one afternoon.

What differed about those two aisles was label placement. They had been relabeled during a refit fourteen months earlier, the same fourteen months, and carried the item code at eye height on the bay rather than at ankle height on the shelf lip. Everywhere else, a picker confirming a code had to crouch.

Why the new starters really were worse

The most instructive part of this case is that the requester’s observation was correct and their inference from it was not, and the two are easy to confuse in a meeting.

New starters genuinely were less accurate, and not because they knew less about picking. Staff who had worked the aisles for a year had memorized the layout and no longer needed to read the labels at all, so a label that was hard to read cost them nothing. New starters read every one. The condition was identical for both groups; only one group was exposed to it.

That is worth stating as a general pattern, because it recurs. Where experience conceals a condition, the newest people look like the problem. Every remedy aimed at them will be aimed at the wrong thing, and the measure will improve on its own as each cohort learns to work around the obstacle, which reads, a year later, as proof that the onboarding worked.

The site manager, who had asked for the program, presented the finding to the regional meeting himself. That is the ordinary outcome when a test is run with somebody rather than on them.

What this does not claim

Testing requests is a professional posture with costs of its own, and the costs fall on the person doing the testing.

• It is not a license to refuse work. A function that tests every request and concludes “not training” every time has replaced one reflex with another, and the second is less forgivable because it claims evidence. Section 06 gives the three endings; refusal is the rarest.
• The requester is usually right about the symptom. This paper argues that a request is a hypothesis, not that it is mistaken. In the case above every observation the site manager made was accurate. What did not follow was the inference.
• The evidence about self-report is imported. That literature examined how people account for their own mental processes, largely in laboratory settings, not how managers explain a delivery number. Applying it here is an inference and is stated as one.
• The fire and the iceberg are illustrations, not evidence. Both have been used in this practice for twenty years and neither carries any weight in the argument. Where an image does the work of a finding in somebody’s report, that report has a problem.

Six lines, on the request in front of you

Twenty minutes, with the requester rather than about them. Lines one to four are the test; five and six decide what you do with it.

Two problems, one method, sixteen papers

Two models carry the method. Gilbert’s six conditions ask what kind of cause; Rummler and Brache’s three levels ask where it sits. Both are sound and this series uses both, and it leads with Gilbert, because his unit is the accomplishment, which is what an organization already counts, and because his six conditions are the categories an organization can actually change.

Where the method comes from

Two further pieces close the set and carry no number, because each points at all the numbered papers. Before the guide sits Arets L&D Business Models, the destination piece for the people who receive the requests: four models a learning function can run, and the route from HR-driven to business-driven. Last sits Solutions that cannot work, the field guide, ordered by the sentence that lands on your desk: seventeen forms a performance problem arrives in, each with an address in one of the numbered papers. Three orders in all: the destination you choose, the question you are on, and the form that landed on your desk.

Sixteen papers, six parts

Number 01 is the map and number 02 is the foundation. From number 05 on, every number is depth on exactly one step of the method. The filled block is the paper you are holding; the two unnumbered pieces close the set.

Where every claim in this paper comes from

The other two books named earlier in this paper are listed in full in number 02, so that they do not recur in every reference list. Number 02 also sets out the evidence rule this series follows: empirical claims rest on meta-analyses and systematic reviews, and nothing here is supported by a single study.