Working the Callout

The previous fifteen chapters assumed you own the machine. This one and the next assume you do not.

You have arrived at somebody else's site, on somebody else's rules, to fix somebody else's problem, and you will leave at the end of it. The person standing next to you is a customer: the relationship is contractual, your findings have commercial consequences for them, and their opinion of the visit is formed by things that have nothing to do with whether you were right.

If you work in-house, almost all of this still applies — swap "customer" for "production" and "site" for "your own plant". The parts that genuinely differ are named where they come up.

The claim

The person at the machine is a diagnostic instrument, and like every instrument in Chapter 4, it has characteristics.

Chapter 15's method opens with defining the symptom, and says that the operator's phrasing is frequently the diagnosis. That is only true if they tell you. You cannot get a useful account out of somebody who feels dismissed, blamed, or talked over — you get a short, defensive, sanitized version, and the detail that would have saved you two hours stays unsaid.

So the content of this chapter is not manners. It is the technique for getting a clean reading off the least reliable and most informative instrument on the job.

Before you arrive: early is information, late is a fault

The thing to understand is that they are not managing your arrival. They are managing their own decisions, and every one of those needs an estimate:

• Do we send the shift home or hold them on the clock?
• Do we start the other line, and lose the changeover if you fix this one?
• Do we call a second contractor?
• Do we tell our own customer the order is late?

Every hour they spend without an estimate is an hour of making those calls blind. That is why a two-hour delay you tell them about at the start costs less than a twenty-minute delay they discover when you fail to appear.

Practically:

• Give an ETA you can beat, and say what it is based on.
• Update the moment it changes, not when you arrive. The update that matters is the one before the deadline you are about to miss.
• Say what you know before you get there. "Bring the machine to a safe state and don't run it" is worth a phone call. So is "have the fault log and the last service report to hand".

The first ninety seconds

Somebody who has been down four hours has already been on the phone to their manager, has product backing up, and has spent the morning being asked when it will be fixed by people who cannot help. Their opening is not a fault report. It is pressure coming off.

Let it run. Do not defend the response time, do not explain the contract, do not start diagnosing, and do not interrupt to correct a technical detail. Ninety seconds of listening costs you ninety seconds. Interrupting costs you the rest of the visit, because you will have taught them that talking to you is not useful.

Then acknowledge the impact before you go anywhere near the machine. "Four hours is a lot of product" is free, it is true, and it changes the next twenty minutes.

Three things not to do while listening:

• Do not agree with criticism of a colleague or your own company. It feels like solidarity and it is a commitment you cannot honor.
• Do not promise anything. Not a time, not a cause, not a cost. You have not seen the machine.
• Do not match their tone. Somebody who is angry is looking for a temperature to settle at, and they will settle at yours.

Listening as a measurement

Once they have finished, ask the Chapter 15 questions: what stopped, when, under what conditions, how often, what changed. Then one more that field work adds:

"What have you already tried?"

This is free elimination, and it is routinely thrown away because the answer is sometimes wrong or naive. Take it anyway. A fault that survived a component swap was never in that component — and the person who feels their attempt was dismissed will not tell you about the second thing they tried, which is frequently the one that matters.

While you are there: the update nobody asked for

The highest-leverage habit in this chapter. Every thirty to forty-five minutes, and immediately whenever the picture changes, tell them:

1. What you have ruled out — this is the part that demonstrates progress when you have not found it yet.
2. What you are doing next.
3. Your current estimate, and explicitly whether it has changed.

"I don't know yet" is a strong answer when it comes with what you have eliminated. On its own it is not an answer at all. The difference is the whole of Chapter 15 step 6, said out loud to somebody who is not a technician.

Delivering bad news

The part is not on the van. It needs a specialist. It will be tomorrow. It is not covered by the contract. It is not the machine — it is the process, or the consumable, or the way it is being run.

The structure that works:

• Say it early. Bad news does not improve with age, and the anger is proportional to how late it arrives, not to how bad it is.
• Say it plainly, once, without cushioning it into ambiguity.
• Say what you have done, so they know what they are getting for the visit.
• Give the options, including the ones you would not choose.
• Say what you recommend, and why. They are entitled to an opinion from the person who has seen it.

The hardest version is when the cause is something they did. Describe the mechanism, not the culpability: "the filter has been in for four months on a three-month schedule, and that is what blocked the silencer" lands very differently from the same fact framed as a failure to maintain. You are not softening it — the mechanism is the useful part, and blame is the part that stops them telling you things next time.

Saying no on somebody else's site

"While you're here, could you just…" is the commonest way a field technician ends up doing work that is outside their scope, outside their competence, outside their authorization, and outside anybody's insurance.

Saying no is harder here than in-house, because the pressure is commercial and you are a guest. Chapter 1's line has not moved — safety is not negotiable and neither is competence — but it now has to be said to somebody who is paying.

What works is to be specific and to offer the route: "That's not something I'm signed off for, and doing it would invalidate the certification. What I can do is raise it so somebody who is comes out." You have refused, you have explained why in a way that is not about you, and you have left them with a next step.

Leaving

Do not disappear. Before you go, walk the person who will still be there after you leave through: what was wrong, what you did, what to watch for, and what is still outstanding. Not the person who signed you in — the person who runs the machine on nights.

If something is unresolved, say so explicitly and say what happens next. A customer who knows a part is on order is managing a plan. A customer who thinks it is fixed is about to be surprised.

The callout, from dispatch to departure.

1. Before you set off, find out what you can and say when you will arrive

   The fault description, what they have tried, whether the machine is safe, and what history exists. Then give an ETA you can beat and say what it depends on. If anything about it changes, the update goes out immediately — not on arrival.

2. Sign in properly and learn the site's rules

   Safety

   Induction, permits, PPE requirements, escort arrangements, and — from Chapter 1 — the emergency number, the first aiders, and which gate an ambulance would use. Thirty seconds at sign-in, and you are a guest who does not know where anything is.

3. Let the first ninety seconds run

   Do not defend, do not diagnose, do not correct, do not promise. Then acknowledge the impact before you look at the machine. This is the cheapest and highest-return ninety seconds of the visit.

4. Ask the Chapter 15 questions, and then what they have tried

   What stopped, when, under what conditions, how often, what changed — and what has already been done to it. Take the last answer seriously even when it is naive, because it is free elimination and because dismissing it closes the source.

5. Establish what has been done to the machine before you touch it

   Safety

   Guards, interlocks, jumpers, half-applied isolations, parts swapped. Ask it in a way that cannot be heard as an accusation, because on an unfamiliar site this is your only route to knowing.

6. Work the method — and keep the hypotheses to yourself

   Chapters 1 to 15 apply unchanged. What changes is that you are being watched: an early guess said out loud becomes a commitment, and revising it reads as being wrong twice rather than as narrowing the search.

7. Update every thirty to forty-five minutes, unasked

   What you have ruled out, what you are doing next, and whether your estimate has changed. The eliminations are what demonstrate progress before you have an answer.

8. Deliver bad news as soon as you have it

   Early, plainly, with what you have done, the options, and your recommendation. If the cause is something they did, describe the mechanism rather than the fault of it — you lose nothing accurate and you keep the source open.

9. Hand over to the person who will still be here

   What was wrong, what you did, what to watch for, what is outstanding and what happens next. Then say you are leaving. A customer who knows a part is on order has a plan; one who thinks it is fixed is about to be surprised.

• Going quiet when things are going badly

  Why

  This is the one that costs most, and it is entirely counter-intuitive from the inside. When the job is going well, updates feel easy. When you are two hours in with nothing, calling to say so feels like advertising failure — so the silence lengthens exactly when the customer needs information most. What you have ruled out is progress, and saying it is the difference between a technician working a hard fault and one who has vanished.

• Interrupting to correct a technical detail

  Why

  They have said something inaccurate about the machine and the correction is right there. It is also worth nothing — you were going to check it anyway — and it costs you the rest of what they were about to say. Correct it later, or never.

• Thinking out loud

  Why

  Talking through your reasoning feels transparent and collaborative, and to a technician it would be. To somebody who is not one, every hypothesis sounds like a conclusion, so narrowing the search reads as a sequence of wrong answers. Share findings, not candidates.

• Dismissing what they already tried

  Why

  It is often naive, occasionally wrong, and rarely the answer — so it gets waved away in the first minute. But it is free elimination, sometimes it contains the detail that matters, and waving it away teaches them not to offer the next one.

• Softening bad news until it is ambiguous

  Why

  Nobody wants to say the part is a week away, so it becomes "it might be a few days, I'll chase it" — which is kinder in the moment and leaves them planning against a date that was never real. The anger when it arrives is about the ambiguity, not the week.

• Saying yes to the extra job

  Why

  It takes ten minutes, they are asking nicely, and refusing feels petty over something small. But scope, competence, authorization and insurance are not about the size of the task, and the ten-minute favor is the one that is uncovered when it goes wrong.

Six questions. Each is about the mechanism — why a communication choice changes what you learn or what it costs — rather than about being pleasant.