The Diagnostic Value of Asking What Changed Recently

Why this matters

Equipment that ran fine for years and suddenly does not almost always has a trigger, and that trigger is almost always something that changed recently, even when the customer insists nothing did. Skipping this question sends you straight into a generic fault tree, checking every plausible cause in order, when a thirty-second answer could have told you exactly where to start. This is the single highest-leverage question in diagnostic work precisely because most techs under-ask it or accept the first "no" too easily.

Why "nothing changed" usually means "nothing I noticed as related"

Customers filter their own answer before they give it to you. They mentally exclude anything they do not believe is connected to the fault, which means the most useful answer is often the one they left out because it "obviously" has nothing to do with it. A new appliance on the same circuit, a renovation two rooms away, a different detergent, a landscaping change that altered drainage, a different thermostat schedule, none of these feel equipment-related to a customer, so they do not volunteer them.

Ask the question more than once, from different angles, before accepting "nothing" as the real answer. See related: reading what the customer didn't say.

The categories of "what changed" worth asking about directly

• Physical changes nearby. Renovation, new construction, new flooring, new furniture blocking a vent or access panel, landscaping or grading changes that affect drainage or airflow.
• New equipment or devices sharing infrastructure. A new appliance on the same circuit or line, a new fixture added to the same supply, anything that increases load or demand on a shared system.
• Changes in usage pattern. More people in the home, a new tenant, a change in schedule (working from home now, a new occupant's habits), increased frequency of use.
• Prior service or DIY activity. A different company's visit, a homeowner's own attempt at a fix or adjustment, even something that seems unrelated like changing a filter or resetting a breaker. See related: the misuse-as-education angle.
• Environmental or seasonal shifts. The transition into a new season is itself a "change," even without anyone touching anything; equipment that idled through one condition can fault the first time it is asked to work under a new one. See related: seasonal-transition technical faults.
• Upstream system changes. Something replaced or serviced elsewhere on the same property that affects supply, pressure, voltage, or signal to the equipment you are diagnosing, even if it seems unrelated at first glance.

How to ask it so you get a real answer

Open-ended and specific beats generic. "Has anything changed?" invites a reflexive "no." Better:

• "In the days or weeks before this started, was anyone working on anything else in the house or building, for any reason?"
• "Did you or anyone else touch this equipment, or anything connected to it, before this started?"
• "Was there any construction, remodeling, or new installation nearby recently, even if it seems unrelated?"

If the first answer is "no" to all of these, ask about timing instead: "did this start suddenly, or gradually?" A sudden onset with no acknowledged trigger is worth a second pass at the same questions; something changed, the customer just has not connected it yet.

What a confirmed trigger buys you

Once you have a real "what changed" answer, you can often skip most of the generic fault tree and go straight to testing the specific hypothesis that trigger implies. A renovation nearby means check for physical disturbance to lines, wiring, or airflow paths in that specific area first. A new appliance on the same circuit means check load and capacity before assuming the equipment itself failed. A different service company's visit means check what they touched before assuming an unrelated coincidental failure.

When there really was no trigger

Sometimes there genuinely was no change, and the fault is straightforward age-related wear or a random component failure. That is a legitimate answer too, do not manufacture a trigger where none exists just because this question usually finds one. If a thorough round of asking turns up nothing, move on to the standard fault tree with confidence that you have not skipped the highest-value question first.

References

• Trade-standard practice for structured root-cause interviewing
• See related: The Diagnostic Question That Narrows It Fastest, Reading What the Customer Didn't Say, Retrofit and Renovation-Induced Faults