It Works for You but Fails for Them: Decision Tree
Why this matters
You run the equipment and it works. The customer runs it and it fails. Both of you are right. The gap is a variable that is present when they use it and absent when you test it, and your job is to find which one. The mistake is assuming it must be operator technique and stopping there, because the difference is just as often the environment at their time of use, the load they put on it, or the state the machine is in when they start. This tree walks the full candidate space so you do not leave with a no-fault-found on a fault that is real every day but yours.
A safety note before you reproduce it
When you deliberately recreate the customer's conditions, you are trying to make a fault happen on purpose. If that fault is a hazard, an overheat, an overpressure, an electrical trip, a gas or water event, set up to observe it safely before you trigger it: know where your shutoff is, keep protection in place, and be ready to kill it fast. Reproduce the condition, not the danger.
Start here: nail down what "works" and "fails" actually mean
Before chasing variables, confirm you and the customer are describing the same event. Have them tell you exactly what failure looks like from their side (what it does, when, how often) so you are not reproducing the wrong thing. A vague "it acts up" has to become a specific, observable event you can aim at.
Branch 1: is it the way they operate it
Start with operator input because it is the most common and the cheapest to test.
If the fault appears when they run it and not when you do, the difference may be technique: force, speed, sequence, or a habit. Do not test it your gentle standard way. Have them operate it while you watch, then match their exact input. The related articles on matching a user's technique and on the fault following the operator cover how to isolate and replicate this. If matching their technique reproduces the fault, you are done here. If it does not, move on; it is not the operator.
Branch 2: is it the environment at their time of use
If technique does not explain it, ask when it fails, not just how. Environmental variables track with time of day and season, and a mid-day service call misses them:
- Temperature and humidity, indoor or outdoor, at the hour it actually fails.
- Sun load on one side of a building in the afternoon.
- Other loads on the same supply that only run at certain times (morning demand, evening peak).
- A door open, a fan on, a nearby process running.
Recreate the timing or the condition, not just the task. A fault that only shows up on the coldest start or the hottest afternoon will hide from you at any other time.
Branch 3: is it the load or demand they place on it
If the environment is not it, look at how hard they push it. You test briefly and lightly; they may run it long, hot, or with several demands at once.
- Duration: a fault that only appears after extended continuous run, past your short test.
- Simultaneous demand: two fixtures, circuits, or functions running together that you tested one at a time.
- Volume: more throughput, cycles, or occupants than your test put through it.
Load it the way they do and hold it there long enough for the fault to develop.
Branch 4: is it the state the machine starts in
If load is not it, consider the starting state. You may test a warm, settled, recently-run machine; they hit it cold, or after it has sat, or in a specific accumulated condition. Cold-start versus warm-start faults are a whole family; reproduce the state they start from, not the state your prior testing left it in.
Branch 5: is it their supply or inputs
If none of the above, the difference may be upstream of the machine entirely: their incoming water pressure, voltage, fuel supply, or the materials and consumables they feed it, which differ from your test conditions or your own supplied parts. Measure the actual supply under their real demand.
Quick reference
| Candidate variable | What differs between you and them | How to test |
|---|---|---|
| Operator technique | Force, speed, sequence, habit | Watch them, match the input |
| Environment | Time of day, temperature, other loads | Recreate the timing and conditions |
| Load and demand | Duration, simultaneous use, volume | Run it as hard and as long as they do |
| Starting state | Cold vs warm, accumulated condition | Start it from their state |
| Supply and inputs | Pressure, voltage, fuel, materials | Measure the real supply under real demand |
Recap
- Confirm you and the customer mean the same failure.
- Rule out operator technique first: watch them, match it.
- Then environment and timing, then load and duration, then starting state, then supply.
- Reproduce the condition safely, and once the fault appears, diagnose from real readings.
References
- Trade-standard practice for reproducing customer-reported symptoms under field conditions
- Manufacturer documentation on supply requirements, duty cycles, and start conditions
- See related: Reproducing a Fault by Matching the User's Technique; The Fault Follows the Operator, Not the Equipment (decision tree); It Only Fails at This One Customer (decision tree)