The check would have taken minutes. Instead, the resulting investigation, rework, and production delay ended up consuming days.
Managers recognize this pattern. An unnoticed problem grows until it demands attention through materials, overtime, worker stress, or customer complaints. In some operations, the consequences include recalls or fines.
These costs can become accepted as “just part of business.” But late discovery does not have to be normal.
On October 17, 2026, USDA announced a recall of more than 1.5 million pounds of ready-to-eat meat and poultry products because of possible temperature abuse during storage. Discovery followed a customer complaint about warm products. The investigation identified temperature deviations in a warehouse cooler.
The announcement does not explain the internal monitoring or notification arrangements. But it raises a key question: How do we identify an unacceptable condition and alert the right people before more work or product is affected?
Control starts with a defined expectation
In the first post of this series on Control, we proposed this definition:
A process is under control when the organization knows what should be happening, can recognize meaningful differences between the expected and actual condition, and can respond before those differences become larger problems.
The previous post explored defining the expected state: required work, timing, and acceptable conditions or results.
Now the question is how departures from expectations become visible. A requirement in a procedure cannot, by itself, tell a manager it was just missed.
Recognizing when something is wrong
Consider a grinder that must be cleaned and checked before the next production run. Several situations could need attention:
Cleaning has not been fully completed.
A check found an unacceptable condition.
Cleaning is taking progressively longer, threatening the schedule.
Production is approaching, but confirmation of cleaning is unavailable.
Recognizing these conditions before production begins creates a different opportunity from discovering them after several batches have been produced.
Recognition depends on current information: a measurement, a recorded result, a worker’s observation, or knowing the absence of something that should already have happened.
Designing work for recognition means connecting each important expectation to information that can reveal whether it is being met. What would show that the activity occurred, that the result was acceptable, or that something needs attention? When must that information be available, and who needs to know if it is missing or outside acceptable limits?
These questions make recognition part of how the work is organized. They reduce dependence on someone remembering to check or discovering the problem later.
Recognizing a problem early creates an opportunity to respond before more work or product is affected.
Someone may already know
An operator hears an unfamiliar sound. A technician finds unexpected wear. An employee cannot complete a required check. This information may provide early warning. But someone noticing a problem does not necessarily mean the organization knows about it.
The observation may stay in a conversation, a notebook, or someone’s memory. Under pressure, a worker may dismiss a concern or find a workaround. If reporting is difficult or routinely leads to blame, useful information may never reach someone who can act.
Human observation is part of how a system senses change. Maintaining control requires making those observations available in time to influence action.
Missing information is a warning, too
A verification never arrives. A required activity isn't completed by its deadline. A measurement becomes too old to support the next decision.
An absent completion record does not prove the work was missed. It means completion has not been established.
The absence of a reported problem doesn't mean conditions are acceptable.
A control system needs to distinguish between conditions that are acceptable, conditions that are not, and conditions that cannot be verified.
Getting the right information to the right people
Managers cannot continually inspect every record or question every worker.
When something needs attention, an automatic notification can alert the person responsible. It might flag a measurement outside acceptable limits, a missed deadline, an unacceptable result, or missing information needed before work can proceed.
The goal isn't simply to generate alerts. It's to give people the opportunity to act while a problem is still manageable, rather than discovering it after the consequences have spread. But notification alone isn't enough. Someone must respond and confirm that control has been restored.
If earlier recognition and better notifications would help your team, let’s talk about how GapCross can support your work.
Knowing what should happen and recognizing when reality differs are essential to maintaining control. But not every difference requires the same response.
In the next insight, we’ll explore that judgment: Not Every Difference Matters, So Which Ones Do?