A shift supervisor on an afternoon food processing packaging line sees output running slowly after forty minutes. She calls the maintenance lead to ask whether there is an open work order. He is on the floor and misses the call, so she leaves a voicemail. Eighteen minutes later, he calls back. The work order had closed an hour ago. The actual issue is conveyor belt tracking, which belongs to the line technician. She calls the technician, who is at lunch. Forty-two minutes pass between noticing the issue and reaching the right person. The CMMS could have identified the issue in about ninety seconds.
This happens constantly on manufacturing floors. This is not a complaint about disorganized supervisors or hard-to-reach maintenance leads. It is an observation about data access. When the person asking cannot quickly see the information needed to route the question, the question goes to whoever seems likely to know. That person may or may not be the right one.
How the Pattern Works
Plant-floor operations questions usually fall into three groups. The first has a clear owner and simple routing, such as sending a safety event to the safety officer. The second has unclear ownership because the relevant data is not immediately available: output dropped 9% on Line 2 in the last hour. Is it a maintenance issue, a process issue, or a material issue? The third is genuinely ambiguous and needs expert judgment, even when the data is available.
Most costly routing calls belong to the second group. The data contains a findable answer, but the supervisor does not have it in front of them. They call whoever seems most likely to know, such as the maintenance lead, quality engineer, or process engineer. Sometimes that person is right. Often, the question has to be redirected.
The Real Cost Is Not the Call
The three-minute call to the maintenance lead is easy to see. The hidden cost is the context switch on the other end.
A maintenance lead interrupted during a repair has to stop, answer, and resume the work. If he misses the call and returns it fifteen minutes later, the original caller may be on another task and need to reconstruct the question. If the callback shows that the routing was wrong, the sequence starts again with someone new.
A shift supervisor at a busy discrete manufacturing plant may start eight to fifteen routing sequences per shift. Some end quickly. Others take several callbacks before reaching the right person. The accumulated interruption cost for the supervisor and everyone called becomes part of the soft overhead that plant managers describe as simply how the floor works.
Why Data Access Changes the Pattern
The second category is unclear because the person asking cannot see the data that would resolve the question. Before calling anyone, a supervisor who sees the Line 2 output drop could ask, "What open work orders exist for Line 2?" and "What was the packaging unit sensor reading during the last hour?" A closed work order with a recent maintenance completion means something different from no work order. A temperature reading two degrees above the upper control limit points toward a process issue rather than a maintenance issue.
Data cannot route every question correctly. Some problems genuinely need expert judgment. But many calls interrupting maintenance leads, quality engineers, and process engineers could be resolved with sixty seconds of data access instead of twenty minutes of phone tag.
This is what we built Tendrel for: questions a supervisor could answer from plant data if that data were available. "Is there anything in the maintenance records that would explain this production drop?" is something a CMMS can answer when the question can be asked in plain language and returned in plain language. No call is needed.
The Expert Call That Needs the Expert
Reducing unnecessary routing does not mean stopping expert calls. Some problems need the process engineer. Others need the maintenance lead at the machine. Experts are more available for those problems when their time is not consumed by routing questions that plant data could answer.
A maintenance lead who receives ten calls per shift, including three that need his judgment and seven that only require redirection, is in a different operating position from one who receives three calls because the other seven questions were answered before anyone picked up. The information has not changed. Its path through the plant has.
Measuring the Pattern
Routing overhead is hard to measure because it is spread across dozens of short interactions that each look normal. To see the pattern, track what prompted every call a shift supervisor made during a shift. Then record whether the information behind the call was already available in plant data.
In early conversations with plant teams before a pilot, we consistently hear that supervisors underestimate how often they call for information stored in a system rather than judgment only a person can provide. The distinction matters. Judgment requires the call. Information requires data access, which can be addressed without anyone's phone ringing.
The forty-two minutes in the opening example is not unusual. It is the median, not the outlier, in plants where operational data is available only to a small group of people. Making that data available changes who needs to be called, not only how quickly a call gets answered.