The most consequential fact in an energy project may be sitting in someone's head rather than in the data room.
Models are necessary. Utility bills, load data, maintenance records, equipment schedules, engineering drawings, and financial assumptions create the analytical foundation for a business case. But institutional energy systems rarely behave exactly as the records suggest.
Years of operating history create workarounds, exceptions, deferred maintenance, seasonal behaviors, recurring failures, undocumented modifications, and practical knowledge that may never make it into a spreadsheet.
WORKING PRINCIPLE
The model describes the system. Operators know how the system actually lives.
A credible decision process needs both.
Clean data can hide a messy operating reality
A model may show a chiller operating for a given number of hours, a boiler at a specified efficiency, or a pump at its design point. The operating team may know that the chiller is rarely loaded as expected, the boiler short-cycles, the controls are bypassed in winter, or a supposedly redundant piece of equipment cannot actually carry the building during a peak condition.
Those differences are not minor details if they change the baseline against which an investment is being evaluated.
The danger is not that models are inaccurate by definition. It is that apparently precise inputs can create false confidence when the physical and operating context behind them has not been tested.
Operating history reveals risks that specifications cannot
Equipment records tell part of the story. The people who have lived with the system often know the rest.
Which components fail repeatedly? Which parts are difficult to source? Which alarms are routinely ignored because they are unreliable? Which systems are sensitive to weather? Which spaces are chronic comfort problems? Which shutdowns require more coordination than the drawings suggest?
This knowledge can change the value of replacement, redundancy, service contracts, controls upgrades, or a completely different infrastructure model.
Deferred maintenance changes the economics
A current cost baseline can understate the true burden of an existing system when major maintenance has been deferred.
An institution may appear to be operating equipment inexpensively because overhauls have been postponed, vacancies remain unfilled, spare parts are cannibalized, or capital repairs are moving from one budget cycle to the next. The resulting historical spend can look efficient while masking obligations that are accumulating.
A business case that compares a fully maintained new alternative with an under-maintained existing system is not making a like-for-like comparison.
Workarounds have costs even when accounting systems do not capture them
Operating teams are often very good at keeping imperfect systems running. Temporary fixes become recurring practices. Staff manually reset equipment. Portable systems appear during failures. Technicians spend time compensating for controls problems. Supervisors learn which sequence of actions keeps an unreliable system stable.
These workarounds can be economically significant without appearing as a distinct budget line. They consume labor, management attention, spare equipment, overtime, and risk tolerance.
They also create fragility because the system may depend on the experience of a few people who know what to do when something goes wrong.
DISCOVERY QUESTION
What do people have to do to make the current system look as reliable as the data says it is?
The answer often exposes hidden operating cost and institutional dependence.
Institutional memory can reveal future constraints
Frontline knowledge is not limited to the condition of existing assets. Experienced staff may understand why past projects were rejected, which shutdown windows are realistic, where site access becomes difficult, which departments resist certain operating changes, or why an apparently obvious technical modification was never made.
That context can prevent a new project team from rediscovering old constraints after significant time and money have already been spent.
Interview before you finalize the model
A useful energy assessment should treat interviews as evidence gathering, not as a courtesy meeting after the analysis is complete.
Facilities staff, operators, maintenance technicians, controls personnel, finance staff, and building users see different parts of the same system. Their observations can identify data gaps, contradict assumptions, explain anomalies, and reveal criteria that the formal project scope missed.
The strongest process is iterative: the model creates questions, interviews create new evidence, and the analysis changes when that evidence matters.
Experience is evidence, but it still needs to be tested
Institutional knowledge should not automatically override measured data or engineering analysis. People can remember selectively, normalize inefficient practices, or interpret the same problem differently.
The objective is not to replace quantitative analysis with anecdotes. It is to use operating knowledge to identify what deserves verification.
A recurring failure described by a technician can be checked against work orders. A concern about peak capacity can be tested against trend data. A claimed maintenance burden can be compared with labor records and invoices.
Good analysis turns operational observations into better questions and better evidence.
The people closest to the system often improve the decision
Energy infrastructure decisions are strongest when formal analysis and practical operating knowledge inform each other.
The spreadsheet can calculate consequences that operators cannot easily quantify. Operators can reveal conditions the spreadsheet does not know exist.
The goal is not to decide which source of knowledge is superior. It is to build a business case that is grounded enough in reality that both recognize the system being described.