The Fundamental Assumption SFG20 Makes That Technology Is Challenging

SFG20's maintenance schedules are built on a single operating assumption: that we do not know the condition of an asset between service visits. Because we cannot observe it continuously, we service it at intervals calculated to prevent failure before it occurs. The interval is set conservatively — frequently enough that failure is unlikely, infrequently enough that the schedule is commercially viable. The interval is fixed regardless of whether the asset is new or old, lightly loaded or running at capacity, in a clean office or a dusty plant room.

This assumption was entirely reasonable in 1990, when SFG20 was created. Building management systems existed but were expensive, proprietary, and limited in scope. IoT sensors did not exist in any commercial sense. The only practical source of asset condition data was the engineer standing in front of the asset with a clipboard. Fixed-interval maintenance was not a limitation of the standard. It was the only rational model given the available information.

In 2026, that assumption is no longer universally true. A modern commercial office estate, an NHS trust with a current BMS, or a data centre will have continuous real-time data on the performance of its primary M&E plant: supply and return temperatures, flow rates, pressure differentials, energy consumption, vibration signatures on rotating plant, runtime hours, fault codes. The engineer with a clipboard is no longer the only source of condition data. On a well-instrumented estate, the BMS knows more about the assets between service visits than the SFG20 schedule assumes.

SFG20 was built for a world where fixed-interval maintenance was the only rational model. On a modern instrumented estate, it is no longer the only model. It is increasingly the expensive default.

sfg20 technology