A pattern worth learning to recognize, because it looks like a seasonal problem and is not one.
Normal behavior on a commercial account is that billed demand and consumption move together. Summer arrives, the cooling load comes on, both rise. Correlation between the two sits high, usually above 0.9, and load factor stays in a reasonable band.
The pattern I want to flag is the opposite. Billed demand sits in a very narrow range across two summers and two winters — a coefficient of variation under about eight percent — while consumption underneath it moves seasonally as you would expect. Correlation between demand and consumption falls to 0.6 or lower. Load factor lands in the thirties.
That is not sustained load and it is not HVAC. HVAC demand is seasonal by definition. A peak that holds nearly constant regardless of season is a short-duration recurring event setting the interval maximum. Motor or compressor inrush is the usual cause, or several machines starting together on the same schedule.
Why it matters commercially: something occupying a few seconds can set the billed demand for an entire month, and where that is the cause it is frequently correctable by sequencing or soft-start at modest cost. The saving then recurs indefinitely.
You cannot prove any of this from statements. You need the interval data. But the monthly figures tell you whether it is worth asking for.
Normal behavior on a commercial account is that billed demand and consumption move together. Summer arrives, the cooling load comes on, both rise. Correlation between the two sits high, usually above 0.9, and load factor stays in a reasonable band.
The pattern I want to flag is the opposite. Billed demand sits in a very narrow range across two summers and two winters — a coefficient of variation under about eight percent — while consumption underneath it moves seasonally as you would expect. Correlation between demand and consumption falls to 0.6 or lower. Load factor lands in the thirties.
That is not sustained load and it is not HVAC. HVAC demand is seasonal by definition. A peak that holds nearly constant regardless of season is a short-duration recurring event setting the interval maximum. Motor or compressor inrush is the usual cause, or several machines starting together on the same schedule.
Why it matters commercially: something occupying a few seconds can set the billed demand for an entire month, and where that is the cause it is frequently correctable by sequencing or soft-start at modest cost. The saving then recurs indefinitely.
You cannot prove any of this from statements. You need the interval data. But the monthly figures tell you whether it is worth asking for.