How to Measure Energy Cost per Unit Produced

A lower bill does not settle the question

Your electricity bill falls during a quiet month. That may be welcome for cash flow, but it does not tell you whether the factory used energy more effectively. Output may have fallen faster than consumption. A different product mix may have required less processing. The bill can also change because of the tariff rather than anything on the floor.

For an operating decision, connect consumption to the work it supported. Start with one line, product family or process where a better measurement would change an action: scheduling a warm-up, investigating idle consumption, comparing a setup or evaluating equipment. Ask planning and finance which decision matters, then check with operators and maintenance how the process actually uses energy.

You need two related measures. Energy intensity describes consumption per defined unit of output. Energy cost per unit adds prices and an allocation policy. Keep them visible separately. Otherwise a price increase can look like deterioration in production, and a convenient allocation can look like an engineering improvement.

Excellence industrielle’s 2024 report describes energy workshops using participating companies’ own data. That is a useful starting discipline: establish what your records can support before comparing your operation with a benchmark. The worksheet here is a proposed calculation method, not a formula supplied by that report. [1]

Draw a boundary the records can support

Write down which equipment and services the measurement includes. A line meter may cover the production machine but exclude compressed air, extraction, cooling and lighting. A building meter includes those loads but may also include offices and another process. Either boundary can be useful if its exclusions are explicit and comparisons use the same basis.

Match the time window to production records. Record meter identifiers, opening and closing readings, timestamps, units and any missing intervals. Kilowatts describe power at a moment or averaged over an interval; kilowatt-hours describe energy accumulated over time. A nameplate rating multiplied by shift hours is an estimate based on assumed loading, not a substitute for measured consumption.

Have qualified personnel assess and install any additional metering under the site’s safety procedures. Do not ask operators to improvise electrical measurements. Before adding equipment, check whether existing meters, utility interval data or approved controller readings provide enough evidence for the initial question. Record their limitations, including whether auxiliary loads are captured.

Choose output that means something

For a single product, accepted good units are often a useful denominator. Energy used on rejected units remains in the numerator, so scrap does not disappear from the result. Define when a unit counts as accepted and how later quality decisions revise the period. Keep gross production, rejects and rework alongside the accepted count so the number can be explained.

A mixed-product line needs more care. One small part and one large part may require very different processing. Compare product families separately where the records permit it, or use an explicitly defined equivalent-output measure. Do not call a change in product mix an efficiency gain. Keep the conversion factors and their owner visible, and revisit them when the process changes.

Warm-up, changeovers, cleaning and idle periods also consume energy. Decide whether to include them in the run, allocate them across a campaign or show them as separate categories. The decision should match the question. Excluding warm-up may help compare steady running; including it may be essential when deciding whether to split an order into several short runs.

Calculate consumption before allocating money

The basic calculation is measured energy within the boundary divided by accepted output for the same period. For electricity, kWh per good unit equals the closing meter reading minus the opening reading, adjusted for a documented meter multiplier where applicable, divided by good units. If the period has no accepted output, report the energy and the reason; do not divide by zero or quietly discard the period.

Consider this illustrative calculation, not a Quindart customer result. A defined line consumes 600 kWh during a run, including its agreed warm-up and idle allowance, and produces 1,200 accepted units. Electricity intensity is 0.50 kWh per good unit. At an assumed energy-only rate of CAD 0.12 per kWh, the variable electricity cost is CAD 72, or CAD 0.06 per good unit.

If the same energy accompanies only 1,000 accepted units, intensity becomes 0.60 kWh per good unit and variable electricity cost becomes CAD 0.072 per good unit at that assumed rate. This does not establish the cause. Scrap, slower output, waiting or a changed process could explain the difference. The connected production record tells the team where to investigate.

Where rates vary by time, multiply each interval’s consumption by the applicable rate and add the results. Keep gas or other fuels in their native measured units until a documented conversion is needed. If the question is total energy cost, include their costs explicitly rather than presenting electricity alone as the factory’s complete energy use.

Keep the tariff and shared-load assumptions visible

The utility bill may contain demand charges, fixed charges and other adjustments as well as energy charges. Their treatment depends on the actual tariff. A monthly peak charge is not automatically a variable cost for each unit produced. Finance can allocate it for product costing, but that allocation should be labelled and should not be presented as a saving that will occur when one run uses fewer kWh.

Shared services create a similar distinction. If a compressor serves several lines, use measured consumption or a defensible allocation driver where possible. Record the method, period and uncertainty. Check that allocations sum to the shared total and that the same energy has not also been included in each line meter. An unexplained residual is a question to investigate, not a number to hide.

Use the worksheet to preserve three outputs: measured energy intensity, variable energy cost per good unit, and fully allocated energy cost under the stated policy. Review changes against product mix, operating time and quality before claiming an improvement. Quindart’s information-first approach joins those records so an energy number leads to a useful investigation rather than another debate about whose spreadsheet is right.