Energy Monitoring ROI: How to Calculate Your Savings

The business case for an energy monitoring system usually dies in a meeting that starts with the wrong question. If the first slide asks “what will this software cost?” the project is already defending itself against a budget number with no savings number attached. The honest question is the reverse: given what we spend on energy and the avoidable waste in our own measured data, how fast does a monitoring system pay for itself? Answering that requires building a genuine financial model — baseline, savings fractions, demand-charge economics, hardware and labor costs, and a defensible payback — rather than relying on a vendor’s “average savings” claim.

The Two Ledgers: Energy and Demand

Commercial and industrial electricity bills have two cost components, and the ROI math must treat them separately. Energy charges (kWh) are volume-based — reducing them is a straightforward per-unit saving. Demand charges ($/kW) are capacity-based: the utility charges for your peak demand in the billing interval, often the highest 15-minute average in the month, regardless of how briefly it occurred. A monitoring system pays on both ledgers, but the demand side is where the multiples live — shaving one 200 kW peak for fifteen minutes can save more than a week’s worth of careful kWh trimming.

The reason monitoring specifically drives both is that both require visibility. You cannot shave a peak you cannot see coming, and you cannot find the base-load leak without knowing where the load is. The monitoring ROI therefore has a structure: avoidable energy savings from found faults and schedule fixes, plus peak reduction from forecast-and-shed control, plus the soft benefits of auditability (compliance, ESG, capital planning) that are hard to monetize but worth documenting.

Step 1: Establish the Baseline From Your Own Data

Never build the business case from industry averages alone; build it from the site’s own utility bills and interval data. Assemble 12 months of data:

• kWh per month and average unit price ($/kWh), from the tariff or the bill line items.
• Peak kW per month and the demand charge ($/kW), noting whether it is ratcheted (the utility holds the year’s highest peak and charges it all year).
• Base load — the minimum hourly consumption in the year. This is the theoretical floor of always-on waste; a monitoring system typically finds faults that move this number.
• Time-of-use structure, if the tariff prices peak and off-peak hours differently, because a monitoring system enables shifting load into cheaper hours.

A worked example grounds the model. A 150,000 ft² office uses 4.2 million kWh per year at a blended $0.12/kWh ($504,000/yr), with a monthly demand peak of 1,100 kW and a demand charge of $12/kW ($13,200/mo, $158,400/yr), for a total electric bill of roughly $662,000 per year.

Step 2: Estimate the Achievable Savings Pool

The honest way to size the savings is not a single “10%” figure but a stack of separately justified reductions:

Savings source	Typical range	How it is justified
Found faults and schedule fixes	5–12% of kWh	Interval data exposes base-load leaks, stuck schedules, runaway equipment
Peak-shaving via forecast + shed	5–15% of demand charge	Monthly kW forecast triggers load-shed before the 15-min peak forms
Setpoint/occupancy optimization	3–8% of HVAC kWh	BAS setpoint resets verified against measurement
Energy-reduction project verification	Confirms planned savings	Baseline model proves the retro-commissioning worked

For the example office, a conservative stack: 8% of kWh from faults and schedules (336,000 kWh ≈ $40,300/yr) and 8% of the demand charge from peak management ($12,700/yr) — roughly $53,000 in the first year, before any capital project savings.

Step 3: Count the Real Costs

The cost side is where business cases get wrecked by omissions. Account for:

• Hardware: meters, CTs, gateways, sensors — and their installation. Field labor is typically the largest line item; a retrofit with split-core CTs is far cheaper per point than a build-out requiring panel work.
• Software: the monitoring platform license (typically per point or per meter) and any cloud/storage fees.
• Integration: connecting the BAS, meters, and utility interval data; often the single biggest hidden cost if the estate speaks many protocols.
• Labor: one-time setup and ongoing review time (a facility manager spending two hours a week on the dashboard is a real cost).
• Ongoing O&M: recalibration, sensor replacement, gateway maintenance.

For the example office, a realistic deployment — 24 panel meters with split-core CTs, two gateways, installation, and the first year of software — lands near $42,000.

Step 4: Compute Payback, NPV, and IRR

With first-year savings of $53,000 and an initial cost of $42,000, the arithmetic payback is under a year on the conservative stack — which is why the example is deliberately realistic rather than a sales pitch. The financial metrics to present:

• Simple payback: upfront cost ÷ annual savings.
• Payback with escalation: if energy prices rise 4% a year, the second-year saving is larger; recalculate to show the real curve.
• NPV and IRR: discount the 5–10 year cash flow at the company’s hurdle rate. At a 10% discount rate, $53,000/yr for ten years has a net present value of roughly $283,000 against the $42,000 outlay — an IRR well above any hurdle rate a CFO is likely to name.

The honest caveat: savings persist only while the monitoring is used. The single biggest ROI risk is not installation cost but abandonment — a system nobody reviews stops finding faults. The business case should include a line for a recurring monthly review (internal or vendor-assisted), because that is what converts the first-year saving into a ten-year cash flow.

Sensitivity: Which Assumptions Matter Most?

Run the model three ways — conservative, central, optimistic — and identify the swing factors. For most sites the sensitivity ranking is:

1. Actual kWh saved (the fault-find pool). This depends on how much always-on waste the site has; a site with a tight BAS and disciplined schedules will find less than a site with six years of ad-hoc control changes.
2. Demand charge structure. A ratchet clause or a high $/kW rate dramatically raises the value of peak management.
3. Energy price trajectory. In markets with rising prices, even modest savings compound fast.
4. The review discipline. As noted, the abandonment risk is the silent killer of the model.

The Thresholds That Matter

A few decision rules help keep the business case honest:

• If the conservative payback is under 24 months, the project should not need heroic assumptions to pass.
• If it is over 60 months, the site probably has either very low energy prices, very tight existing controls, or an over-scoped hardware plan — shrink the deployment, start with the top feeders, and prove the model before expanding.
• Never claim demand savings on a site that has no demand charge (some tariffs are energy-only); the model must match the actual tariff.

Building the Case, Not the Fantasy

The ROI calculation is ultimately a discipline of separating what you can measure from what you hope. The baseline is your own data; the savings stack is a set of separately justifiable mechanisms; the costs include every hour of labor; and the payback is presented as a range with its swing factors named. Done that way, the business case survives the CFO’s questions and, once the system is installed, the same baseline model verifies whether the projected savings actually materialized — closing the loop from prediction to proof. Integrar IoT’s platform provides the interval data collection, demand forecast, and baseline-model savings verification that make both the business case and its audit trail come from one consistent source.