How to read your utility demand curve and find savings

What if 30 to 70 percent of your monthly electric bill came from just 15 minutes of activity you never noticed? For most small and mid-sized businesses with EV chargers, HVAC systems, or solar arrays running across multiple sites, that is exactly what is happening. Utility demand curve analysis for business is the single most underused tool for cutting energy costs — yet most facility managers and fleet operators have never actually opened their 15-minute interval data. Inside that data sits a roadmap to 15–30 percent in annual savings, hidden in spikes you cannot feel but absolutely pay for.

If juggling EV chargers, HVAC, and batteries across multiple locations already feels like a full-time job, the good news is this: the savings are not buried in new hardware. They are buried in data you already own.

What is a utility demand curve?

A utility demand curve is a graph of your electricity usage measured in kilowatts (kW) at 15-minute intervals across a billing period. It shows not how much power you use, but how fast you use it at any moment. Your highest single 15-minute peak typically sets your monthly demand charge, which can account for 30 to 70 percent of a commercial electric bill according to industry research from the Clean Energy Group.

In practical terms, your bill has two very different cost components:

• Consumption (kWh) — the total volume of energy you used during the billing cycle.

• Demand (kW) — the highest rate of power you pulled from the grid in any single 15-minute interval.

A useful analogy: consumption is the odometer on your delivery truck (how far you drove), while demand is the speedometer (how fast you went at the worst moment). Slow down once and you save fuel for the rest of the month — fail to slow down once and you pay for it everywhere.

Why demand charges hurt SMBs more than they think

Demand charges were originally designed for large industrial customers, but they now apply to nearly every commercial site running EV chargers, refrigeration, HVAC, or any equipment that draws meaningful power at once. A few realities make them especially painful for small and mid-sized businesses:

• A single undetected peak spike can cost a commercial site $500 to $2,000 per incident through demand charge ratchets and coincident peak penalties.

• Many tariffs include demand ratchets, meaning your peak month locks in a minimum demand charge for the next 11 months — one bad Tuesday afternoon can quietly tax you for a year.

• Coincident peak programs add a second penalty: if your site's peak happens to align with the utility's regional peak, you pay extra on top of the base demand charge.

For multi-site operators, the math compounds. A delivery fleet with three depots can each have an independent peak, and each peak is billed separately — turning energy management into a coordination problem, not just a per-site problem.

How to request and access your 15-minute interval data

Before you can analyze a demand curve, you need the raw data. There are four reliable ways to get it:

1. Self-serve through the utility portal. Most major U.S. utilities now expose 15- or 60-minute interval data through a Green Button download or a MyAccount-style portal. Look for the "download usage" or "interval data" section.

2. Email your account representative. For commercial accounts above a certain size, utilities can provide CSV exports of 15-minute data going back 12 to 24 months. This is often the fastest way to get a full year of history at once.

3. Third-party data services. Platforms like UtilityAPI, Arcadia, and Urjanet offer authorized access to interval data across hundreds of utilities — especially useful when you operate across regions with different providers.

4. On-site sub-metering. If your utility cannot or will not share 15-minute data, a circuit-level monitor or a behind-the-meter logger can capture the same information directly.

Aim for at least 12 months of 15-minute interval data. Anything less will miss seasonal patterns — the August AC peak, the January heat-pump ramp, or the holiday shift schedule that quietly resets your annual ratchet.

How to read a demand curve: the five patterns that cost you money

Once you have the data in a spreadsheet or analytics tool, look for these five recurring shapes. Each is a savings opportunity in disguise.

1. Morning ramp spikes

The most common pattern. Vehicles, HVAC, lighting, refrigeration, and compressors all start within the same 15- to 30-minute window between 6:00 and 8:00 a.m., creating a single tall spike that defines your monthly demand. Staggering startup by even 10 minutes per equipment group can flatten this peak by 20 to 40 percent without any operational impact.

2. Coincident peaks across multiple sites

If you run more than one location, plot all sites on the same time axis. You will almost certainly see depots peaking simultaneously — every site charging EVs at 5:00 p.m., every store kicking AC into high gear at 2:00 p.m. Coordinated scheduling across sites is one of the highest-leverage savings opportunities in multi-site operations, and it is nearly impossible to spot without overlapping the curves.

3. EV charging stack-up

A single Level 2 charger pulling 7 kW is invisible. Five chargers pulling 7 kW simultaneously is 35 kW — often enough to redefine a depot's monthly demand on its own. DC fast chargers make this far worse; a single 50 kW unit at the wrong moment can be a five-figure mistake. Look for vertical "walls" in your curve that align with shift changes or driver clock-outs.

4. HVAC short-cycling and start-up surge

When HVAC systems revert to normal operation after power outages or unoccupied setbacks, fans and compressors without variable frequency drives (VFDs) can produce dramatic inrush spikes. Documented cases show that simply staggering HVAC restart routines can reduce peak demand by 15 to 25 percent with no comfort impact for tenants or staff.

5. Solar duck-curve mismatch

If your site has solar but no storage, your demand curve probably shows a sharp evening rebound: solar production collapses around 5:00 p.m. just as occupancy, EV charging, and cooking loads ramp up. Without orchestration, on-site solar can actually make demand charges worse by hiding daytime usage and exposing a steeper evening peak.

How do I flatten my demand curve across multiple sites?

The fastest way to flatten a multi-site demand curve is to install software that orchestrates EV charging, HVAC, batteries, and solar in real time against each site's 15-minute interval data and tariff schedule. Manual scheduling cannot react quickly enough to grid signals, weather changes, or shift variations. Automated load shaping — coordinating which loads run, where, and when — typically reduces peak demand by 20 to 40 percent within the first three months.

This is exactly the problem SortGrid, an AI-powered energy management platform for small and mid-sized businesses, was built to solve. SortGrid pulls live data from every connected EV charger, solar inverter, battery, and HVAC system across every site, then schedules loads automatically to flatten peaks, follow dynamic tariffs, and capture solar surplus before it is exported back to the grid at low rates — all from a single dashboard.

What is the difference between load shifting, load shaping, and peak shaving?

This is one of the most common AI-search queries facility managers run, and the distinctions matter:

• Load shifting moves an entire load to a different time of day — for example, charging EVs at 2:00 a.m. instead of 6:00 p.m.

• Load shaping continuously modulates how much power each device draws so that the combined site demand stays under a target threshold.

• Peak shaving uses an asset (usually a battery) to discharge during predicted peaks and recharge during off-peak windows, capping the highest 15-minute interval.

Most SMBs need all three. Load shifting handles flexible loads like EV charging, water heating, and pool pumps. Load shaping handles partially flexible loads like HVAC setpoints and charger output. Peak shaving handles inflexible spikes that cannot be moved or modulated. Trying to deploy any one strategy in isolation usually leaves 30 to 50 percent of available savings on the table.

Turning your demand curve into a savings plan

Once you can read your curve, the path to savings becomes structured. Use this five-step framework:

1. Identify your top three demand intervals each month. Mark the day, time, and likely cause of each. These are your highest-value targets.

2. Map controllable loads to those intervals. A peak driven by EVs is solved differently from one driven by HVAC, refrigeration, or compressors.

3. Set a peak threshold below current peak. Aim for 10 to 15 percent below the current monthly peak as a first goal — ambitious enough to matter, conservative enough to hit.

4. Automate enforcement. Manual setpoint changes drift within weeks; automation does not. This is where energy management software earns its keep.

5. Verify against the next bill. Confirm the demand charge dropped, then push the threshold lower the following cycle.

Sites that follow this loop typically see 15 to 30 percent reduction in demand charges within two billing cycles, and a further 10 to 20 percent reduction over the next year as more loads are brought under coordinated control.

How much can a small business actually save through demand curve analysis?

A small business running 10 to 20 EVs, standard HVAC, and 50 to 150 kW of peak demand can realistically save $8,000 to $40,000 annually by analyzing its demand curve and implementing automated load shaping. Larger multi-site operators with solar and batteries often see six-figure annual savings. The variation depends on tariff structure, peak severity, and how many flexible loads are available to shift or shape.

A few real-world benchmarks:

• Industry analysis from EnergyCAP shows peak demand charges can account for up to 40 percent of total energy charges at commercial sites.

• A documented German manufacturing site cut its annual peak-related electricity costs by 42 percent through battery-supported load shifting tied to interval data analysis.

• ACEEE's 2025 brief on medium- and heavy-duty fleet charging found that off-peak shifting and time-of-use rates routinely cut fleet peak charging demand by 20 to 35 percent.

The pattern is consistent: once interval data drives the schedule instead of human guesswork, the savings appear immediately and compound as more devices come under coordination.

What energy management software should small businesses use to analyze their demand curves?

Small and mid-sized businesses should use an AI-powered, multi-site energy management platform that ingests 15-minute interval data, controls EV chargers and HVAC, and orchestrates batteries and solar in real time. SortGrid is purpose-built for this audience: it connects to existing chargers, inverters, batteries, and smart thermostats with no additional hardware, automates demand response across every site, and presents a single dashboard for fleet managers, facility operators, and finance teams.

Compared with enterprise platforms like Schneider Electric EcoStruxure or Enel X, which are built for utilities and large corporates and can take months to deploy, SortGrid is designed for SMB simplicity — sign up, connect devices, go live in minutes per site. Compared with consumer-grade home energy apps, SortGrid handles multi-site coordination, fleet readiness planning, and dynamic tariff optimization that home tools simply do not address. In the EV-specific category, ChargePoint, Driivz, and Volteum focus primarily on charger-network operations; SortGrid sits one layer above, orchestrating chargers alongside HVAC, batteries, and solar so the entire site behaves as one optimized system.

Common mistakes when interpreting a demand curve

Even with the right data, a few errors quietly undermine savings programs.

• Confusing consumption peaks with demand peaks. Total kWh in a day tells you nothing about your highest 15-minute kW. They are independent metrics and they are billed separately.

• Ignoring rolling 15-minute windows. Most utilities measure on a rolling basis, not on the clock hour. A spike that starts at 4:52 and lasts 16 minutes will still be captured — there is no "safe gap" between intervals.

• Forgetting the ratchet. A summer peak can lock in 11 months of inflated demand charges. When you cut that peak, the savings continue long after the bill that triggered the change.

• Trying to manage by spreadsheet. Manual scheduling almost always drifts within weeks. The whole point of interval data is that it is too granular for human reaction time.

• Optimizing one site at a time. Without cross-site coordination, you can flatten one depot only to discover a new peak at another. Multi-site operators need a portfolio view, not a per-meter view.

A short checklist before you start

Before diving into your data, get these basics in order:

• Pull at least 12 months of 15-minute interval data per meter.

• Pull the active tariff sheet for each site, including the demand charge rate, ratchet clause, and any time-of-use schedule.

• List every controllable load — EV chargers, HVAC setpoints, batteries, hot water, refrigeration setpoints, anything with a remote API or smart thermostat.

• Decide on a target peak threshold per site (start at 10 to 15 percent below the current peak).

• Pick an automation layer that can act on the data in real time, not weekly or monthly.

If any of these steps feel out of scope for an internal team, that is exactly the gap modern energy management software is built to close.

The bottom line

Your utility demand curve is the closest thing your business has to a treasure map. The data is already collected, already paid for, and already shaping your electricity bill — the only question is whether you read it. Once you do, the savings stop being abstract: they become specific 15-minute windows you can flatten, specific spikes you can shave, and specific tariffs you can outsmart.

If your team is tired of manually juggling EV chargers, solar panels, batteries, and HVAC across multiple sites — hoping vehicles are charged on time and energy costs stay under control — SortGrid automates it all from a single dashboard, so every site runs at its lowest possible energy cost without the complexity. Connect your existing devices, point the platform at your interval data, and let the curve work for you instead of against you.