How to right-size commercial battery storage

Most commercial battery installations are wrong before they're even installed. Industry data and field reports suggest that 30–40% of commercial battery energy storage systems are oversized, while a smaller but more painful share are undersized for the very peaks they were bought to shave. Either way, the result is the same: capital wasted, payback stretched, and an asset that quietly underperforms for the next 10–15 years. Commercial battery storage sizing is the single biggest lever for ROI on a battery project — and it's the one most fleet operators, facility managers, and multi-site SMBs get wrong because they size on rules of thumb instead of data.

This guide walks through how to size a commercial battery the way the best operators do it: from interval load data, tariff structure, and use-case priorities — validated with dispatch simulation before a single cell is purchased.

What does right-sizing a commercial battery actually mean?

Right-sizing a commercial battery means choosing the kW (power) and kWh (energy) combination where the avoided utility charges, demand response revenue, and resilience value outweigh the lifetime cost of the system — and no further. It is not about buying the biggest battery your roof or budget can hold. It is about finding the smallest system that captures the majority of the achievable savings on your specific load profile and tariff.

A right-sized battery has two distinct dimensions:

• Power rating (kW): how much demand the battery can shave at any instant. This drives demand charge reduction.

• Energy capacity (kWh): how long it can sustain that discharge. This drives time-of-use arbitrage, solar self-consumption, and backup duration.

Get the kW wrong and you can't cut the peak. Get the kWh wrong and the battery runs out mid-event. Get the ratio wrong and you pay for capacity you'll never cycle.

Why most commercial battery installations are sized wrong

Oversizing usually comes from three places: vendor incentives to sell larger systems, sizing based on monthly kWh totals instead of 15-minute interval data, and conservative "just in case" buffers stacked on top of already-conservative inputs. The result is a battery that cycles shallowly, never reaches the duty cycle it was specified for, and stretches simple payback from a realistic 4–6 years toward 8–10.

Undersizing is rarer but more visible. It typically comes from sizing on average load instead of peak load, ignoring coincident peaks across EV charging, HVAC, and process loads, or designing for today's site without accounting for planned EV fleet growth, heat pump retrofits, or new tenants. An undersized battery hits its state-of-charge floor mid-peak event and the demand spike happens anyway — wiping out the demand-charge savings that justified the investment.

The common root cause: sizing without dispatch simulation. Static spreadsheets cannot model how a battery will actually behave across 8,760 hours of variable load, weather, tariff, and solar generation.

The 5-step methodology for commercial battery storage sizing

This is the framework used by experienced C&I storage developers and by software platforms that automate the same logic at scale.

Step 1: Pull 12 months of 15-minute interval load data

Monthly utility bills are useless for sizing. They show kWh totals and a single monthly peak — but they hide the shape of your load. You need interval data: the kW recorded by your utility meter every 15 minutes (or 5 minutes in some markets) for at least 12 months.

Most US and European utilities will provide this on request through a customer portal or a Green Button download. If your meter is not yet smart-metered, that is the first conversation to have — sizing without it is guesswork.

With interval data you can answer the questions that actually matter:

• What is my real coincident peak, not just my monthly maximum?

• How many hours per month do I exceed 90% of peak?

• How spiky are my peaks — short bursts of 15–30 minutes, or sustained plateaus of 2–4 hours?

• How does load shape change between weekdays, weekends, and seasons?

Step 2: Define your primary use case (and rank the rest)

A battery cannot be optimized for every revenue stream simultaneously. Pick a primary use case and rank secondary ones. The four common commercial use cases are:

1. Peak shaving / demand charge reduction — short, high-power discharges to clip monthly peaks. Demand charges can account for 30–70% of a commercial electricity bill, and peak shaving is usually the highest-ROI use case for SMBs in markets with significant kW-based charges.

2. Time-of-use arbitrage — charging when energy is cheap, discharging when it is expensive. Most attractive in markets with dynamic or strongly differentiated TOU tariffs.

3. Solar self-consumption — storing midday solar surplus for evening use instead of exporting at low feed-in rates.

4. Backup and resilience — keeping critical loads alive during outages.

The ranking matters because it changes the sizing math. A peak-shaving-first battery wants high power and 1–2 hours of energy. A solar-self-consumption-first battery wants moderate power and 4+ hours of energy. A backup-first battery is sized to a critical-load schedule, not an economic optimum.

Step 3: Map your tariff structure in detail

A battery is a tariff arbitrage machine. Sizing it without modeling the tariff is like sizing a truck without knowing the freight. You need to capture:

• Demand charges — $/kW, including whether they are non-coincident, coincident peak, or ratcheted (where a single bad month locks in higher charges for 11 more).

• Time-of-use energy rates — peak, off-peak, and shoulder $/kWh windows by season.

• Capacity charges — based on Peak Load Contribution or 4CP/5CP coincident peak readings.

• Dynamic or real-time pricing — if you are on, or eligible for, a dynamic tariff, hourly volatility becomes a major sizing input.

• Export rates and net-metering rules — what you get paid for exported solar, which determines how aggressively a battery should soak up surplus.

The more granular and volatile the tariff, the more value a correctly sized battery captures — and the more expensive a wrong size becomes.

Step 4: Run dispatch simulation across 8,760 hours

This is the step most projects skip and the one that separates a right-sized battery from an expensive guess. Dispatch simulation runs your interval load data, tariff, solar generation profile, and candidate battery sizes through an hour-by-hour (or 15-minute) optimization for an entire year.

For each candidate kW/kWh pair the simulation answers:

• How much demand charge is avoided each month?

• How much TOU arbitrage and solar self-consumption value is captured?

• How many cycles does the battery accumulate, and what does that imply for degradation?

• What is the marginal value of the next 25 kWh or the next 50 kW?

Plot the savings curve and you almost always see the same shape: steep gains up to a point, then a flat plateau where each additional kWh barely moves the needle. The right size is the knee of that curve — typically 30–40% smaller than what a vendor proposal will recommend.

Step 5: Stress-test against degradation, growth, and edge cases

A battery loses capacity every cycle. A lithium-iron-phosphate (LFP) system typically retains 70–80% of nameplate capacity after 10 years of daily cycling. Your sizing must account for end-of-life capacity, not day-one capacity, or the system will silently underperform in years 7–10.

Also stress-test for:

• Load growth — adding 10 EVs, a heat pump, or a new tenant over the next 3 years.

• Tariff changes — capacity charges and dynamic pricing are tightening across the EU and many US ISOs in 2026.

• Outage duration assumptions — if backup matters, model realistic outage scenarios, not best cases.

This is where modular battery architectures earn their premium: they let you start at the right size today and add modules later without replacing the inverter or controls.

How big a battery do I need for my business?

For most small and mid-sized commercial sites, a right-sized battery falls in one of three brackets:

• Small site, peak-shaving focus (single retail unit, small workshop, 50–150 kW peak): typically 30–100 kWh / 30–60 kW, sized to clip the top 15–25% of monthly peak demand.

• Mid-size site, mixed solar + peak shaving (warehouse, depot, multi-tenant building, 200–500 kW peak): typically 150–400 kWh / 100–200 kW, balancing demand-charge reduction with solar self-consumption.

• EV fleet depot or multi-load commercial site (500 kW–1 MW peak with chargers, HVAC, and on-site solar): typically 400 kWh–1.5 MWh / 250–600 kW, with sizing driven by coincident charging windows and tariff volatility.

These ranges are starting points, not answers. The only way to land on the right number for a specific site is interval data plus dispatch simulation. A facility with 300 kW peak and a flat load shape needs a very different battery from a 300 kW facility with spiky EV charging on top of a baseline.

Sizing for peak shaving vs. solar self-consumption vs. backup

The biggest sizing mistake is treating these three goals as the same problem. They are not.

Peak shaving is a power problem. The battery must deliver a specific kW for the duration of the peak event — usually 30 minutes to 2 hours. Energy capacity beyond what is needed to cover the longest plausible peak is wasted on this use case alone. A peak-shaving battery is typically specified at a 0.5C–1C ratio (1 hour of energy per kW of power, or less).

Solar self-consumption is an energy problem. The battery must absorb midday surplus and discharge it through the evening shoulder. Power requirements are modest, but energy capacity must match a typical day's surplus. These systems usually run at 0.25C–0.5C (2–4 hours of energy per kW).

Backup is a duration problem. The battery must support critical loads for a defined outage length — often 4 hours or more — at lower power than peak shaving. Sizing here is driven by the critical-load schedule and outage duration target, not economics.

When all three matter, you do not stack three batteries. You run dispatch simulation that respects the priority ranking and finds a single system that captures most of the value across all three. This is exactly what software-driven dispatch is built to do.

Why software-driven dispatch simulation is the missing step

A spreadsheet can estimate demand-charge savings to within ±30%. That is not good enough when a battery represents a $150,000–$500,000+ capital decision with a 10–15 year lifetime. Modern energy management platforms simulate dispatch on real interval data and real tariff structures — and then run that same logic in real time once the battery is installed.

SortGrid, an AI-powered energy management platform for small and mid-sized businesses, runs dispatch simulation across solar, battery, EV charging, and HVAC loads for every site in a portfolio. Before you commit to a battery size, you can model how a 100 kWh, 200 kWh, or 400 kWh system would have performed against your last 12 months of actual load and tariff data — and see the diminishing-returns curve at the kWh level. After installation, the same platform dispatches the battery automatically against live tariffs, solar generation, EV charging schedules, and HVAC demand, capturing the savings the simulation predicted.

This matters because the gap between simulated savings on the right-sized system and realized savings on a default-controlled oversized system is often 25–40%. Right-sizing only pays off if you can actually dispatch the battery against the assumptions you sized it on.

Common commercial battery storage sizing mistakes

1. Sizing on monthly kWh instead of 15-minute kW. Monthly bills hide the shape of the load and lead to systematic oversizing.

2. Ignoring coincident peaks. EV charging, HVAC, and process loads can stack. A battery sized for the largest single load misses the actual site peak.

3. Designing around nameplate capacity instead of end-of-life capacity. A system that just barely covers peak on day one will fail to cover it in year 8.

4. Treating peak shaving and solar self-consumption as additive sizing. Adding the kWh required for each use case independently almost always produces an oversized system.

5. Skipping dispatch simulation. Without a year-long hour-by-hour model, sizing is a guess dressed up as a calculation.

6. Ignoring tariff direction. Capacity charges and dynamic pricing are tightening across most markets. A battery sized to today's tariff is undersized for 2027.

7. Buying integrated hardware without an open control layer. A battery you cannot dispatch intelligently behaves like a smaller, dumber battery.

The ROI math: how right-sizing changes payback

With installed lithium battery prices for commercial systems now in the $180–$300/kWh range for containerized projects and pack prices below $100/kWh, the economics have shifted hard. Payback periods that were 7–10 years in 2020 are routinely 3–5 years in 2026 for right-sized systems on demand-charge-heavy tariffs.

But payback is brutally sensitive to sizing. A simple example, holding everything else constant:

• A correctly sized 200 kWh / 150 kW system on a site with $18/kW demand charges might capture ~85% of the achievable demand-charge savings at a payback of roughly 4 years.

• Oversizing the same site to 350 kWh / 200 kW captures maybe 92% of the savings — but at 75% more capital cost. Payback stretches toward 6–7 years.

• Undersizing to 100 kWh / 75 kW captures only ~55% of the savings because the battery hits its energy floor mid-event. Payback can actually be longer than the oversized case.

The knee of that savings curve is where the project lives or dies. Dispatch simulation finds the knee. Vendor proposals usually do not.

From sizing to optimization: making your battery actually pay

Right-sizing a battery is the first half of the job. Operating it intelligently is the second. A correctly sized battery that runs on default vendor firmware — simple time-of-use schedules, fixed peak thresholds — typically delivers 60–75% of its theoretical savings. The rest is left on the table because static control logic cannot react to real-time tariff changes, solar forecasts, EV charging demand, or unexpected load spikes.

This is where multi-asset energy management platforms close the gap. SortGrid coordinates battery dispatch with solar self-consumption, EV charging schedules, and HVAC pre-conditioning across every site from a single dashboard. When a tariff spike is forecast, the platform pre-charges the battery from solar or off-peak grid power. When EV charging is queued, the platform load-balances chargers against battery state of charge to avoid tripping a new monthly peak. When the day-ahead forecast shows a cheap energy window, HVAC pre-cools the building so the battery can be saved for the expensive afternoon.

For multi-site SMBs — fleet operators with 3–10 depots, multi-property landlords, retail chains, parking operators — this kind of coordinated dispatch is the difference between a battery that pays back in 4 years and one that pays back in 7. Sizing is the floor. Optimization is the ceiling.

Right-size first, then optimize

The summary is short and the consequences are long:

• Get interval data. Monthly bills are not enough.

• Pick a primary use case. Peak shaving, solar self-consumption, or backup — one of them leads.

• Model the tariff in detail. Demand charges, TOU, capacity, dynamic pricing.

• Run dispatch simulation across 8,760 hours. Find the knee of the savings curve.

• Stress-test for degradation, load growth, and tariff change.

• Buy at the knee, not above it. And operate the battery with software that can actually capture what the simulation promised.

If your team is tired of guessing battery sizes from vendor PDFs, juggling EV chargers, solar, and HVAC across multiple sites, and watching demand charges creep upward every quarter, SortGrid simulates the right battery size against your real load and tariff data — and then dispatches it automatically alongside your solar, chargers, and HVAC, so every site runs at its lowest possible energy cost without the complexity.