← Back to Energy guides

12 min read

How to Reduce Your Business Peak Demand Charges

Once you understand that peaks are priced differently from totals, the practical question is what your UK site can change next week without a six-figure programme. This guide emphasises operational levers—sequencing, controls, maintenance, and honest half-hourly detective work—that shave the top off the demand curve. Read {link('maximum-demand-charges-explained.html', 'maximum demand charges explained')} first if the vocabulary still feels abstract.

Next step: If you use under about 50,000 kWh a year, you can get a quote in under 90 seconds online — fast, no obligation. Larger supply, half-hourly metering, or prefer chat? Use the contact page.

Find the guilty half-hours

Sort a year of HH data by kW, not only by cost. Map the top events to shift patterns, weather, maintenance tests, and reopen days. If you are not half-hourly yet, read half-hourly versus non-half-hourly meters.

Stagger, soften, and supervise starts

Introduce modest delays between large loads—often a few minutes prevents stacking into one settlement period. Soft starters and VSDs cut inrush on motors. Document programmes in BMS/PLC comments so the next service visit does not “optimise away” your work.

Tactics vs typical impact

Maintenance debts that inflate peaks

Stuck dampers, fouled condensers, and compressed-air leaks force plant to work hardest at the worst moments. A focused maintenance week sometimes beats another LED rollout.

Implementation checklist

• Baseline peak kW and keep a dated change log.
• Pilot on one worst day-type before rolling network-wide.
• Tell shifts why stagger matters—otherwise someone “helps” by starting everything early.

If reactive power is part of the story, involve power factor correction before you lock a long fix. For procurement alignment, read how to read a business energy contract.

When peaks snap back after a “good” month

Demand programmes fail quietly: someone changes a BMS schedule for a one-off event and never reverts it; a contractor “speeds up” morning preheat; EV chargers lose their stagger after a firmware update. Build a quarterly review that overlays occupancy, production, and the top ten half-hours. If peaks return, treat it like a near-miss safety incident—investigate, document, retrain.

Cold snaps and heat waves also lie in wait. Pre-commit sensible rules: which loads may defer, which cannot, and who can authorise exceptions. If you operate across regions, remember microclimates and building fabric differ; copying a London playbook to a Scottish warehouse without adjustment invites both comfort complaints and expensive spikes.

Half-hourly analytics without hiring a data science team

You do not need a warehouse of servers—just discipline. Export CSVs monthly, keep a single canonical spreadsheet with the top 50 kW intervals, and annotate each with a human cause code (weather, test, reopen, fault). Over time, patterns become obvious: the same compressor test every Tuesday, the same cleaning shift energising everything at once. That annotated file is gold at renewal when a supplier claims your profile is “too volatile” to price fairly.

When you trial a change, run an A/B style comparison: four weeks before vs four weeks after, matched by day-type where possible. Share the chart with operators so they see cause and effect. People change behaviour when they recognise their own fingerprints on a spike—much faster than when finance emails a scolding about “discretionary load.”

Battery, on-site generation, and the peak-shaving promise

Storage and solar can help, but only with honest limits: export caps, inverter behaviour, maintenance access, and who operates dispatch when prices move. If a vendor promises automatic peak shaving, ask what happens during a comms outage and whether manual override exists. Pair technical diligence with solar for business economics so the project competes fairly against cheaper operational fixes.

Talking to finance in pounds, not jargon

Translate kW into annual cost using your blended rates and pass-through sensitivity—not a generic national average. Show a band: conservative, central, optimistic. Finance funds ranges with assumptions, not hero single points. If you cannot explain the peak in one sentence to a non-engineer, you are not ready to buy hardware to fix it.

Finally, celebrate wins visibly. When stagger policies shave twenty kW off a crest, tell the team. Momentum beats another spreadsheet nobody opens. Keep the before/after chart pinned beside the health and safety board for a month—energy becomes real when people see their own fingerprints on the curve.

Related guides

Keep reading: three-phase power explained, battery storage for business, and the energy hub index.