Peak Demand, Hidden Cost: Reading Half-Hourly Data Before You Renew Capacity

Capacity charges are easy to overlook until they appear as a fixed monthly sum that barely moves when consumption falls. For UK organisations with half-hourly meters, the agreed capacity figure behind that charge deserves as much scrutiny as the unit rate. It is a reservation of network capacity, and it is shaped by the site's peak demand rather than its average load.

Why the agreed capacity line matters more than the unit rate

Most energy reviews begin with pence per kilowatt hour. That is understandable, but it can miss the part of the bill that reflects how much capacity the site asks the network to hold ready. Distribution use of system charges often include a capacity element based on agreed kVA, and exceeding that figure can trigger excess capacity charges. The commercial risk runs in both directions: too much reserved capacity means paying for headroom that is rarely used, while too little creates exposure to short, expensive peaks and awkward conversations when operations change.

Half-hourly data is the practical antidote to guesswork. It shows 48 settlement periods a day, so rather than relying on a monthly maximum or a supplier's summary, an energy manager can see exactly when demand rose, how long it stayed high and what else was happening on site.

What half-hourly data can tell you that a monthly bill cannot

A monthly bill may show a maximum demand figure, but it rarely explains the shape of the event behind it. Half-hourly data reveals whether the peak was a single late-afternoon spike, a sustained cold-weather period or a short-lived start-up surge. That distinction matters because capacity is not set by energy consumed; it is set by the highest level of demand the site needs the network to support.

Start by obtaining at least 12 months of half-hourly data, ideally covering the full winter 2026/27 planning period and the previous heating season. Convert kW to kVA using the site's power factor where available; if it is not, ask the supplier or metering agent for a measured figure rather than assuming a generic value. Then rank the half-hours by demand and look at the top 20 to 50 events. Patterns usually emerge quickly: the 16:00 to 18:00 overlap when heating, lighting and ventilation coincide; Monday morning restart loads; or a piece of equipment that runs only during certain production cycles.

Finding the spikes that actually set the risk

A useful first pass is to group peak events by cause. Some will be weather-driven and recurring. Some will be operational and avoidable. Others will be one-off commissioning or maintenance events that should not drive a permanent capacity reservation. The aim is not to explain away every spike, but to separate the demand the site genuinely needs from the demand it could manage.

Look also at duration. A peak that lasts one settlement period may be less commercially significant than a plateau that persists for two hours, even if the single half-hour is higher. Network capacity is reserved in kVA, but the operational response differs: short spikes may be addressed through sequencing or soft-start controls, while sustained peaks may point to a genuine need for more capacity or a change in working patterns.

Is your capacity oversized, tight, or about right?

A simple utilisation ratio helps frame the conversation. Divide the highest observed kVA in the period by the agreed capacity. If the site never exceeds 65 to 70 per cent of its agreed capacity, even in cold weather, there is a reasonable case to review whether the reservation is too generous. If it regularly runs above 90 per cent, or has already breached the agreed level, the site is operating with little margin and should either justify the headroom or plan a controlled increase.

The ratio is a starting point, not a verdict. A site with a high utilisation ratio may still need its current capacity because the peak is unavoidable and safety-critical. A site with a low ratio may still need headroom for planned expansion. The evidence should be presented as a profile, not a single number: maximum demand by month, the number of half-hours above 80 per cent and 90 per cent of agreed capacity, and a note on which events were exceptional.

Turning evidence into a negotiation

When approaching a supplier or distribution network operator, a data-led case is stronger than an assertion that the bill looks high. A concise pack should include the 12-month maximum demand profile, the dates and times of the highest events, the operational explanation for each, and any planned mitigation such as load shifting, equipment replacement or revised start-up procedures. If the aim is to reduce agreed capacity, show how the site will manage the remaining risk and what trigger would prompt a review. If the aim is to increase capacity, show the growth or operational change that justifies it.

This is also the point to test the tariff structure rather than only the capacity figure. Ask how excess capacity is charged, whether the agreed level can be varied during the year, and what notice period applies. For sites with seasonal profiles, a single annual capacity figure may be a poor fit; the data may support a staged review rather than a once-a-year negotiation.

A practical sequence for the next renewal

Begin with a complete half-hourly dataset and a clear power-factor assumption. Identify the top demand events and label them as weather, operational or exceptional. Compare those events with the agreed capacity and calculate utilisation by month. Discuss the findings with operations before proposing any change, because a capacity reduction that ignores a legitimate winter peak is a false economy. Finally, build the evidence into the renewal conversation and record the assumptions used, so the next review starts from a documented baseline rather than a fresh guess.

The late-afternoon spike is not inherently a tax. It becomes one when a site pays for capacity it does not need, or pays excess charges for peaks it could have seen coming. Half-hourly data gives UK businesses the evidence to tell the difference.