Model vs Measured Performance

A model estimates energy use under standard conditions; a meter records energy use under real conditions. When the two differ, the difference is often called the performance gap, and it is usually a sign that conditions differed, not proof that anyone made a mistake. Real weather is warmer or colder than a standard year, real occupants heat different rooms to different temperatures, and real systems age, drift out of tune or develop faults.

That distinction matters because it changes the response. If a gap comes from behaviour, the remedy may be controls, advice or a different heating pattern. If it comes from construction, the remedy may be fabric work. If it comes from a modelling error, the remedy is to correct the model. Treating every gap as a construction failure wastes money, and treating every gap as behaviour excuses real defects.

Measured performance itself needs care. A single winter bill reflects that winter's weather and that household's habits, while a full year of meter readings smooths some variation but still reflects one household in one year. Smart meter data, where available, can show patterns across days and seasons, such as high overnight use or heating running when the home is empty, but it does not by itself explain why those patterns occur. Temperature logging in rooms adds context, showing whether a home is heated evenly or whether one room dominates demand.

Comparing model and measurement fairly means aligning the comparison: same period length, same fuel scope, and an allowance for weather differences. A model that assumes a standard heating season cannot be judged against three cold months without adjustment, and a model that covers heating and hot water cannot be judged against a bill that also includes cooking and appliances unless those are separated out.

Start by checking the model inputs before questioning the physics. Was the wall construction right, was the heating system right, were the controls right. Then check the measured side: is the meter data complete, does it cover the same fuels, and does it include any unusual period such as building work, a change of occupants or a long absence. Only after both sides are checked does it make sense to investigate construction quality, airtightness or system commissioning.

Where a gap persists after those checks, targeted investigation helps: a heating system service and controls check, a look at ventilation and draughts, and, where appropriate, an airtightness or thermographic survey by a qualified person. Each test narrows the possible causes, whereas a broad claim that the home underperforms narrows nothing.

EU and UK systems both face this comparison problem, but they frame it differently. In the UK, an EPC rating is explicitly a standardised comparison, so a gap against bills is expected and is not by itself evidence of a faulty EPC. In EU member states, national certificates and renovation plans similarly rest on standardised methods, and several countries are developing measured-data tools alongside certificates without replacing the standardised rating. The lesson is the same: standardised and measured figures answer different questions, and both are useful if their question is kept in view.

Regulators and researchers also use aggregated gaps, across many homes, to improve methods and defaults. An individual home's gap is a prompt for investigation, while a pattern across many similar homes is evidence that a method, a default or a construction practice needs review. Do not generalise from one home to a whole housing type, or from a housing type to one home.

A common mistake is to quote a performance gap as a single percentage and treat it as a verdict. A percentage without a period, a fuel scope and a weather note can mislead in either direction, making a sound home look poor or a poor home look sound. Another mistake is to chase the gap by adding insulation where the real cause is an uncontrolled heating system, or by replacing a heating system where the real cause is a draughty, uninsulated envelope.

A further mistake is to use a short monitoring period to judge a model. A few weeks of data in mild weather, or during a holiday, say little about annual performance. A full heating season, with notes on occupancy and settings, is a far stronger basis for comparison.

Keep the model report, the meter data and any notes on occupancy, weather and system settings together, with dates. When you ask a professional about a gap, that bundle lets them test explanations against evidence instead of debating impressions.

Questions worth asking include: over what period and for which fuels are we comparing, what was different about real conditions versus standard conditions, and which single investigation would narrow the causes most. A careful answer to those three questions usually points to the next sensible step.