renewableenergyforbusinesses

Degree Days for Energy Monitoring: Is Your Heating Use Really Falling?

Updated 28 October 2026 · Renewable Energy for Businesses

Want to cut what your site uses and pays for energy?

Request a quote

Ask about your site

By submitting you agree to our privacy policy. We never sell your details.

What is a heating degree day?

A heating degree day scores how cold the weather was in a form that lines up with heating demand. CIBSE's Technical Memorandum 41 (TM41) describes degree days as a tool for analysing weather related energy consumption in buildings: a running total of the gap between the outdoor temperature and a reference, or base, temperature over a set period.

With a base of 15.5°C, a day averaging 10.5°C outside adds 5 degree days, and a day warmer than the base adds none. That is the idea at its simplest.

Heating use rises in a cold winter and falls in a mild one whether or not anything in the building has changed. Setting kWh against degree days separates the weather from the building, which turns a lower bill into evidence of a saving. It is the measuring step behind any plan on the energy management page.

Is 15.5°C the right base temperature for your building?

The Met Office's gridded heating degree day datasets, listed on data.gov.uk, use 15.5°C: they count the mean number of degrees by which the air temperature falls below 15.5°C, day by day. A published base is a convention and a starting point, not a measurement of your building.

The base is not the thermostat setting. People, lighting, equipment and sunshine all add heat, so a building needs no heating until the outside air is some way below the indoor set-point. TM41 calls the definition of the base temperature a key issue and ties it, in buildings, to the energy balance of the building and system. So the right base varies:

  • A crowded office full of IT, or a site with warm process plant, points to a lower base.
  • A building kept warmer for vulnerable occupants, or one with little internal heat, points to a higher base.

If the base is badly wrong, the points on the chart tend to bend rather than follow a straight line, which is a cue to try another. An upgrade that cuts heat loss, such as insulation, can lower the base itself, so recheck it after fabric work.

Where can a UK business get degree-day data?

For a practical primer, CIBSE points readers to the Carbon Trust's 2006 Good Practice Guide 310, Degree days for energy management. For the figures themselves:

SourceWhat it publishesBest used for
DESNZ, Energy Trends table ET 7.1 (GOV.UK)Average temperatures and heating degree days, with deviations from the long-term mean, in weather tables published monthly on the last ThursdayWas last month colder or milder than normal?
DESNZ, Digest of UK Energy Statistics table 1.1.9 (GOV.UK)Mean heating degree days for 2002 to 2025, published each JulyComparing winters over the long run
Met Office gridded datasets (listed on data.gov.uk)Annual values of heating degree days below 15.5°C; viewing them needs registration on the Met Office websiteComparing places, not tracking months
Degree Days.net (commercial website)Describes itself as a site that generates degree days for locations around the worldFigures for a named location

The government series are there for context: DESNZ says they provide contextual information for consumption patterns in energy, and the readings come from Met Office weather stations. Degree Days.net is named as an example of a commercial tool, not a recommendation. Whatever the source, use a location near the building, keep one source and one base throughout, and match the degree-day periods to your meter-reading dates.

How do you test heating use against degree days, step by step?

The method uses ordinary monitoring-and-targeting tools. The contents of a CIBSE-listed textbook, Energy Management and Operating Costs in Buildings, cover the same ground, from linear regression and degree days to a common time base and cumulative sum deviation.

  1. Collect at least 12 months of actual readings for the meter serving the heating. Estimated reads blur the picture; GOV.UK's business energy efficiency campaign notes that smart meters send automatic readings to the supplier.
  2. Put every period on one footing. Convert to kWh, then use equal periods or kWh per day, so a 35-day period is not set against a 28-day one.
  3. Match the degree days to exactly the same periods, from one source, at one base, for a nearby location.
  4. Plot and fit. Degree days along the bottom, kWh up the side, one point per period, and a straight trend line through them.
  5. Read the intercept: the use at zero degree days, meaning the load that does not depend on the weather, such as hot water, catering or process heat on the same meter.
  6. Read the slope: the extra kWh for each additional degree day. This is the heating response that insulation, controls and plant upgrades should push down.
  7. Check the scatter. Points far from the line need explaining: an estimated read, a shutdown, a fault or a change of hours.
  8. Set the target. Expected use for any later month is the intercept plus the slope multiplied by that month's degree days. Judge actual use against that, not against last year's bill.
  9. Run a CUSUM. Each month, add actual minus expected to a running total. A flat line means no change, a steady fall means a real saving and a rise means waste, and the month where the line turns dates the change.

What does the method look like in practice? A hypothetical example

The figures below are invented and rounded to show the method. They describe no real building, and every month is treated as the same length.

Suppose a small office's baseline line is 4,000 kWh a month plus 100 kWh per degree day. The intercept is its hot water and kitchen gas; the slope is its heating response. In January its heating controls are reset so that weekends are no longer heated.

MonthDegree daysExpected kWh (4,000 + 100 × degree days)Actual kWhActual minus expectedCUSUM
October15019,00019,00000
November25029,00029,500+500+500
December30034,00034,0000+500
January30034,00031,000-3,000-2,500
February25029,00026,500-2,500-5,000
March20024,00021,500-2,500-7,500

October to December sit on the line. From January the CUSUM turns down and keeps falling, which dates the improvement to the controls reset. January to March used 79,000 kWh against an expected 87,000 kWh: a weather-adjusted saving of 8,000 kWh, or about 9%.

Now the trap. Suppose the same months a year earlier were colder, at 350, 300 and 250 degree days, and the office used exactly what its line predicts: 102,000 kWh in all. The raw bills now show a fall of 23,000 kWh, more than a fifth. Most of that is milder weather; only the 8,000 kWh belongs to the controls. A colder winter can do the reverse and make a genuine saving disappear from the bills.

What can throw a degree-day analysis off?

  • Occupancy and hours. A new shift, a floor let to a tenant or more home working moves the intercept and the slope. Date every change, and rebuild the baseline if the building is used differently.
  • Hot water and process heat on the same meter. They sit in the intercept, so a kitchen closing can look like a heating saving. A separate meter on that load gives a cleaner line; the sub-metering guide covers meter types and approval.
  • The wrong base. A curved scatter is the warning sign. Summer months with no heating can also flatten the low end, so fitting heating-season months on their own can help.
  • Cooling. TM41 covers the dual concepts of heating and cooling degree days. In an air-conditioned building, analyse cooling electricity against cooling degree days on its own chart rather than mixing it with heating.
  • A change of fuel. When a heat pump replaces a gas boiler, the meter, the fuel and the slope all change, so electricity per degree day cannot be compared with gas per degree day. Start a new baseline from the first full heating season after the switch.

How does this fit ISO 50001-style energy performance indicators?

An energy management system has to show that performance improved, not just that bills fell. BS ISO 50006:2023, published by BSI, gives guidance on how to establish, use and maintain energy performance indicators (EnPIs) and energy baselines (EnBs) in any organisation, including those using ISO 50001, and on how to demonstrate improvement. The ISO 50001 guide covers the management system itself.

The degree-day line can do both jobs. The baseline-year regression is an energy baseline that already allows for weather, and actual use measured against its prediction, monthly or as a CUSUM, is a weather-adjusted indicator. That beats the shortcut of dividing kWh by degree days. In the hypothetical office, October works out at about 127 kWh per degree day (19,000 over 150) and December at about 113 (34,000 over 300), with no change in the building. The ratio swings because the weather-independent load is spread over fewer degree days in mild months; the regression keeps intercept and slope apart.

Degree days are one driver among several in a wider monitoring plan, alongside operating hours and output. The energy management page shows where weather-adjusted tracking fits among the other measures, and readers can send an enquiry through the form on that page.

Sources

All pages checked 28 October 2026. If a publisher's own page disagrees with this one, that page is the authority.

Common questions

Can degree days be used with daily or half-hourly data?

Yes. Half-hourly readings can be summed to daily totals and set against daily degree days, which shows a change within weeks rather than months. Weekends and holidays are best analysed separately or left out, because the building runs differently on those days.

Do I need special software for degree-day analysis?

No. A spreadsheet with a scatter chart and a linear trend line gives the intercept and slope, and one extra column gives the running CUSUM total. Monitoring platforms automate the same steps, so when comparing them ask which base temperature and which weather location they use.

Why has my kWh per degree day gone up in a mild month?

Usually because the weather-independent load, such as hot water, is being divided by fewer degree days. A simple ratio rises in mild months even when nothing in the building has changed. Compare actual use with the regression line instead.

How do I show a board that a heating saving is real?

Fix the baseline line before the change, then report actual use against the line's prediction, with the CUSUM chart showing when the gap opened. Keep the same data source and base temperature throughout, and list any changes in hours or occupancy so the result can be checked.

Request a renewable energy for businesses quote

By submitting you agree to our privacy policy. We never sell your details.

Related guides

Request a quote
Request a quote