Three kinds of input
Every figure on a non-domestic EPC falls into one of three categories, and the certificate does not tell you which is which.
| Category | What it means | Can it be improved without works? |
|---|---|---|
| Measured or documented | A tested or calculated value backed by paperwork – an air permeability certificate, a U-value calculation, a commissioning record | No – it is already the real value |
| Observed | The assessor could see it and recorded what they saw – glazing type, luminaire count, visible plant | Sometimes, if what was visible understated what is there |
| Assumed (a default) | Nothing could be seen and nothing was produced, so a published age-band default was entered | Yes – this is the opportunity |
Where defaults get recorded
These are the inputs that most often carry an assumption on a commercial building, in rough order of how much they tend to cost.
| Input | The default when nothing is evidenced | Why it is usually wrong |
|---|---|---|
| Air permeability | A cautious default leakage rate for an untested building | Buildings constructed or refurbished under recent Part L were tested – the certificate exists, often in the O&M manual |
| Wall construction | Age-band U-value for an unknown build-up, assumed uninsulated | Cavity fill, refurbishment insulation and dry-lining are all invisible from the outside |
| Roof insulation | Age-band default, or the minimum implied by the build date | Loft top-ups and warm-roof refurbishments are rarely documented on site |
| Glazing | Age-band assumption on frame and glass performance | Later replacement units frequently outperform the age assumption |
| Lighting | Generic luminaire efficacy and no credit for controls | An LED replacement programme with a schedule and product data is one of the biggest single gains |
| Heating and cooling efficiency | Default seasonal efficiencies for the plant type and age | Manufacturer data and commissioning certificates give the actual figures |
| Controls | No credit for controls that cannot be demonstrated | A BMS with zoning, optimum start and compensation earns credit only if it is shown to the assessor |
What an assumption costs you
Take a 1980s office of 1,800 m² rated D at 84 points, where the walls were recorded as unknown-construction cavity, no air test was available, and the lighting was entered generically.
The cavity was filled during a 2006 refurbishment; the invoice is in the client's files. An air test was carried out after a 2019 recladding; the certificate is in the O&M manual. The lighting was relamped to LED in 2022, and the contractor's schedule lists every fitting.
None of that was available on the day of the assessment, so none of it is in the model. The building is rated for a wall it does not have, a leakage rate it does not suffer and lighting it no longer uses. It has not been mis-assessed – it has been under-evidenced, and the remedy is a folder rather than a contractor.
What overturns each assumption
| Assumption on the certificate | The document that settles it |
|---|---|
| Unknown wall construction | Construction drawings, the specification, a cavity-fill guarantee or invoice, or an opening-up record |
| Unknown roof build-up | As-built details, refurbishment specification, or a measured insulation depth with photographs |
| Default air permeability | The air permeability test certificate |
| Generic lighting | The lighting schedule with luminaire and lamp data, and the controls specification |
| Default boiler or chiller efficiency | Manufacturer performance data and the commissioning certificate |
| No credit for controls | BMS schematics, points list, and the controls commissioning record |
| Default ventilation performance | Specific fan power data and AHU commissioning sheets |
| Unknown glazing | Window schedule, or the supplier's U-value and g-value data |
Assembling these is the subject of the evidence pack.
The honest limits
- This only works where an assumption was actually made. A well-evidenced certificate has little headroom, and that is a good outcome to discover cheaply.
- Evidence has to be genuine and contemporaneous. Reconstructing it after the fact is not the same thing, and an assessor will not accept it.
- SBEM publishes no per-measure point value, so nobody can honestly promise you a specific number of points from a specific document in advance. We give indicative ranges and say plainly that they are indicative.
- It may confirm bad news. Sometimes the default was generous.
Find out how much of your rating is assumption
We read the certificate on the register, identify which inputs carry defaults, and tell you what evidence would settle each one – before you spend anything on works.
Common questions
Can my EPC rating improve without doing any work?
Yes. Where the assessor recorded a default because no evidence was available, supplying that evidence at a re-assessment replaces the assumption with the real value. Nothing about the building changes; the model's description of it does.
Why would an assessor assume the worst?
Because the conventions require it. An assessor may only enter what they can see or evidence. Where construction is concealed and no documentation is offered, published age-band defaults are used, and those defaults are deliberately cautious.
Would a different assessor give me a better rating?
Not on the same evidence. Two competent assessors given the same information should reach materially the same result. The variable is the information, not the assessor.
How much difference can evidence make?
It varies with the building. The largest movements we see come from air permeability testing, lighting schedules and construction records on buildings where the walls were recorded as unknown.
Sources
- Simplified Building Energy Model (SBEM) technical manual – National Calculation Methodology, uk-ncm.org.uk
- Non-domestic building services compliance guide and Approved Document L – GOV.UK
Legislation is cited to the version in force at the date below. We publish a corrections policy and amend pages when the law changes.