A lower EPC does not automatically mean the property has become less efficient. Compare the inputs, assumptions and evidence before paying for improvements.
A landlord can renew an Energy Performance Certificate expecting the same result and discover that the rating has dropped. A previous C may become a D, or a comfortable D may move several points further away from C. That can affect letting plans, mortgage conversations and the amount of work the certificate appears to recommend.
The useful question is not simply whether the new rating is lower. It is why the inputs changed. The building may have changed, the assessment method may be different, or the new certificate may contain assumptions because evidence was unavailable. Until those possibilities have been checked, a recommendation list is not a reliable shopping list.
The quick answer
An EPC is a standardised model of the property, not a meter reading of its actual bills. Existing homes are normally assessed using Reduced Data Standard Assessment Procedure, or RdSAP. Government guidance explains that RdSAP uses assumptions to reduce the volume of information an assessor must collect. A rating can therefore change when the recorded construction, dimensions, heating, insulation, glazing, evidence or assessment methodology changes, even if the landlord has not deliberately made the home worse.
Seven reasons an EPC rating can go down
1. The two certificates used different assessment routes
A newly built home can initially receive a full SAP assessment based on drawings, specifications and a complete dataset. A later certificate for the same home may be produced using RdSAP after an on-site visit. The government’s Home Energy Model consultation specifically notes that a rating can change when a property moves from SAP to RdSAP because some original data is replaced by RdSAP assumptions. The newer rating is not necessarily evidence that the physical home deteriorated. It may reflect a different permitted dataset.
2. Evidence available for the older EPC was not available this time
Insulation may be hidden behind plasterboard, under a floor or inside a roof build-up. A heating appliance may have lost its model label. Windows may have been installed years ago without documents being kept with the property file. If an assessor cannot inspect a feature or accept suitable evidence, the methodology may require an assumed or default entry. That does not prove the feature is absent, but the assessor cannot simply accept an unsupported description from the owner.
3. The property was measured or divided differently
Floor area, exposed perimeter, extensions, conservatories, roof rooms and the age bands of different parts of the building feed into the model. A small difference in total floor area will not always transform the result, but several changed inputs can move a property near a band boundary. Compare the dimensions and building-part descriptions rather than assuming that both assessors recorded the same geometry.
4. Heating, hot water or controls were recorded differently
The boiler or heat-pump model, hot-water cylinder, secondary heating and heating controls all matter. A previous certificate may have recorded thermostatic radiator valves, a programmer and a room thermostat, while the latest assessment records a simpler control arrangement. An appliance can also be matched to a different database entry if its exact model cannot be confirmed. Check the descriptions against what is actually installed and retain commissioning, installation and product records.
5. A physical change reduced the modelled performance
Not every drop is an evidence problem. A failed boiler may have been replaced with a less efficient system. A room may have been brought into the heated envelope, increasing the area and heat loss. Insulation might have been removed during repair work, or a low-carbon system may have been replaced. Start with the possibility that the home genuinely changed, then distinguish that from changes in the recorded data.
6. The calculation conventions or software version changed
SAP and RdSAP are periodically updated. The rating is based on standardised occupancy and heating patterns and fixed model assumptions, including the cost basis behind the current headline Energy Efficiency Rating. Methodology and convention changes can alter the result without tracking the household’s actual bill. This is one reason certificates from different dates should be compared at input level, not only by looking at the coloured band.
7. One of the certificates may contain a data issue
Assessors and accreditation schemes have quality-control processes, but no large data system is immune from recording problems. Government EPC accuracy research published in 2026 examined weaknesses in assumptions, defaults and assessment processes. A difference does not prove that the latest assessor made an error, or that the earlier certificate was right. It means the conflicting entries deserve a structured review.
Compare the two EPCs before buying improvements
Download both certificates from the official energy certificate register and compare the underlying descriptions. The point score is the outcome; these entries are where the explanation normally begins.
- Total floor area, extensions, roof rooms and conservatory treatment.
- Wall, roof and floor construction, insulation descriptions and whether entries are assumed.
- Window type, glazing age, draught-proofing and the proportion of glazing.
- Main and secondary heating, appliance model, heating controls and hot-water cylinder insulation.
- Solar PV, solar thermal or other renewable technology and whether the system details are complete.
Build an EPC evidence file for the property
The best time to collect evidence is before the next assessment. Acceptance depends on the methodology, the item and the assessor’s accreditation requirements, so ask what will be usable rather than arriving with an unsorted folder on the day. Useful records can include:
- Building Regulations completion documents, plans and specifications for extensions or conversions.
- Insulation guarantees, installation records, invoices with a clear specification and dated photographs where acceptable.
- Window and door certificates showing installation date and specification.
- Boiler, heat-pump, hot-water and heating-control model details plus commissioning records.
- MCS certificates, commissioning information and system specifications for eligible renewable installations.
Where a material entry is assumed, not inspected or not rated, a specialist EPC assessment can establish whether accepted documentary evidence, airtightness testing or U-value testing is relevant. Testing is evidence, not a manufactured score. It must be suitable for the property, accepted within the assessment route and used by a qualified assessor when the certificate is lodged.
Do not treat the recommendation list as an instruction to buy everything
An EPC recommendation list provides modelled options, not a property-specific works specification. A landlord close to C may need better evidence, a control improvement or another proportionate measure. A different property may genuinely need fabric or heating work. Spending thousands before identifying the points gap and the inputs behind it can lock the landlord into disruption that did not address the decisive issue.
A practical D-to-C route for landlords
- Confirm the current certificate, score and expiry date.
- Compare it with the previous certificate and mark every changed or assumed entry.
- Collect accepted evidence for the entries most likely to affect the result.
- Model the credible evidence route before specifying physical measures.
- If evidence does not close the gap, compare proportionate controls, heating, fabric or funded routes for that property.
The enforceable domestic private-rented minimum in England and Wales remains EPC E unless a valid exemption applies. The government has confirmed the policy direction of EPC C by 2030, with implementing legislation still required. That makes early landlord EPC planning valuable, but it does not justify assuming that every D requires the same package of work.
New HEM-based EPCs arrive in 2027
Domestic EPC reform is scheduled for the second half of 2027. The proposed Home Energy Model EPC will use multiple headline measures and a more modular approach to existing-home data. The detailed HEM:EPC methodology and band boundaries are still being finalised. Do not assume that a current score can be translated directly into a future fabric, heating or smart-readiness band. The sensible action now is to improve the property evidence and understand the current D-to-C route, not to invent a future score.
Questions landlords ask about a lower EPC
Does a lower EPC mean the property has become less efficient?
Not necessarily. The property may have changed, but the result can also move because the assessment route, dimensions, recorded systems, available evidence, assumptions or methodology changed. Compare the inputs before drawing a conclusion.
Can an EPC assessor correct a certificate?
Raise a specific suspected data issue with the assessor or their accreditation scheme and provide the relevant evidence. A disagreement with the rating alone is not proof of an error, and changes must follow the applicable assessment and quality-assurance process.
Will better evidence always improve the EPC rating?
No. Evidence can confirm a better, equal or worse input than the permitted assumption, and not every type of evidence can be used for every entry. Review the likely effect and acceptance route before paying for specialist work.
Does an airtightness or U-value test guarantee EPC C?
No. Testing records performance; it does not manufacture points. The result may improve, leave unchanged or worsen the relevant model input. It only helps when the test is suitable, the result is accepted and the complete assessment reaches the required band.
The bottom line
When an EPC rating falls, do not begin with a product quotation. Begin with the two certificates, the recorded inputs and the property evidence. That review can reveal a genuine building change, a different permitted assumption or a small number of entries that deserve further investigation.
Cucumber Eco reviews the current EPC, score and available evidence before recommending testing or improvements. Start your free EPC assessment.
