Ask an asset manager what proportion of their stock condition data covers roofs properly and you will get a pause. The honest answer, for most providers, is that roofs are recorded as inspected as far as visible from ground level, which for a rear slope, a valley or a flat roof over a stair core means not at all.
This is not a criticism of stock condition surveyors. It is a consequence of physics. You cannot see a roof from the pavement, and nobody is going to scaffold four thousand homes to look at them.
What the regulator actually expects
The Regulator of Social Housing's Safety and Quality Standard, in force since April 2024, requires registered providers to have an accurate, up-to-date and evidenced understanding of the condition of their homes, based on a physical assessment, and to use that understanding to inform their investment plans. Consumer regulation judgements since then have repeatedly pointed at stock condition data as a reason for a C3 or C4 grading.
The word that matters is evidenced. A survey that records a roof as not inspected is honest, but it is not evidence of condition. Multiply that across the roofs on an estate and a provider can hold a complete survey that says very little about the component most likely to cause a damp and mould case, an excess cold hazard or a falling-element hazard under Awaab's Law Phase 2.
Why the gap persists
Three reasons, and all of them are about access rather than effort.
Cost. Access equipment for an inspection costs more than the inspection. A scaffold tower for one property is hundreds to thousands of pounds; a cherry picker is a few hundred a day and cannot reach a rear terrace slope. Against a survey budget of tens of pounds per property, roof access has never been affordable at scale.
Time. A ground-based surveyor covers perhaps a dozen properties a day. Adding roof access to each visit would cut that by more than half.
Geometry. Terraces expose their rear slopes only to neighbouring gardens. Blocks with central stair cores have flat roofs invisible from anywhere on the ground. Valleys and hips on complex roofs are hidden by the roof itself.
None of these is solved by trying harder. They are solved by changing the vantage point.
Twenty to fifty roofs on one estate, flown to a standard shot list, reviewed by a named person, delivered as component rows your asset system can import.
What aerial evidence changes
A drone flown to a standard shot list captures every slope, ridge, hip, valley, chimney, gutter run and flashing on a property in a few minutes, from the same angles on every property. On a clustered estate a pilot can cover ten to thirty roofs in a day. That does two things a ground survey cannot.
It gives coverage. The rear slopes and the flat roofs are in the data.
It gives comparability. Because every roof is photographed the same way, the observed condition of the covering on number 12 can be set against number 14 and number 40, which is what a programme needs.
What it does not give is a hands-on assessment or a hazard rating. The imagery is reviewed by a person, observations are recorded against a fixed defect list, and the condition value is an observation on the provider's own scale. Classification under HHSRS, and any decision about statutory timescales, stays with the provider's assessors.
Closing the gap one estate at a time
A provider does not need to fly its whole stock in a year. A practical programme looks like this.
- Find the unknowns. Query the stock condition database for roofs recorded as not inspected or ground-level only. Group them by estate and archetype.
- Rank by risk. Pre-1919 slate terraces with shared chimneys, flat-roofed blocks over stair cores, and anything with a recent leak or damp case go first.
- Fly by estate. Book sweep days where the properties are clustered, so the per-roof cost stays low. Send resident notices the week before.
- Import the rows. Insist on component rows per UPRN in your own import template: covering, structure, chimney, rainwater goods, flashings, fascia and soffit, rooflights, parapet, with material, observed condition, priority code and photo references. The data should land in the same database as the rest of the survey, not in a separate folder of PDFs.
- Work the P1s and P2s through the existing hazard and repairs process, and let the P3s and P4s inform the next programme.
- Keep the baseline. A re-survey in two or three years is far more valuable if it can be compared to a first pass.
What to tell the board
The board does not need to hear about class marks or shot lists. It needs three sentences. We know which roofs we have not seen. We have a costed plan to see them, ranked by risk, that does not involve scaffolding the stock. And the results go into the same system as the rest of our condition data, so the regulator can see our understanding is evidenced.
That is a better position than most providers are in today, and it is reachable within a single financial year.