Roof inspection of 70 housing units

A contractor's mistake repeated across an entire estate

Our client, a facility management company, was facing a significant problem. A contractor had carried out flashing work across a newly built housing estate, but the work was defective - and the error had been replicated across every unit. The buildings were all the exact same design, so the same poor workmanship appeared identically across all ~70 homes.

Some of the defects were at ground level and easy to spot. Others were on the roof pitch, completely out of sight from the ground. The client needed a clear, definitive picture of exactly what they were dealing with: which units were affected, how severe each defect was, and where the rectification team should start.

70 identical units, one replicated defect

The scale of the problem was simple but daunting. With roughly 70 units all sharing the same roof design, the defective flashing was repeated dozens of times. The client's immediate need was to identify every affected unit and understand the severity of each defect - but the roof-pitch defects could not be assessed from ground level.

The client also needed a practical, actionable output. A report that simply listed issues would not be enough. They needed something their rectification team could take straight into the field: a document that told them, unit by unit, what needed to be fixed.

A sub-250g drone purpose-built for residential work

To operate safely in a residential area - close to people, parked cars, and private property - we chose a sub-250g C0-class drone. This ultra-lightweight category of drone is designed to fly over people without putting them in danger. If it fell on someone, it would not cause injury. It was also exceptionally quiet: at the height we were flying, it was impossible to hear from the ground, keeping a very low profile throughout the day.

Before the campaign, the management company that runs the estate communicated with all residents. Everyone in the area was informed that the inspection was happening, why, and who would be carrying it out. This meant no surprises for anyone looking out their window.

On site, we began by capturing overview imagery of the area to orient ourselves back in the office. Then we moved into a systematic campaign: we would find the first one or two defects, refine the exact angle and distance needed to capture them clearly, and then replicate that capture position across the entire estate. Every unit was photographed at the same distance and angle, showing only the defect area - a piece of slate and a piece of gutter, nothing more.

A critical part of the workflow was orientation. When you are taking the same image dozens of times, you need a reliable system for knowing exactly where you are. We used extensive orientation imagery alongside the defect captures to ensure every image could be matched to its correct unit without any ambiguity.

A GDPR-first approach to residential drone inspection

Operating a drone in a residential area means privacy is paramount. We designed our capture workflow specifically to eliminate any risk of capturing personal data. Every image was tightly framed on the defect itself - so tightly that the only things visible in any given photo were a section of slate and a piece of gutter. No cars, no people, no windows, no details of private property beyond the defect being inspected.

After capture, all imagery was reviewed in the field. Anything that should not have been there was deleted immediately. The result was a dataset that contained no personal data whatsoever - only the defects the client needed to assess.

Severity ratings that prioritised the rectification work

Every defect was assessed against a standard five-point severity scale used across the industry. At the low end, severity zero means there is no problem but something worth noting - for example, a mark on a wall that looks like a crack from a distance but is actually harmless. At the top end, severity five means a danger to human life - a loose slate over a front door, for instance. Severity four is a serious structural issue such as a large hole in the roof.

Each affected unit received an image, a defect category, a written description for context, and a severity level. Units where the defect was not present were marked as good. The result was a large report that listed every single unit in the estate, with the imagery and assessment alongside each entry.

For the client, this report was more than a record of findings. It became a complete rectification picking list. Their team could take it into the field and work through it unit by unit: fix the defects on the units flagged as affected, verify the work, and know with certainty when every issue had been resolved.

What the client received