Key takeaways
- You are not buying a flight, you are buying the report. Ask for a sample of a real past report before you award the work, not a marketing page.
- A defensible finding has five parts: where it is, what was seen, how confident the operator is, what it means, and what should happen next. Findings missing the confidence level are the ones that cause arguments.
- Page count is not quality. The most common weak report is a large contact sheet of good photographs with no locations, no severity and no conclusion.
- The report has to survive being read a year later by someone who was not on site, because that is usually when it matters.
Almost every conversation about hiring a drone operator is about the flight. Which aircraft, which sensor, which certificate, how quickly they can get to site. Very little of it is about the thing the client actually receives, which is a document.
That is backwards. The flight lasts an afternoon. The report is what gets attached to a work order, argued over in a claim, handed to an engineer, or pulled up eighteen months later when somebody asks whether that crack was there before. The quality of the report, not the quality of the aircraft, is what decides whether the money was well spent.
Here is what separates a report you can act on from one that only looks like work.
What a defensible finding actually contains
Everything else in this article follows from one idea. A report is a list of findings, and a finding is only useful if it answers five questions without the reader having to ask.
- Where. Not "north face" but an identifier a crew can walk to: a structure number, a chainage, a grid reference, a panel or string ID, a marked location on a site plan or orthomosaic. If two people could disagree about which spot is meant, the location is not specific enough.
- What was observed. A description of what is visible, kept separate from the interpretation of it. "Rust staining approximately 300 mm below the upper weld seam" is an observation. "Weld failure" is a conclusion, and it belongs on its own line so the reader can see the reasoning.
- How confident. Confirmed, probable, or requires verification. This is the part most often missing and the part that causes the most trouble later, because a hedge buried in prose reads as certainty when someone skims it.
- What it means. Severity, ranked consistently across the whole report, with the ranking defined somewhere in the document rather than left to the reader's judgement.
- What happens next. A recommended action and a rough urgency. Monitor at next inspection, verify by hand, repair, or escalate.
A report where every finding has those five parts can be turned into a work plan on the day it arrives. A report missing the confidence level and the recommended action becomes a meeting instead.
The tells of a report built to look thorough
Weak reports are rarely empty. They are usually full, which is what makes them hard to spot until you try to use one.
The contact sheet
Two hundred sharp photographs, arranged neatly, with captions like "east elevation" and no findings section. This is raw capture presented as a deliverable. The photographs may be excellent, but the analysis has been quietly transferred to the client, who now has to look through all of them and decide what matters. If you wanted the images, that is fine. If you wanted an inspection, you did not get one.
Findings with no location
A defect described well enough to be alarming and not well enough to be found. This forces a second site visit by someone with the report in hand, trying to match a photograph to a wall. Any finding that cannot be navigated to is close to worthless operationally.
Uniform severity
Forty findings, all marked medium. It means nothing was prioritised. The reader still has to triage the whole list, which was the job they were paying to have done.
Conclusions with no visible evidence
A statement of condition with no image, no measurement and no sensor data behind it. Even where the operator is right, nothing in the report supports the claim if it is ever challenged.
No statement of what was not covered
Every flight has gaps. Areas obscured by equipment, a face that could not be approached safely, a section cut short by wind or light, anything below grade or inside an enclosure. A report that never mentions a limitation is implying complete coverage that no aerial survey actually delivers. We wrote about the physical limits in what a drone survey cannot tell you, and the same principle applies to the document: the exclusions belong in writing.
Thermal findings with no capture conditions
A thermal anomaly is only interpretable alongside the conditions it was captured in. Time of day, ambient temperature, wind, cloud and irradiance all change what the imagery shows, and a radiometric image without them is a picture rather than a measurement. Our piece on the conditions that make or break a thermal inspection covers why. If a report contains thermal findings and no conditions log, the findings cannot be defended.
The parts nobody reads until they matter
The front and back of the report carry the information that makes it hold up over time.
- Capture metadata. Date, times, aircraft and payload, and the weather and light on the day. This is what lets a future reader judge the imagery rather than guess at it.
- Scope, stated as scope. What was asked for, what was covered, what was excluded and why.
- Method. Altitude, overlap, whether ground control was used, and what software processed the data. Enough that the survey could be repeated the same way next year and compared like for like.
- Who flew it and under what authority. Pilot certification level and, where the site required it, the operational authorisation the flight was conducted under.
- Data and file inventory. What was delivered alongside the PDF, in what formats, and where it lives.
That last one matters more than it sounds. A report that references a point cloud nobody can locate a year later is a report that has lost half its value.
Whether it is repeatable
Most inspection programmes are not one-time. The value compounds when this year's survey can be laid against last year's, so the question is whether the report was built to be compared.
That requires a stable finding numbering scheme, consistent location references between visits, a flight plan recorded well enough to be re-flown, and a severity scale that does not quietly change between reports. Ask an operator directly what a second inspection of the same asset looks like and how the two are compared. An operator running a repeat programme will answer immediately. One who has only ever done one-off jobs will describe a fresh survey, which means change detection is being done by eye.
Reading a sample report in ten minutes
Ask for a redacted sample of a real past report before awarding the work. Then run these checks.
- Pick any finding at random. Could a crew who has never been to the site walk to it using only what is on the page?
- Does that finding state a confidence level, or does it read as certain?
- Is the severity scale defined anywhere, and do the assigned severities look like they used it?
- Is there a section that says what was not covered?
- If there are thermal findings, are the capture conditions recorded?
- Could you hand this to a contractor and get a quote back without a phone call?
The last one is the real test. A report that produces a priced remediation quote with no further conversation has done its job. A report that produces a meeting has moved the work back onto your desk.
Write it into the scope, not into hope
The straightforward fix is to specify the deliverable in the request rather than assume it. Name the five parts a finding must contain, ask for the severity scale up front, require a limitations section, and require the conditions log if thermal is in scope. Operators who work this way will not blink. Those who do not will tell you so by how they respond.
The same logic applies across the whole document, and the clauses worth writing down are collected in what to put in a drone survey RFP. If you want to go further back in the process, how to vet a commercial drone contractor in Alberta covers the compliance and insurance side of the same decision.
If you would like to see how we structure an inspection deliverable, ask and we will walk you through a redacted one. You can see the services it comes out of on our drone inspection services page.