Satellite and aerial roof measurements are accurate enough to bid and order materials from when the imagery is current, the roof edges are clearly visible, and pitch is derived rather than assumed. Accuracy degrades with stale imagery, heavy tree cover, low sun angle, complex or stacked geometry, and recent construction. A good report discloses its imagery date and a confidence signal so you know when to verify in the field.
- →Accuracy is a function of imagery quality and roof complexity, not brand
- →Imagery date is the single most important disclosure on a report
- →Tree cover, low sun angle, and stacked geometry are the usual culprits
- →Derived pitch beats assumed pitch; a report that hides which one it used is a red flag
- →Cross-checking two independent derivations catches most large errors
"How accurate is it?" is the right question and usually gets the wrong answer — a single percentage quoted with no conditions attached. Accuracy is not a property of a product; it is a property of a specific roof, on a specific date, in specific imagery. Here is how to reason about it.
The five things that actually drive accuracy
| Factor | Effect | What to look for |
|---|---|---|
| Imagery date | A roof rebuilt, added to, or re-decked since capture will not match | The capture date printed on the report |
| Resolution | Edge placement error scales directly with ground sample distance | Crisp ridge and eave lines in the diagram |
| Tree cover | Obscured facets get interpolated rather than measured | Coverage notes or a flagged facet |
| Sun angle and shadow | Shadows on a facet edge shift where the boundary is placed | Heavy shadow in the aerial view |
| Geometry complexity | Dormers, stacked planes, and low-slope tie-ins are the hardest to resolve | Facet count versus what you see on the elevation photos |
Area versus pitch versus lengths
A report contains three classes of number and they do not fail in the same way.
- Footprint area is the most robust. Outlining a roof from overhead is a well-constrained problem when the edges are visible.
- Pitch is where reports diverge. Pitch can be derived from elevation data or assumed from neighborhood norms. Derived pitch is a measurement; assumed pitch is a guess that then multiplies through your entire squares figure.
- Linear quantities — ridge, hip, valley, eave, rake — depend on getting the facet breakdown right. A report that merges two planes into one loses the valley between them.
Slope multiplies footprint area into surface area. A pitch that is off by 2/12 on a large roof moves your squares by more than most people expect — which is why a report that tells you it derived pitch per facet is worth more than one that reports a single roof-wide number.
When you should not bid from a report alone
- The imagery predates a known storm, addition, or re-roof at that address
- Mature tree canopy covers a meaningful share of the roof
- The structure is new construction that may not appear in imagery yet
- The roof has stacked or interior planes you cannot see from overhead
- There are multiple structures on the parcel and the report may have measured the wrong one
- The report's confidence signal is low, or it does not provide one
None of these mean skip the report. They mean treat it as the base and verify the specific thing that is uncertain — one facet, one pitch, one tie-in — rather than re-measuring the whole roof by hand.
How to read a report critically in two minutes
- Check the address and the aerial view — is that the building you bid?
- Check the imagery capture date against what you know about the property
- Count the facets in the diagram against the elevations you can see
- Look at pitch: is it reported per facet or one number for the whole roof?
- Sanity-check total squares against footprint and pitch; a steep roof should carry meaningfully more surface than footprint
- Read the ridge and valley lengths — a cut-up roof with almost no valley is a sign a plane was merged
- Read the notes and caveats section before you price anything
Why cross-checking matters
The most reliable accuracy safeguard is not a better single measurement — it is two independent derivations compared against each other. When two methods agree closely, confidence is high. When they diverge materially, something about that roof is hard, and that is exactly the case where you want to be told rather than handed a confident-looking single number.
RoofGenius reports disclose the imagery date, include a confidence score, and flag disagreement between cross-checks above a threshold. When a property is too uncertain to measure responsibly, we say so instead of selling a number we would not bid from.
Accuracy for insurance work specifically
For claims, defensibility matters as much as precision. A desk reviewer cannot re-measure your roof; they can only check whether your quantities have a documented source. A report with per-facet area, per-facet pitch, labeled linear quantities, and a diagram gives them that. A number typed into a spreadsheet does not, regardless of how accurate it actually is.
| Use case | What accuracy needs to support | Report features that matter |
|---|---|---|
| First-call ballpark | A price range | Area and rough pitch |
| Retail bid | Material order and labor pricing | Per-facet area, pitch, waste factor |
| Material order | Bundle counts and accessory lengths | Ridge, hip, valley, eave, rake, starter |
| Insurance supplement | A quantity a reviewer can verify | Diagrams, per-facet table, imagery date |
| Complex or commercial | Plane-by-plane detail | Facet breakdown plus field verification |
Field verification without re-measuring
When you are on site anyway, spot-check three things and you will catch nearly every material error: one eave length with a tape, one pitch with a level or gauge, and a quick count of penetrations and facets against the diagram. If those three agree with the report, the rest almost always does too.
See a real report before you rely on one
Read a full sample measurement report — including the confidence disclosure and caveats — then order a $10 report on an address you already know well and compare it against your own numbers. That is the only accuracy test that means anything for your market.