Construction progress with drones: three KPIs that make projects steerable

Many sites collect data — orthophotos, point clouds, ground photos. Few steer with it. The difference is not the aircraft alone, but the metrics that site and project management can read every week.

This article shows three KPIs that work in practice — and the data pipeline they need.

Why overall “percent complete” often misleads

A single “72% done” says little when:

• sections have different criticality,
• subcontractors work out of sync,
• risk sits in a narrow corridor.

You need section reference and a cadence (e.g. weekly), or the survey stays nice documentation without control value.

KPI 1 — Progress per section

Question: How far is section A / B / C — not only the whole project?

How you measure it

• Define sections along alignment, building or trade.
• Recapture the same corridor (similar height, similar GSD).
• Derive progress from comparisons (area, volume, visible build state) — not gut feel.

Benefit: Site management prioritises crews and material where the section really lags.

KPI 2 — Target vs actual (early warning)

Question: Where does the as-built state deviate from plan or from the last epoch — early enough to correct?

How you measure it

• Reference: design, alignment or epoch t0
• Delta: position, height, volume (cut/fill), open areas
• Thresholds: e.g. “amber” from X cm / Y m³, “red” above

Benefit: Fixing small deviations is cheaper than large change orders. Drone data makes deviation visible and shareable — for client and contractor alike.

KPI 3 — Risk density per section

Question: Where do uncertainties cluster (access, subcontractors, weather windows, adjacent infrastructure)?

How you measure it

• Tag annotations and open items per section.
• Count repeats of the same risk type across epochs.
• Combine with progress: high lag + high risk density = escalate.

Benefit: Resources go where delay and risk meet — not only where shouting is loudest.

The pipeline behind the KPIs

Without a clean chain, metrics stay cosmetics:

1. Capture — recurring flights, documented parameters
2. Process — ortho / point cloud / derivatives in one system
3. Compare — epoch vs epoch or design vs as-built
4. Escalate — clear roles: who reads KPIs, who decides, who closes items

GDDC connects flight operations, processing and — where useful — platform access (including geospatial / RailSight-style logic for corridors and sections).

Practical cadence (recommendation)

Weather and permits belong in the same plan — otherwise the cadence breaks in the first bad-weather week.

Next step

If you already fly but without steering effect: start with two sections and the three KPIs above. Scale only once the weekly rhythm works.

Contact · Geospatial · Services · All news