A hedge fund analyst counting cars in Walmart parking lots to estimate quarterly sales. An insurer measuring a wildfire's exact perimeter before a single adjuster reaches the site. A commodities trader watching crop stress in Ukraine's wheat fields two weeks before a harvest report confirms it. None of these people work in aerospace. They work in finance, insurance, and agriculture — and their edge comes from a data source that used to belong exclusively to governments and scientists: earth observation data.
Earth observation (EO) has quietly moved from a niche research tool into a commercial data layer that businesses query the way they'd query a database. The satellites are still in orbit doing what they've always done — capturing imagery, radar returns, and spectral signatures of the planet's surface. What's changed is who can buy that data, how fast it arrives, and what can be built on top of it.
What Earth Observation Data Actually Is
Earth observation is the practice of collecting information about the planet's physical, chemical, and biological systems using remote sensing instruments, most commonly satellites. The output isn't just "photos from space" — it's a range of data types, each suited to different business questions.
- Optical imagery: Standard visible-light and near-infrared photos, similar to what a camera captures, used for counting objects, tracking construction, or assessing vegetation health.
- Synthetic Aperture Radar (SAR): Radar-based imaging that works day or night and through cloud cover, useful for flood mapping, ship detection, and ground deformation monitoring.
- Multispectral and hyperspectral data: Imagery captured across dozens or hundreds of narrow wavelength bands, revealing details invisible to the human eye — crop stress, mineral composition, methane plumes, water quality.
- Thermal infrared: Surface temperature data used for wildfire detection, urban heat mapping, and industrial monitoring (like tracking activity at a power plant or refinery).
- GNSS and altimetry data: Precise positioning and elevation measurements used for sea-level monitoring, ice sheet tracking, and infrastructure surveying.
A single satellite pass generates raw pixels, not answers. The business value comes from a processing stack that sits on top: geometric correction, cloud removal, change detection, and increasingly, machine learning models trained to recognize specific objects or patterns — a parked car, a stressed corn plant, a new shipping container in a port.
The Supply Chain, From Orbit to Dashboard
It helps to think of EO as a layered stack rather than a single product:
| Layer | What happens | Example providers |
|---|---|---|
| Satellite operators | Launch and operate the sensors, capture raw data | Planet Labs, Maxar, ICEYE, Capella Space |
| Ground stations & downlink | Get raw data from orbit to earth | AWS Ground Station, KSAT, Leaf Space |
| Processing & analytics | Turn raw imagery into structured, queryable insight | Orbital Insight, Descartes Labs, Sinergise |
| Vertical applications | Package insight for a specific industry use case | Insurance loss estimators, ag-tech platforms, ESG monitors |
Most businesses never touch layers one and two. They buy access to layer three or four through an API or a subscription dashboard — the same way a company consumes weather data or credit scores without operating a weather balloon or running a credit bureau.
This layering matters because it determines who a business actually contracts with, and what they're paying for. A company buying "satellite data" rarely signs a contract with a satellite operator directly. More often, they're buying a subscription to an analytics platform that has already licensed raw imagery from one or more operators, blended it with other data sources, and wrapped it in an interface built for a specific job — a risk score, a change-detection alert, a yield forecast. Understanding which layer you're actually buying from helps set expectations: a raw imagery subscription requires in-house analytical capability, while a vertical application layer trades flexibility for immediacy.
Why This Shifted From Government Program to Business Tool
For decades, high-resolution earth observation was the province of national space agencies and defense programs — expensive to build, expensive to launch, and tightly controlled. Two structural shifts changed that.
The first is cost. Small satellites (smallsats and cubesats) built from increasingly standardized, mass-produced components can be launched for a fraction of what a traditional imaging satellite cost. Rideshare launch programs let operators put dozens of small satellites into orbit on a single rocket, splitting the cost across many customers. This made it economically viable to build large constellations rather than a handful of exquisite, expensive satellites.
The second is revisit frequency. One or two satellites can only photograph a given spot on Earth every so often. A constellation of dozens or hundreds of smaller satellites can image the same location multiple times a day. For a business question like "did construction on this factory actually start this week" or "how much oil is sitting in this storage tank right now," revisit frequency matters more than raw image resolution. Daily or near-daily imagery turns EO from an occasional research input into an operational monitoring feed.
Together, these shifts turned earth observation from a slow, expensive, government-run capability into a commercial data category with subscription pricing, self-serve APIs, and vertical-specific products — closer in feel to a SaaS analytics tool than a space program.
There's a third, less-discussed factor: declining launch costs and increasing launch cadence have made it economically rational to treat satellites as semi-disposable assets rather than irreplaceable, decade-long investments. A traditional government imaging satellite might be designed to operate for 10 to 15 years, with every component over-engineered against failure because a replacement launch was rare and expensive. Commercial smallsat operators increasingly plan for shorter operational lifespans and more frequent replacement launches, which changes the economics entirely — it becomes acceptable to iterate quickly on sensor technology, accept a higher failure rate on individual units, and treat the constellation as a whole as the durable asset rather than any single satellite.
Why It Matters Now
The practical reason earth observation data matters to businesses today isn't a single headline event — it's the accumulation of three trends reaching a usable threshold at the same time.
First, data volume and revisit rates have crossed a threshold where near-daily global coverage is achievable, not theoretical. Second, cloud infrastructure now makes it feasible to store and process petabytes of imagery without a business needing its own data center. Third, and most importantly, machine learning has matured enough to automate the labor-intensive part of the process: turning millions of pixels into a structured answer like "there are 412 cars in this parking lot" or "this field's vegetation index dropped 18% week over week." That automation is what makes EO usable by a business analyst rather than only by a remote sensing PhD.
The result is that earth observation has moved from being an input to research reports into being an input to operational decisions — pricing an insurance policy, timing a trade, routing a supply chain, or verifying a sustainability claim.
Practical Business Applications
Different industries use earth observation data for different jobs, but a few patterns repeat across sectors.
Insurance and Risk
Insurers use EO to assess property risk before underwriting a policy (roof condition, proximity to flood zones, vegetation encroachment near wildfire-prone structures) and to accelerate claims after a catastrophe. Instead of waiting for adjusters to physically reach a disaster zone — which can take days or weeks after a hurricane or wildfire — insurers can use pre- and post-event satellite imagery to estimate damage extent and prioritize claims processing within hours.
Agriculture
Multispectral imagery reveals crop health long before it's visible to the human eye, through indices like NDVI (Normalized Difference Vegetation Index) that measure plant vigor. Farmers and agribusinesses use this for irrigation planning, yield forecasting, and early pest or disease detection. Commodity traders use the same data, aggregated across regions, to anticipate harvest volumes ahead of official crop reports.
Finance and Investing
This is the most publicized use case: "alternative data" hedge funds analyzing satellite imagery of retail parking lots, oil storage tank fill levels (visible via shadow analysis), shipping traffic at ports, and mining or industrial activity to estimate company performance ahead of quarterly earnings releases.
Logistics and Supply Chain
Companies track port congestion, ship positions via SAR (which works through cloud cover, unlike optical imagery), and infrastructure construction progress. This helps logistics firms reroute shipments, and helps investors and procurement teams anticipate delays before they show up in official shipping data.
ESG and Sustainability Monitoring
Regulators, investors, and NGOs increasingly use EO to independently verify corporate sustainability claims — deforestation linked to supply chains, methane emissions from industrial facilities, or illegal mining activity — rather than relying solely on self-reported data.
Government and Infrastructure
Beyond the traditional defense and intelligence use cases, municipal and national governments use EO for urban planning, illegal construction detection, disaster response coordination, and infrastructure monitoring (pipelines, dams, power lines).
Energy and Natural Resources
Energy companies and their investors use thermal and optical imagery to monitor well pad activity, pipeline right-of-way encroachment, and flaring events at oil and gas facilities. Mining companies use multispectral and hyperspectral data for exploration — identifying mineral signatures across large tracts of land before committing to expensive ground surveys — and for monitoring tailings dam stability, where undetected structural failure can be catastrophic. Utilities use EO to track vegetation encroachment near power lines, a leading cause of wildfire ignition in many regions, allowing preventive trimming before a line fault occurs.
The common thread across all of these applications is that EO doesn't replace ground-level data collection — it prioritizes it. A utility still needs crews to trim vegetation; satellite data tells them which of thousands of miles of line to send crews to first. An insurer still needs adjusters; satellite data tells them which claims are likely most severe so those get attention first. The business value is triage and prioritization at a scale that ground-based inspection alone can't match.
What It Takes to Actually Use This Data
Buying access to satellite imagery is the easy part. Turning it into a repeatable business input requires a few things most companies underestimate.
- A clear, narrow use case. "We want satellite data" is not a project. "We want to detect new construction starts on competitor retail sites within 30 days of groundbreaking" is a project. EO is expensive and complex enough that vague ambitions rarely survive contact with a budget review.
- Historical baseline data. Detecting change requires knowing what "normal" looked like. A single snapshot rarely answers a business question; a time series usually does.
- Ground truth for validation. Machine learning models trained to detect objects or conditions from imagery need to be validated against real-world outcomes — actual crop yields, actual claims payouts, actual sales figures — or their outputs can't be trusted for decision-making.
- Integration into existing workflows. A dashboard nobody looks at doesn't change decisions. The highest-value EO deployments feed directly into systems teams already use: underwriting software, trading models, supply chain platforms.
- A realistic view of latency and cost. Even "daily" imagery has gaps from cloud cover, tasking priority, and processing time. High-resolution, low-latency tasking (ordering a fresh image of a specific location) is still a premium, per-request cost, not a flat-rate commodity.
- A build-versus-buy decision made deliberately, not by default. Some companies default to building in-house analytics pipelines because it feels like more control. For most businesses outside of quantitative finance or defense, buying a packaged vertical application is faster to value and cheaper over a multi-year horizon than hiring a remote sensing team to replicate what a vendor has already built and validated. The cases where building in-house makes sense are usually ones where the analysis is a core competitive differentiator, not a supporting input.
Limitations and Open Questions
Earth observation data is powerful, but it is not a frictionless oracle, and vendors have an incentive to understate its constraints.
- Cloud cover and weather. Optical satellites can't see through clouds. In persistently cloudy regions, useful optical imagery may be available far less often than marketed "daily revisit" numbers suggest. SAR solves this for many use cases but requires different (and often more specialized) analytical skills to interpret.
- Resolution limits and regulation. Very high-resolution commercial imagery (sub-30cm) is subject to government licensing restrictions in most countries, and the sharpest imagery isn't always commercially available at scale.
- Model accuracy and bias. Object detection and change detection models are trained on specific geographies and conditions. A model tuned to detect crop stress in Iowa cornfields may perform poorly on smallholder farms in a different climate and soil type — a real risk for any business assuming global consistency out of the box.
- Data fusion complexity. The most valuable insights typically come from combining EO with other data sources — weather, ground sensors, financial filings, shipping manifests. That fusion work is nontrivial engineering, not a checkbox feature.
- Cost at scale. Tasking a satellite for a specific, timely image of a specific location remains expensive relative to freely available lower-resolution archive imagery. Businesses building EO into a real-time operational process need to budget for this explicitly, not assume archive-tier pricing applies.
- Privacy and geopolitical sensitivity. As resolution and revisit rates improve, the line between commercial monitoring and surveillance gets blurrier, and different jurisdictions are still working out where that line sits.
What to Watch Next
A few developments will shape how usable and how valuable this data becomes over the next few years:
- Hyperspectral commercialization. Hyperspectral imaging, which captures far more spectral detail than standard multispectral sensors, is moving from research missions toward commercial constellations, with implications for mineral exploration, agriculture, and methane detection.
- On-orbit processing. Rather than downlinking raw data and processing it on the ground, some operators are pushing analytics onto the satellite itself, reducing latency between capture and insight — relevant for time-sensitive use cases like disaster response or maritime tracking.
- Consolidation and platform standardization. As with most young data industries, expect consolidation among smaller operators and analytics providers, along with growing pressure toward standardized APIs that make it easier for businesses to switch providers or blend data sources.
- Regulatory clarity on resolution and tasking. Governments are still refining rules on what resolution can be sold commercially and to whom, which will directly affect what's available to business customers versus government and defense buyers.
- Foundation models for geospatial data. Similar to large language models, general-purpose machine learning models trained specifically on satellite imagery are emerging, which could reduce the custom-model-building burden that currently makes EO adoption expensive for smaller companies.
FAQ
What is earth observation data used for in business?
It's used to monitor physical, real-world conditions that indicate business risk or opportunity — crop health for agriculture, property damage for insurance, parking lot activity for retail analysis, port congestion for logistics, and emissions or land-use change for ESG compliance and verification.
How is earth observation data different from regular satellite photos?
Regular photos are single optical images. Earth observation data includes multiple sensor types — optical, radar (SAR), thermal, and multispectral — often delivered as time series and processed through analytics models, rather than as one-off images meant for visual inspection.
Do I need a data science team to use satellite data?
For basic use cases, no — many vendors now sell packaged analytics (like a wildfire risk score or a crop health index) through simple dashboards or APIs. For custom, high-value use cases, you'll typically need in-house or contracted expertise to validate model outputs against ground truth and integrate the data into existing workflows.
How much does earth observation data cost?
Costs vary widely: broad, lower-resolution archive imagery can be inexpensive or even free from public sources, while high-resolution, freshly tasked imagery of a specific location can run from tens to thousands of dollars per image, depending on resolution, sensor type, and urgency.
Can satellites see through clouds?
Optical satellites cannot. Synthetic Aperture Radar (SAR) satellites can, because they use radar wavelengths rather than visible light, which is why SAR is heavily used for flood monitoring, maritime tracking, and other applications in persistently cloudy regions.
Is earth observation data legal to use for competitive intelligence?
Generally yes for publicly available commercial satellite imagery of non-restricted areas, since it doesn't involve trespassing or accessing private systems. However, resolution limits, export controls, and use-case-specific regulations (particularly involving foreign facilities or sensitive infrastructure) vary by country and should be checked against current law.
What industries benefit most from earth observation data today?
Insurance, agriculture, and commodities trading currently show the most mature, widely adopted use cases, followed closely by logistics, ESG monitoring, and infrastructure/government applications — with financial services broadly using EO as one input among several "alternative data" sources.
Teams evaluating whether earth observation data fits their workflow, or building a first pilot around it, can get hands-on help from Woyce Technologies.