Alternative data: satellite imagery, web scraping, credit card signals

Before Amazon reported Q3 2021 earnings, some hedge funds already knew the company had a strong quarter. Not from inside information. From satellite imagery of Amazon's fulfillment centers.

This is alternative data. And it's reshaping how sophisticated organizations use data for decisions.

What is alternative data?

Alternative data refers to information generated outside traditional business systems, not financial statements, not sales reports, not official surveys. It's data collected as a byproduct of digital activity, physical movement, or economic transactions.

The market for alternative data exceeded $14 billion in 2024 (Neudata) and keeps growing fast. Early adopters were hedge funds seeking investment alpha. Mainstream adopters now include retailers, manufacturers, and strategy teams looking for a competitive edgecompetitive edgeA lasting edge over competitors: a resource, capability or position they cannot easily replicate, letting a firm earn above-average returns over time.View full definition →.

The four major alternative data categories

1. Satellite and aerial imagery

Planet Labs and Spire Global operate fleets of small satellites providing frequent imagery of almost any location on Earth. (Orbital Insight, an earlier pioneer here, was acquired by Privateer in 2022 and folded into that company.) Applications:

• Counting cars in retail parking lots as a proxy for foot traffic before sales data lands
• Measuring factory utilization from smokestack activity
• Tracking oil storage levels from shadow analysis on floating-roof tanks
• Monitoring crop conditions for commodity forecasting

Walmart tracks Amazon's warehouse activity. Agricultural commodity traders watch crop conditions before USDA reports. The data is public (satellites observe publicly visible activity) and commercially available via API.

1. Credit card and transaction data

Vendors aggregate billions of anonymized card transactions to build spending indices by merchant, category, and geography. Bloomberg ran Second Measure as a transaction analytics product for years, letting Terminal subscribers track something like DoorDash's real revenue trend week by week before quarterly earnings. Bloomberg wound down the Second Measure product in 2024, so if you evaluated it in the past, check current options: Earnest Analytics, Facteus, and Consumer Edge all sell comparable card-panel spending data today.

The use case still holds. During Netflix's password-sharing crackdown, card-panel data showed subscriber and revenue shifts in near real time, months before Netflix confirmed the impact in its own earnings.

1. App usage and digital signals

data.ai and Apptopia track app downloads, usage frequency, session length, and rating trends for millions of apps. (data.ai, formerly App Annie, was acquired by Sensor Tower in 2024, and its data now sits under the Sensor Tower umbrella.) When a competitor's daily active users fall 15% month on month, that's a leading indicator of customer dissatisfaction, often visible six months before it shows up in churn rates.

1. Job posting and hiring data

Already covered as a CI tool. At scale, Revelio Labs tracks hiring velocity for public companies and sells it as a predictor of revenue growth. A firm hiring 50 data scientists in Q1 usually signals a major technology investment six to twelve months out.

Building an alternative data capability

For most organizations outside financial services, alternative data works at three levels:

Level 1, Consume: Buy access to one or two providers for specific use cases. Low investment, immediate value. The right starting point.

Level 2, Integrate: Build data pipelines that pull alternative data feeds into your existing analytics stack. The data becomes a persistent signal, not a one-off report.

Level 3, Produce: Generate your own alternative data as a byproduct of operations and monetize it. This is the Mastercard/Amazon level, where your operational data becomes a product for others.

Most organizations should start at Level 1 and reach Level 2 within about 18 months. Level 3 is a strategic bet for later.

Legal and ethical framework

Alternative data sits in a complex legal and ethical space:

• Terms of Service: Many websites prohibit scraping in their ToS. Violating ToS may not be illegal, but it can trigger IP bans and reputational damage.
• Privacy (GDPRGDPREU regulation governing how organizations collect, store and use personal data, with fines tied to global revenue for breaches.View full definition →/CCPA): Consumer-level transaction data must be anonymized and aggregated. Individual-level data from any EU source requires consent.
• Insider trading risk: Alternative data used in investment decisions falls under SEC oversight. Financial institutions need clear protocols for what data can inform trading decisions.

A CDO rolling out alternative data needs a governance framework before the first feed goes live. The governance cost is low. The cost of ignoring it is not.

Key Takeaways

• Alternative data is byproduct information (satellite imagery, card transactions, app usage, hiring signals) that often leads official reporting by months.
• The vendor market shifts fast: Second Measure was discontinued by Bloomberg in 2024, data.ai was absorbed by Sensor Tower in 2024, and Orbital Insight was acquired by Privateer. Verify a provider is still active before you build on it.
• For card-panel spending data today, look at Earnest Analytics, Facteus, or Consumer Edge.
• Start by consuming one or two feeds, then integrate the useful ones into your pipelines; producing and selling your own dataown dataData collected directly from your own customers and prospects through your own channels: your most reliable and privacy-compliant source.View full definition → is a later, bigger bet.
• Set up governance for ToS, privacy, and (for investors) insider-trading risk before the first feed goes live.