What is alternative data?
What is alternative data? Podcasts, newsletters and the other sources funds use
A plain guide to alternative data: what counts, the main types, how investment teams use it and how to judge whether a dataset is usable.
featured in
Alternative data is information from outside the traditional financial sources (company filings, market prices and official economic releases) that investors and analysts use to understand companies, sectors and the economy. Card transactions, app downloads, shipping records, satellite images, web traffic and what people say in public all count.
The term comes from investment management, where hedge funds started buying these datasets to see trends before they reached reported numbers. It's now used by asset managers, private equity, corporate strategy teams and lenders.
Traditional data vs alternative data
| Traditional data | Alternative data | |
|---|---|---|
Examples | Filings, earnings releases, prices, official statistics | Transactions, web and app data, location, imagery, public text and audio |
Who publishes it | Companies, exchanges, governments | Third parties, platforms, the public |
Timing | Scheduled (quarterly, monthly) | Often daily or continuous |
Structure | Standardized | Varies; usually needs cleaning and mapping to companies |
Availability | Same for everyone | Bought or collected; coverage differs by vendor |
Common types of alternative data
Transaction data
Anonymized card and receipt data used to estimate company revenue before it's reported.
Web and app data
Traffic, downloads, reviews, prices and job postings collected from websites and app stores.
Location and imagery
Foot traffic from mobile devices, and satellite or aerial images of car parks, ports and farmland.
Supply chain data
Shipping manifests, freight movements and customs records.
Social and public text
Posts, reviews and forum discussions, often reduced to counts or sentiment.
Spoken and long-form media
Podcast transcripts and newsletters: what executives, analysts and commentators say about companies, attributed to a speaker and a time.
Podcasts and newsletters as alternative data
Executives appear on podcasts between earnings calls, often for an hour at a time. Investors, analysts and former employees discuss the same companies on business shows. Specialists write long Substack posts about sectors they know well, and Bluesky reacts to news within minutes. Together these are a large, public, on-the-record record of what people who know a company are saying about it.
Until it's transcribed and structured, that material is hard to use. As data it becomes rows: the company (by ticker and CIK), the passage, the speaker with title and company, the show or newsletter, and the timestamps for when it was published and when it became available. From those rows you can count mentions per day, rank companies by change in attention, find every podcast appearance by a company's executives, and read the exact words behind a spike.
How investment teams use alternative data
Idea generation
Find companies whose activity or attention is changing before it shows up in reported results.
Thesis checks
Test an investment view against independent evidence, such as what customers, suppliers or executives say.
Event detection
Spot sudden changes (a spike in mentions, a drop in traffic) and investigate the cause.
Risk monitoring
Track narratives, controversies and management commentary for holdings.
Model features
Feed cleaned, point-in-time series into quantitative models or internal AI tools.
How to evaluate an alternative dataset
Six questions to ask any provider
Coverage
Which companies, sources and regions are covered, and how was the list chosen? Is coverage stable over time, or did it grow in a way that could bias a backtest?
History
How far back does the data go, and was history collected live or reconstructed later?
Point in time
Does each record show when it became available, as well as when the event happened? Without that, backtests use information you couldn't have had.
Mapping to companies
Are records keyed to tickers and permanent identifiers such as CIK, and how are ambiguous names handled?
Matching quality
How is a mention or event attributed to a company, and can you see the confidence and the underlying text?
Sourcing and compliance
Is the data public, on the record and collected lawfully? Can each record be traced to its source?
Challenges with alternative data
- Noise: most raw data says nothing about a company's results. Signals have to be tested, not assumed.
- Look-ahead bias: using publish dates instead of availability dates makes backtests look better than reality.
- Entity ambiguity: product names, common words and shared names (Apple, Shell, Target) need careful matching.
- Survivorship and coverage drift: sources added or dropped over time change what a series measures.
- Legal and ethical sourcing: personal data, terms of service and material non-public information all need review.
Alternative data in lending
The same term is used in credit: lenders use rent, utility and bank-account data to score applicants who have little credit history. That meaning is separate from the investment use described in this guide.
FAQ
Common questions
Yes, when it's collected lawfully and doesn't contain material non-public information or personal data used without a legal basis. Investment firms usually review each dataset's sourcing through a due-diligence questionnaire before buying.
Card transactions, web traffic, app downloads, job postings, satellite images, shipping records, social media posts, and podcast and newsletter content about companies.
Hedge funds were the first large buyers. Asset managers, private equity firms, investment banks, corporate strategy teams and lenders now use it too.
Usually as files or through an API, updated daily or weekly. Good datasets include identifiers for joining to other data and timestamps for when each record became available.
Big data describes volume and processing. Alternative data describes the source: anything outside traditional financial data. Many alternative datasets are large, but size isn't what makes them alternative.
Contact us
Have a question or a custom use case?
Tell us what you're trying to monitor or search across, and we'll get back to you within a business day.
Add podcasts and newsletters to your research
Subalytics turns podcast, Substack and Bluesky content into ticker-keyed, point-in-time data for financial institutions.