How to choose an alternative data provider

alternative data providers all sell the same promise: a read on a company's performance before it shows up in a quarterly filing. What they actually sell is very different, because the promise is only as good as the raw signal underneath it. A foot-traffic estimate from mobile location pings, a revenue estimate from card-panel data, and a hiring signal scraped from a job board are three different kinds of evidence, each with its own coverage gaps and its own way of being wrong. Buying by outcome ("we want a revenue estimate") without asking about the underlying source is how research teams end up trusting a number that quietly stopped being representative months ago.

Start with the underlying data source, not the output

Every provider in this category turns raw signal into an investor-facing metric, but the raw signal is the thing to scrutinize.

• Mobile location data. Advan Research licenses anonymized device location pings from apps and aggregators and turns them into store visit counts, dwell time and cross-visitation patterns — a real-world foot-traffic proxy for same-store sales.
• Card transaction panels. Bloomberg Second Measure and Earnest Analytics both build on anonymized, aggregated US credit and debit card panels to estimate company-level revenue growth, market share and customer retention. They compete on panel composition and sector coverage rather than on a fundamentally different data source.
• Satellite and geospatial imagery. RS Metrics uses satellite and aerial imagery analyzed with computer vision to measure physical, real-world activity — parking-lot car counts, commodity stockpile volumes, construction activity — independent of any company disclosure.
• Web-scraped content. Thinknum continuously monitors company websites, job boards, app stores and marketplaces, turning changes over time into hiring, pricing and product-listing signals. YipitData combines web-scraped data with licensed card-panel data across several sectors.
• Public-record aggregation. Quiver Quantitative is the outlier here: its data — congressional trading disclosures, lobbying spend, government contracts, patent filings, social sentiment — is public record, just scattered across many sources, rather than proprietary transaction or imagery data. That makes it a much lighter-weight, lower-cost entry point than the rest of the category.

Match the data source to the question you're asking

A location-data provider tells you about physical foot traffic — useful for a retailer, useless for a pure e-commerce or B2B company. A card-panel provider tells you about consumer spend — strong for retail, restaurants, travel and subscription businesses, weaker for anything not paid for on a consumer card. A web-scraping provider tells you about a company's public digital footprint — hiring, pricing, product listings — which is a leading indicator of intent, not a direct revenue read. Before you evaluate any vendor, write down the specific company-level question you need answered, and check whether that company's business model actually generates the kind of signal the provider collects.

Coverage and company type

Public-record and web-scraped providers (Quiver Quantitative, Thinknum) can cover private companies as well as public ones, since they are not dependent on a card network or exchange listing. Card-panel and location-data providers are effectively limited to businesses with a meaningful US consumer-facing footprint. If your research covers private-company diligence, or non-US or non-consumer businesses, ask directly what share of your target list a vendor's panel or scrape actually reaches — coverage gaps are the most common reason an alternative-data subscription goes unused after the first quarter.

Delivery model and who it's built for

Most of this category is institutional: Advan Research, Earnest Analytics, RS Metrics, Bloomberg Second Measure, Thinknum and YipitData all sell through negotiated institutional licensing to hedge funds, private equity and corporate strategy teams, with data delivered through feeds, dashboards and often analyst-written reports rather than a self-service checkout. Quiver Quantitative is the exception, with a genuine free tier and self-serve paid subscriptions aimed at retail and prosumer investors, alongside separate bulk licensing for institutional buyers. If you are not an institutional buyer, Quiver is realistically the only self-serve entry point in this specific set.

How pricing scales

Nearly every provider here prices by custom quote rather than publishing tiers, reflecting how this data is actually sold: multi-year institutional contracts scoped to specific sectors, tickers or datasets, not a per-seat SaaS subscription. Expect the quote to depend on breadth of coverage (how many tickers or categories), depth of history, delivery format (dashboard access vs. raw feeds vs. API), and whether you need analyst support layered on top of the raw numbers. Budget for a sales cycle, not a self-serve signup, with Quiver Quantitative again the exception at the low end.

A shortlist by situation

• You cover consumer retail, restaurants or subscription businesses and want a revenue proxy. Compare Bloomberg Second Measure and Earnest Analytics on panel size and sector depth for your specific names.
• You need a physical-world signal independent of any transaction data — commodities, industrials, real estate. RS Metrics's satellite-based approach is built for exactly this.
• You want foot-traffic and cross-visitation patterns for specific retail locations. Advan Research's mobile location data is the closest fit.
• You track hiring, product launches or pricing moves as leading indicators, including for private companies. Thinknum's web-scraped datasets cover this ground.
• You want broad public- and private-company estimates with analyst-written context. YipitData pairs web-scraped and card-panel data with sector research reports.
• You are a retail or prosumer investor, or want a low-cost way to try alternative data before an institutional contract. Quiver Quantitative is the only provider here with a real free tier.

Questions to ask vendors

1. What is the underlying raw data source, and how is it collected and licensed?
2. What percentage of our specific coverage universe (tickers, sectors, private names) does your panel or scrape actually reach?
3. How far back does the historical time series go, and has the underlying panel composition changed materially over that period?
4. How do you handle survivorship and panel-composition bias as merchants or apps enter and leave the sample?
5. Is data delivered as a finished metric, a raw feed, or both — and what does it cost to get the raw feed for our own modeling?
6. Can we backtest the signal against a set of companies where we already know the actual reported outcome?

Common mistakes

• Treating a derived metric (e.g. "estimated revenue growth") as ground truth rather than an estimate with its own error bars and panel-composition risk.
• Buying location or card-panel data for a company with limited US consumer-facing exposure, where the panel simply does not see most of the business.
• Skipping a backtest against companies with known, already-reported outcomes before committing to a multi-year contract.
• Assuming a "free" public-record tool like Quiver Quantitative has no maintenance cost — someone still has to validate and monitor the underlying sources it aggregates.

Related reading

See Advan Research vs RS Metrics for two very different physical-world data sources, and Bloomberg Second Measure vs YipitData for two card-panel-based revenue-estimate providers. Every tool in this category: every tool in this category.