DataRevenue.io

How to monetize your company's data

Short answer: To monetize company data, you turn it into products that other organizations pay to use, without selling the underlying data or exposing personal information. The main models are licensing de-identified datasets, selling data products or feeds, licensing AI training data, publishing insights and benchmark reports, and using data-rich assets to support financing. The process runs in seven steps: inventory, rights and consent review, valuation, de-identification, packaging, pricing, and distribution through marketplaces or direct buyers. Many companies run the program through a separate data company that licenses the data from the operating business, which keeps ownership.

Last updated October 10, 2026. Published by Data Revenue.

Key takeaways

  • Gartner defines data monetization as "the process of using data to obtain quantifiable economic benefit" (Gartner Glossary, accessed October 2026).
  • External monetization usually means licensing, not selling. The company keeps ownership and grants limited rights of use.
  • Buyers should receive aggregated, de-identified products under contracts that ban re-identification.
  • AI developers, investors and other companies are already paying for data. Reddit disclosed $203.0 million of data licensing contracts in 2024.
  • Cloud data marketplaces from Snowflake, AWS and Databricks let companies list and charge for data products.
  • Value is never guaranteed. Start with an indicative valuation before investing in a program.

What data monetization means

Gartner's glossary splits data monetization into two kinds (Gartner Glossary, accessed October 2026):

  • Indirect (internal): using data to improve your own decisions and performance, such as pricing, retention or cost reduction.
  • Direct (external): licensing data, selling information products and services, or exchanging data with partners for better terms.

This guide focuses on direct monetization: earning new revenue from organizations outside your business.

It is still early for most companies. In a McKinsey global survey, 41 percent of respondents said their companies had begun to monetize data, and most of those had started within the previous two years (McKinsey, December 1, 2017). That survey is now several years old, but it remains one of the few published benchmarks on adoption.

The five main ways to monetize company data

1. Licensing de-identified datasets

You license a defined dataset, such as transaction histories, product usage or operational records, to a buyer for a set term and purpose. The buyer gets the right to use it, not ownership. This is the most common starting point because it requires the least new technology.

Typical buyers: investment research firms, insurers, consultancies, researchers and companies in adjacent industries.

2. Data products and feeds

Instead of a one-time file, you deliver a recurring product: a daily or weekly feed, an API, or a curated table that updates automatically. Recurring feeds support subscription pricing. Reddit, for example, describes charging customers for access to real-time data streams through its data APIs above certain usage limits (Reddit, Inc. Form S-1, February 22, 2024).

3. AI training data

AI developers license large volumes of lawfully sourced text, images, audio, video and domain-specific records. Public examples include:

  • Reddit: data licensing arrangements entered in January 2024 with an aggregate contract value of $203.0 million and terms of two to three years (TechCrunch, February 22, 2024). Reuters reported that the Google contract alone is worth about $60 million per year, according to one source (Reuters via Yahoo, February 22, 2024).
  • Shutterstock: a six-year agreement giving OpenAI a license to training data from its image, video and music libraries and associated metadata (Shutterstock press release, July 11, 2023).
  • The Associated Press: OpenAI licensed part of AP's text archive, with access to stories going back to 1985. Financial terms were not disclosed (PBS NewsHour, July 13, 2023).
  • News Corp: a multi-year deal that The Wall Street Journal reported could be worth more than $250 million over five years (Variety via Yahoo, May 22, 2024).

These are large content owners. Most companies' AI data opportunities are narrower: specialist, domain-specific data that general-purpose web data cannot replace.

4. Insights and benchmark reports

Rather than releasing records, you sell conclusions drawn from them: industry benchmarks, price indices, demand trends or performance comparisons. Because only aggregated statistics leave the business, this model carries the lowest privacy risk, although it usually earns less per customer than raw data licensing.

5. Data-backed financing

Lenders have started to treat data-rich intangible assets as collateral. Law firm Ropes & Gray describes data asset financing as a logical extension of the asset-backed lending that grew during the pandemic, and sets out a framework covering how to define data assets, value them, apply data regulation and structure the deal. Its key observation:

"...data is an asset type that almost all borrowers accrue as part of the ordinary course of business." (Ropes & Gray, June 2023)

The clearest precedents are airline loyalty programs, which combine customer relationships, partner contracts and member data. United raised $6.8 billion secured by its MileagePlus program in 2020 (United Airlines Holdings Form 8-K, June 26, 2020), and American announced $10.0 billion of AAdvantage financing in March 2021 (American Airlines Group, March 10, 2021). Financing against data is not yet routine for mid-sized companies, but a documented valuation and licensing history make the conversation with lenders possible.

How to monetize your data: seven steps

Step 1: Inventory your data

List every significant data source: transactions, customers, products, operations, sensors, content, support interactions and partner feeds. For each, record approximate volume, years of history, update frequency, format and owner.

Step 2: Review rights and consent

Check what you are allowed to do. Review privacy notices, customer and supplier contracts, platform terms, consent records and any data you received from third parties. This step decides what can be licensed, for which purposes and in which regions. Get legal advice for anything unclear.

Step 3: Value the data

Estimate what the data could be worth before investing in a program. A professional valuation considers the cost, income and market approaches; Data Revenue's valuations follow IVS 210 methodology. To check whether it is worth going further, you can see an indicative value range in about five minutes with the free 15-question Data Snapshot.

Step 4: De-identify and aggregate

Remove or transform direct identifiers, aggregate where possible, and test the result against re-identification risk using the standard that applies to the data (see the privacy section below).

Step 5: Package the product

Turn raw tables into something a buyer can use: a documented schema, a data dictionary, sample files, a methodology note, refresh schedule and quality metrics. Good packaging shortens a buyer's evaluation.

Step 6: Set pricing and terms

Common structures include annual subscriptions, per-use or per-record fees, tiered access, revenue shares and minimum guarantees. Key license terms include permitted use, term, territory, exclusivity, attribution, audit rights, security requirements and an explicit ban on re-identification.

Step 7: Distribute to buyers

  • Marketplaces: Snowflake Marketplace lets providers publish data products and charge for them through paid listings (Snowflake documentation, accessed October 2026). AWS Data Exchange supports subscription pricing for terms of 1 to 36 months, plus private offers with custom pricing (AWS Data Exchange FAQs, accessed October 2026). Databricks Marketplace is an open marketplace for datasets, notebooks, dashboards and machine learning models, powered by Delta Sharing (Databricks, June 28, 2022).
  • Direct sales: larger and more specialised deals, especially AI training licenses and exclusive arrangements, are usually negotiated directly with buyers.

Privacy and compliance

Responsible data monetization is built on one principle: buyers receive aggregated, de-identified data products, never your customers' identities. The rules differ by law and data type.

GDPR (European Union)

Recital 26 of the GDPR says data protection principles do not apply to anonymous information, but pseudonymised data that could be attributed to a person using additional information is still personal data (Regulation (EU) 2016/679, Recital 26). In practice, replacing names with codes is not enough to take data outside the GDPR.

CCPA and CPRA (California)

Under the California Consumer Privacy Act as amended, information counts as "deidentified" only if the business takes reasonable measures to prevent it being linked to a consumer or household, publicly commits to keep it in deidentified form and not to re-identify it, and contractually obligates any recipients to comply (California Civil Code section 1798.140(m)). That last requirement is why data licenses must flow the re-identification ban down to buyers.

HIPAA (US health data)

Health information is de-identified under HIPAA when it does not identify an individual and there is no reasonable basis to believe it could. A covered entity can establish this in one of two ways: an expert determination that the risk of identification is very small, or the safe harbor method of removing 18 listed categories of identifiers (45 CFR 164.514(a) and (b), via Cornell Legal Information Institute).

Contract protections every license should include

  • A prohibition on re-identifying, or attempting to re-identify, any individual or household
  • A prohibition on combining the data with other data to identify people
  • Limits on permitted use, onward sharing and resale
  • Security requirements, audit rights and breach notification
  • Deletion or return of data at the end of the term

The separate data company structure (Data HoldCo)

Many companies keep data monetization out of the operating business by creating a separate data company, often called a Data HoldCo.

How it works

  1. The operating company keeps ownership of its data and its customer relationships.
  2. It licenses the data to the data company under a written intercompany license with defined permitted uses.
  3. The data company builds and sells the products. It de-identifies, packages and licenses data products to outside buyers.
  4. Revenue flows back to the operating company under the license, with the operating company keeping the majority of the revenue.

Why companies use it

  • Focus: the operating team keeps running the core business while the data company handles buyers, contracts and compliance.
  • Risk separation: licensing obligations sit in a separate entity.
  • Clarity for investors and lenders: the data business has its own contracts, revenue and records, which makes it easier to value and, potentially, to finance.

A well-known precedent

Separating a data-rich asset into its own entity while keeping ownership has a high-profile precedent. United's 2020 MileagePlus financing required moving the program into a newly formed subsidiary, while United kept full ownership and operational control of it (Kirkland & Ellis, July 7, 2020). A Data HoldCo applies the same idea to a company's data on a much smaller scale. Tax, transfer pricing and legal structure vary by jurisdiction, so take professional advice.

Indirect monetization: the Caterpillar example

Not all data value shows up as data sales. Caterpillar reports more than 1.6 million connected and reporting assets in 2025, which it says help customers manage fleets and make more informed equipment decisions, and services revenues of $24 billion in 2025 (Caterpillar 2025 Annual Report). In March 2025, Caterpillar Group President Bob De Lange said:

"We had grown services revenue from $14 billion in 2016 to last year $24 billion." (Caterpillar, J.P. Morgan virtual HQ visit transcript, March 21, 2025)

Caterpillar attributes its services growth to several efforts, including enhancements to its eCommerce platforms, so this is not evidence that data alone drove the result. It does show how equipment data can support a services business alongside product sales.

How big is the data monetization market?

Grand View Research estimates the global data monetization market at $3.9 billion in 2025 and projects $17.6 billion by 2033, a compound annual growth rate of 20.4 percent from 2026 to 2033. It segments the market by method, including data as a service and insight as a service (Grand View Research, June 2026). Market-size estimates vary by research firm and definition, so treat any single figure as a directional indicator rather than a measure of what your data could earn.

Common mistakes to avoid

  • Selling instead of licensing. Once ownership transfers, you lose control and future revenue.
  • Skipping the rights review. A product built on data you cannot lawfully license is a liability.
  • Treating pseudonymised data as anonymous. Under GDPR it is still personal data.
  • Granting broad exclusivity too early. Exclusive deals can block better offers later.
  • Pricing without a valuation. Without a defensible value, the first buyer usually sets the price.

Where to start

Start small: pick one well-documented dataset, confirm your rights, and get an indicative value. The free 15-question Data Snapshot lets you see an indicative value range in about five minutes. Any range is indicative, not a guarantee of value or revenue.

Frequently asked questions

Can a small or mid-sized company monetize its data?

Yes, if the data is unique, well documented and lawfully licensable. Size matters less than uniqueness, history, quality and buyer demand. Many mid-sized companies hold specialist data, such as industry transactions, equipment readings or service outcomes, that buyers cannot get elsewhere.

Is it legal to sell or license customer data?

It depends on your privacy notices, contracts, consents and the laws that apply, such as GDPR, CCPA and HIPAA. The usual approach is to license aggregated, de-identified data products under contracts that prohibit re-identification. Always get legal advice before licensing data that originated with individuals.

What is a Data HoldCo?

A Data HoldCo is a separate data company that licenses data from an operating company, turns it into de-identified data products and licenses those products to outside buyers. The operating company keeps ownership of its data and keeps the majority of the revenue.

How much money can data licensing generate?

There is no typical figure, and no revenue is guaranteed. Public examples range from undisclosed deals to Reddit's $203.0 million of data licensing contracts disclosed in 2024, but those involve very large platforms. An indicative valuation of your own data is the best way to judge whether a program is worth pursuing.

Do I need a data marketplace to sell data?

No. Marketplaces such as Snowflake Marketplace, AWS Data Exchange and Databricks Marketplace make it easier to reach buyers and handle delivery and billing, but many larger deals, especially AI training licenses, are negotiated directly.

How do I stop buyers from re-identifying people in my data?

Use layered protections: de-identify and aggregate the data to the standard that applies, such as HIPAA expert determination or safe harbor, and sign licenses that prohibit re-identification, restrict combining the data with other sources, limit onward sharing and give you audit rights.

Sources

  1. Gartner, "Data Monetization," Gartner Information Technology Glossary, accessed October 2026. https://www.gartner.com/en/information-technology/glossary/data-monetization
  2. McKinsey & Company, "Fueling growth through data monetization," December 1, 2017. https://www.mckinsey.com/capabilities/quantumblack/our-insights/fueling-growth-through-data-monetization
  3. Reddit, Inc., Form S-1 Registration Statement, February 22, 2024. https://www.sec.gov/Archives/edgar/data/1713445/000162828024006294/reddits-1q423.htm
  4. TechCrunch, "Reddit says it's made $203M so far licensing its data," February 22, 2024. https://techcrunch.com/2024/02/22/reddit-says-its-made-203m-so-far-licensing-its-data
  5. Reuters (republished by Yahoo), "Exclusive: Reddit in AI content licensing deal with Google, sources say," February 22, 2024. https://tech.yahoo.com/ai/articles/exclusive-reddit-ai-content-licensing-041012503.html
  6. Shutterstock (PR Newswire, via AAP), "Shutterstock Expands Partnership with OpenAI, Signs New Six-Year Agreement to Provide High-Quality Training Data," July 11, 2023. https://www.aap.com.au/aapreleases/cision20230711ae52967
  7. PBS NewsHour (Associated Press), "ChatGPT-maker OpenAI signs deal with AP to license news stories," July 13, 2023. https://www.pbs.org/newshour/economy/chatgpt-creator-openai-signs-deal-with-ap-to-license-news-stories
  8. Variety (republished by Yahoo), "News Corp Inks OpenAI Licensing Deal Potentially Worth More Than $250 Million," May 22, 2024. https://malaysia.news.yahoo.com/news-corp-inks-openai-licensing-211417483.html
  9. Ropes & Gray, "Data Asset Financing: New Asset, New Opportunity," June 2023. https://www.ropesgray.com/en/insights/alerts/2023/06/data-asset-financing-new-asset-new-opportunity
  10. United Airlines Holdings, Inc., Form 8-K (event date June 25, 2020), June 26, 2020. https://www.sec.gov/Archives/edgar/data/100517/000110465920077050/tm2023500d1_8k.htm
  11. American Airlines Group Inc., Exhibit 99.1 (pricing of AAdvantage notes and term loan), March 10, 2021. https://www.sec.gov/Archives/edgar/data/4515/000000620121000024/aex991securednotespricingp.htm
  12. Kirkland & Ellis, "Kirkland Represents United Airlines on $6.8 Billion Financing Secured by MileagePlus Program," July 7, 2020. https://kirkland.com/news/press-release/2020/07/kirkland-reps-ua-on--financing-mileageplus
  13. Snowflake, "About Snowflake Marketplace," documentation, accessed October 2026. https://docs.snowflake.com/en/collaboration/collaboration-marketplace-about
  14. Amazon Web Services, "AWS Data Exchange FAQs," accessed October 2026. https://aws.amazon.com/data-exchange/faqs/
  15. Databricks, "Introducing Databricks Marketplace, an Open Marketplace for All Data and AI Assets," June 28, 2022. https://www.databricks.com/blog/2022/06/28/introducing-databricks-marketplace-an-open-marketplace-for-all-data-and-ai-assets.html
  16. European Union, Regulation (EU) 2016/679 (General Data Protection Regulation), Recital 26. https://eur-lex.europa.eu/eli/reg/2016/679/oj
  17. California Civil Code section 1798.140(m) (definition of "deidentified"), California Legislative Information. https://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?lawCode=CIV&sectionNum=1798.140
  18. 45 CFR 164.514, "Other requirements relating to uses and disclosures of protected health information," Cornell Legal Information Institute. https://www.law.cornell.edu/cfr/text/45/164.514
  19. Caterpillar Inc., 2025 Annual Report (SEC Form ARS), 2026. https://www.sec.gov/Archives/edgar/data/0000018230/000130817926000360/cat015318-ars.pdf
  20. Caterpillar Inc., J.P. Morgan virtual headquarters visit transcript, March 21, 2025. https://s25.q4cdn.com/358376879/files/doc_events/2024/03/JPM-Virtual-HQ-Visit-Transcript_3-21-2025_vF.pdf
  21. Grand View Research, "Data Monetization Market Size And Share Report, 2026-2033," June 2026. https://www.grandviewresearch.com/industry-analysis/data-monetization-market