A data broker is a business that collects, aggregates and sells or licenses personal data about individuals with whom it has no direct relationship. Data brokers assemble records from public sources, commercial transactions and online tracking, link them into detailed profiles, and supply those profiles to marketers, lenders, insurers, employers, landlords, investigators and government agencies.
In Privacy, data brokers are a central example of how personal data moves far beyond the context in which people disclosed it. The industry operates largely out of sight of the people it describes, which is why regulators treat it as a priority for transparency, consent and deletion rights.
How data brokers collect and combine personal data
A data broker builds value through aggregation and linkage. Individual facts, such as an address, a car registration or a magazine subscription, are rarely sensitive on their own. Combined across hundreds of sources and matched to the same person through identity resolution, they form a profile that can reveal income, health interests, religion, family situation and daily movements.
Brokers typically obtain data from three channels: public records such as property deeds, court filings and voter rolls where access is permitted; commercial sources such as loyalty programs, warranty registrations, app developers and other brokers; and online tracking through cookies, mobile advertising identifiers and software development kits embedded in apps. The table below summarizes these sources.
| Source type | Examples | Data obtained |
|---|---|---|
| Public records | Property, court and business registries | Names, addresses, ownership, legal history |
| Commercial transactions | Loyalty cards, catalogs, subscriptions | Purchases, brand preferences, spending level |
| Online tracking | Cookies, pixels, app SDKs | Browsing interests, device IDs, app usage |
| Location feeds | Apps sharing GPS or Wi-Fi signals | Visited places, home and work locations |
| Other brokers | Data exchanges and resellers | Appended attributes and modeled scores |
Brokers also create inferred data, which is information nobody provided. Statistical models assign people to segments such as “expecting parent” or “likely diabetic”, or estimate household income. These modeled attributes can be wrong, yet they are sold as facts and influence real decisions.
Types of data brokers and what they sell
The term covers several business models that differ in customers, products and legal exposure. Some firms combine models, but the distinctions matter because different laws apply to each, particularly when a product is used for eligibility decisions about credit, housing or employment.
| Category | Main product | Typical customers | Key risk |
|---|---|---|---|
| Marketing brokers | Audience segments and lists | Advertisers, retailers | Profiling and manipulation |
| People-search sites | Individual background reports | General public, investigators | Stalking and harassment |
| Risk mitigation brokers | Identity verification, fraud scores | Banks, insurers, platforms | Opaque errors in decisions |
| Location data brokers | Device movement histories | Analysts, agencies, advertisers | Revealing sensitive visits |
| Consumer reporting agencies | Credit and tenant reports | Lenders, landlords, employers | Denial of services |
How data brokers relate to controllers and processors
Under the GDPR, a data broker that decides why and how personal data is processed is a data controller in its own right, not a mere processor. It must have a lawful basis, inform data subjects and honor their rights, even though it never interacts with them. The obligation under Article 14 GDPR to inform people when data was not obtained from them directly is especially demanding for brokers.
The distinction from neighboring roles is easy to confuse, so the table below compares them.
| Role | Relationship with the person | Decides purposes | Sells data onward |
|---|---|---|---|
| Data broker | None | Yes | Yes, as the core business |
| First-party controller | Direct, such as a customer | Yes | Sometimes, often restricted |
| Data processor | None | No, follows instructions | No |
| Ad tech intermediary | Usually none | Partly | Often, through bidding data |
Regulation of data brokers in the US and the EU
Regulation differs sharply between jurisdictions. In the EU, the GDPR applies fully to brokers, which makes broad, consent-free profiling difficult to justify and has led regulators to fine several firms. In the United States, there is no comprehensive federal privacy law, so oversight comes from sector laws, Federal Trade Commission enforcement and state statutes.
California gives residents rights under the CCPA to know, delete and opt out of the sale of their data, and the Delete Act created a centralized mechanism for requesting deletion from all registered brokers at once. Vermont, Texas and Oregon require broker registration. The Fair Credit Reporting Act applies when broker data is used for credit, insurance, employment or housing eligibility, which imposes accuracy and dispute duties.
| Instrument | Jurisdiction | Main effect on brokers |
|---|---|---|
| GDPR | European Union | Lawful basis, notice and data subject rights |
| CCPA and CPRA | California | Rights to know, delete and opt out of sale |
| Delete Act | California | Registry and single deletion request mechanism |
| Fair Credit Reporting Act | United States | Accuracy and dispute rules for eligibility data |
| State registration laws | Vermont, Texas, Oregon | Public registry and annual disclosures |
Risks, harms and limits of opting out
The harms of the broker industry stem from scale, invisibility and secondary use. People cannot correct profiles they do not know exist, and the same dataset can serve harmless advertising and harmful purposes such as locating a domestic violence survivor, targeting vulnerable consumers with scams, or enabling warrantless access by agencies that buy data instead of obtaining it through legal process.
- Re-identification: supposedly anonymous location trails often point to a single home and workplace.
- Inaccuracy: modeled attributes may mislabel people and affect prices or eligibility.
- Sensitive inferences: visits to clinics or places of worship reveal special category data.
- Opt-out churn: removed records frequently reappear when brokers refresh their sources.
Individual opt-outs offer limited and temporary protection. Hundreds of brokers exist, each with its own request process, and many repopulate profiles from new data. Structural measures such as data minimization by first-party companies and centralized deletion mechanisms address the problem more effectively than one-by-one requests.
In simple terms
A data broker is a company that gathers scraps of information about you from many places, such as public records, store loyalty cards and apps on your phone, and pieces them together into a file. It then sells that file to other companies. You usually never hear of the broker, and it never asked you directly.
Think of it as a collector of puzzle pieces who completes your picture and then sells copies. The main takeaway is that your information can travel much further than you expected, and some laws now let you ask brokers to delete it.
A practical example
A nurse in California searches her own name and finds a people-search site listing her home address, phone number, age and relatives. She works in an emergency department and has had a threatening encounter with a patient, so the listing worries her. She submits removal requests to six people-search sites, but two months later her profile reappears on three of them.
She then uses the state’s centralized deletion mechanism under the Delete Act, which sends one request to every registered data broker and requires periodic re-processing. Within the following reporting cycle, her listings disappear from the registered sites, although one offshore site that is not registered keeps showing her details. The case shows both the value of centralized deletion and its limits.