Rosalie Avalon home

ROSALIE AVALON

Search the journal
, ,

The Price of Being Known

A human profile assembled from receipts, a payment card and location data, transforming into stacked dossiers beside the title The Price of Being Known.

How data brokers turn everyday information into profiles that can expose far more than we intended.

I picked up an unfamiliar product at Albert Heijn because a promotion made it free. I scanned my Bonus card. Within a week, advertisements for that product appeared on Facebook.

I had never used it. I had not Googled it.

My suspicion immediately fell on the card. I could not prove the connection, but the experience raised a question: what happens when ordinary details of our lives become information that other companies can buy?

From a purchase to a profile

Albert Heijn describes using purchase information for advertising outside its own channels with the relevant consent. Separately, it acknowledges sharing consented website interaction data with Meta. AH says purchases from cards not activated for Mijn Albert Heijn Persoonlijk are excluded from personalised advertising.[1][2]

These disclosures establish possible mechanisms, not what happened to my particular purchase. A coordinated campaign could also explain both the free product and subsequent advertising.

The larger issue is how observations become commercially useful profiles.

A purchase, postcode, subscription or device identifier may reveal little alone. Connect them, add information from other sources, and the resulting profile can say much more.

How companies assemble the bundle

Data brokers collect or obtain information, match records and organise people into categories. Some information is supplied directly; some is observed; some is inferred. The US Federal Trade Commission’s 2014 industry study documented brokers combining online and offline information and making potentially sensitive inferences about income, religion and health.[3]

The business remains visible in current commercial offerings. Experian’s US marketing service invites companies to submit customer identifiers, including names, addresses, emails and mobile advertising IDs. It then matches records and appends selected attributes, bundles or audience categories. Experian advertises demographic, lifestyle, behavioural and financial attributes.[4]

In everyday language: a business can purchase additional information or predictions to attach to the customers it already knows.

Imagine shopping habits being combined with an estimated income and household characteristics. A marketer could use that combination to predict which offer might appeal. This is an illustrative example, not a verified profile about me.

A profile can influence treatment even when its assumptions are wrong.

What buyers actually receive

There are several markets. A broker may supply records or attributes. An advertising service may sell access to a selected audience without disclosing its members’ names. Other products measure whether advertising was followed by purchases.[5]

The distinction matters. A company can retain identifying information internally while still earning money from what it knows about people.

Removing a name does not necessarily prevent recognition. Coded identifiers can allow organisations to match records about the same person or device. The FTC warns that hashing, which transforms an identifier into a code, does not automatically make information anonymous.[6]

Nor must everything sit in one enormous file. Different organisations can hold different pieces, with matching services connecting them when commercially useful.

The danger grows when information collected in one context becomes material for decisions in another.

When the profile becomes dangerous

Private movements can expose intimate details. In January 2025, the FTC finalised an order restricting sensitive-location practices at Gravy Analytics and Venntel. It alleged that Gravy sold lists associated with visits to healthcare facilities and places of worship.[7]

Such information can expose private associations and create risks of discrimination, intimidation or physical danger. Those are risks, not evidence that every affected person experienced them.

Information can affect financial treatment. A January 2026 FTC settlement involving General Motors and OnStar addressed allegations about driving information supplied to Verisk and LexisNexis. The complaint linked the practices to insurance denials, cancellations and increased premiums. Sharing with those companies ended in March 2024; Smart Driver ended that June.[8]

The concern is the journey: information generated through a connected vehicle became material for assessing a person elsewhere.

Lists can help identify people to exploit. In a 2021 deferred-prosecution agreement, Epsilon admitted that employees in a business unit knowingly sold consumer lists to fraudulent mailing operators during 2008–2017. The schemes included false prize and sweepstakes solicitations.[9]

These cases involve different practices and legal outcomes. They show why the buyer, purpose and onward use matter as much as the original collection.

The danger of being misread

A profile is a representation of a person, not the person themselves. A household purchase does not establish who used it. Visiting a location does not establish why someone went there. A predicted characteristic is not a confirmed fact.

Yet a category can still shape which messages or offers someone receives.

Errors become harder to challenge when people do not know which organisation supplied the information or which inference affected the decision. The FTC’s broker study identified transparency and correction problems across the industry it examined.[3]

The imbalance is clear: organisations can evaluate people using information those people may never have seen.

Profiling can also influence the choices placed in front of us. A 2025 FTC inquiry described capabilities for using personal information to affect prices, personalised discounts and product presentation. Its public examples were hypothetical, and the study was not a blanket finding of illegality. The concern is the unequal knowledge behind an offer: the seller may know which financial incentive is likely to work, while the customer cannot see how the selection was made.[12]

What did we agree to?

Not every subscription sells data, and a discount can be a real saving. The question is whether customers understand the information exchange surrounding the benefit.

Giving details to enter a competition or receive a service does not automatically authorise unrelated uses. EU rules recognise several lawful bases for processing, each with conditions; consent, where relied upon, must meet specific requirements.[10]

Consumers can request information about their personal data, its sources and recipients. They can also object to processing for direct marketing.[11]

Useful questions include: What have you inferred about me? Where did that information come from? Who received it? How can I correct it?

Those rights matter, but the responsibility should not rest entirely on individuals investigating an invisible supply chain.

The free product was visible. The possible commercial afterlife of my shopping record was harder to see.

People deserve to know when information about ordinary living becomes something others can buy, combine and use to make decisions about them.


Sources

  1. Albert Heijn: detailed privacy policy, purchase-based advertising and activation conditions.
  2. Albert Heijn response published by Radar, February 2026.
  3. FTC: 2014 data-broker study, combining information, sensitive inferences and transparency.
  4. Experian US: Enrichment, matching identifiers and appending attributes, bundles and audiences.
  5. ICO: marketing services of data brokers.
  6. FTC: hashing does not automatically make data anonymous, July 2024.
  7. FTC: final Gravy Analytics and Venntel order, January 2025.
  8. FTC: GM/OnStar final complaint, recipients, alleged insurance effects and sharing dates.
  9. US Department of Justice: Epsilon deferred-prosecution agreement, January 2021.
  10. European Commission: legal grounds for processing data and valid consent.
  11. European Commission: access, objection and other individual data-protection rights.
  12. FTC: surveillance-pricing inquiry interim findings, January 2025.

Additional supporting sources

Previous essay
Next essay