Precise location data reveals where people live, work, worship and seek medical care, and a handful of location points is usually enough to identify a person even when names have been removed. Regulators now treat it accordingly: the US Federal Trade Commission has brought actions against companies selling location data tied to visits to clinics and places of worship, California classes precise geolocation as sensitive personal information, and EU law restricts processing of device location data. This guide covers what counts as location data, where it comes from inside an organisation (mobile apps, SDKs, Wi-Fi and network logs, vehicles, employee devices), the main legal rules, the technical controls that reduce risk (coarsening, retention limits, sensitive location suppression, access control) and how to govern third parties. It is for privacy engineers, security architects and product teams. Status checked 27 October 2026; laws and orders change, so confirm the current text before relying on any point.
Key takeaways
- "Anonymised" location data rarely is. Research on 1.5 million people's mobile phone records found that four spatio-temporal points were enough to uniquely identify 95 per cent of them; treat raw location traces as personal data.
- Sensitive places drive enforcement. FTC actions against X-Mode/Outlogic and InMarket targeted location data that could reveal visits to medical clinics, places of worship and shelters, and required programmes to exclude sensitive locations.
- Consent must be specific and informed. A location permission prompt for an app feature is not consent to selling or profiling; the FTC alleged both companies failed to obtain informed consent for how the data was used.
- Precision is a design choice. Collect the coarsest location that serves the purpose, keep it for the shortest time, and separate precise traces from identifiers.
- SDKs and partners are your exposure. Location data often leaves through third-party SDKs and data partners; inventory them and assess whether suppliers obtained valid consent.
What counts as location data
| Source | Typical precision | Where it appears in an organisation |
|---|---|---|
| Device GPS and fused location services | Metres | Mobile apps, field service apps, fleet and delivery tools |
| Wi-Fi and Bluetooth beacons | Metres to tens of metres | Retail analytics, office occupancy systems, wireless controllers |
| Mobile network cell data | Hundreds of metres to kilometres | Telecom operators, call detail records |
| IP address geolocation | City or region, sometimes less precise | Web logs, fraud and risk engines, sign-in logs |
| Vehicle telematics | Metres, continuous | Fleet management, connected car services, insurers |
| Photo and file metadata (EXIF) | Metres | User uploads, document stores |
| Physical access control and badge logs | Building or door | Facilities and security systems |
Two legal definitions show how regulators draw the line. The EU ePrivacy Directive (2002/58/EC, Article 2(c)) defines location data as data processed in an electronic communications network or service that indicates the geographic position of a user's terminal equipment. The California Consumer Privacy Act as amended (Cal. Civ. Code 1798.140(w)) defines precise geolocation as data derived from a device and used or intended to locate a consumer within an area equal to or less than a circle with a radius of 1,850 feet, except as prescribed by regulations.
Why location data is hard to anonymise
A location trace is a pattern: home at night, work in the day, the same gym on Tuesdays. In "Unique in the Crowd: The privacy bounds of human mobility" (Scientific Reports), de Montjoye and colleagues studied 15 months of mobile phone data for about 1.5 million people at hourly and antenna-level resolution and found that four randomly chosen spatio-temporal points uniquely identified 95 per cent of individuals. Removing names or replacing them with an advertising identifier does not change that: the home location alone often points to one household. Practical consequences:
- Pseudonymised traces are personal data. Plan for them as such under GDPR and US state laws.
- Aggregate, do not just strip identifiers. Counts per area and time window, with small counts suppressed, are far safer than individual traces.
- Consider formal methods for published statistics. See the Differential Privacy Guide and the Anonymisation and Re-identification Risk Guide.
The legal landscape
This is a summary for practitioners, not legal advice. Status as checked on 27 October 2026.
| Regime | What it says about location data | Practical effect |
|---|---|---|
| EU ePrivacy Directive 2002/58/EC, Article 9 | Location data other than traffic data may only be processed when made anonymous or with the consent of users, to the extent and for the duration needed for a value added service; users must be told the type of data, purposes, duration and whether it goes to a third party; consent can be withdrawn at any time, and users must be able to refuse processing temporarily, simply and free of charge | Applies to providers of electronic communications services in the EU; national laws implement it. See the ePrivacy Directive Guide |
| EU GDPR | Location data is named in the definition of personal data as an identifier | Lawful basis, transparency, minimisation, DPIA for systematic tracking |
| California CCPA/CPRA | Precise geolocation (1,850 feet radius) is sensitive personal information (1798.140(ae)(1)(C)); consumers have a right to limit its use and disclosure (1798.121) | "Limit the Use of My Sensitive Personal Information" handling for businesses in scope |
| Washington My Health My Data Act, RCW 19.373.080 | Unlawful to implement a geofence around an entity providing in-person health care services to identify or track consumers seeking care, collect consumer health data, or send health-related notifications, messages or ads | No geofenced advertising or tracking around clinics. See the Health Data Privacy Beyond HIPAA Guide |
| FTC Act Section 5 (US federal) | Unfair or deceptive practices, applied to the sale and use of location data | Enforcement actions with bans and programme requirements (next section) |
Several other US state comprehensive privacy laws also treat precise geolocation as sensitive data with consent or opt-out requirements and their own radius definitions; check each state that applies to you. For an overview of regimes, see the Global Privacy Regulatory Map.
What FTC enforcement requires
Two FTC actions announced in January 2024 set out what the regulator expects from companies handling location data. Both were announced as proposed orders; the FTC later announced final orders (X-Mode in April 2024, InMarket in May 2024), so check the FTC case pages for the final terms.
| Case | FTC allegations | Order requirements |
|---|---|---|
| X-Mode Social / Outlogic (announced 9 January 2024) | Sold precise location data tied to mobile advertising IDs that could reveal visits to medical and reproductive health clinics, places of worship and domestic abuse shelters; had no policy to remove sensitive locations until May 2023; did not ensure users were informed or consented; did not honour some Android opt-outs | Ban on sharing or selling sensitive location data; a programme to maintain a comprehensive list of sensitive locations; a supplier assessment programme to confirm informed consent; deletion of previously collected data unless consent is obtained or data is deidentified or made non-sensitive; easy withdrawal of consent and deletion; privacy programme and retention schedule |
| InMarket Media (announced 18 January 2024) | Collected location data through its own apps and an SDK in third-party apps, matched it to about 2,000 audience segments for advertising, did not obtain informed consent, and kept geolocation data for five years | Ban on selling or licensing precise location data; ban on products that target consumers based on sensitive location data; deletion, consumer notice and opt-out; a sensitive location data programme and SDK supplier assessment programme; privacy programme and retention schedule |
The pattern is the checklist: know where location data comes from, verify consent at source, exclude sensitive places, limit retention and control downstream use.
Technical controls
Collect less precision
- Mobile permissions: request approximate rather than precise location where the feature allows (Android's coarse location permission, iOS approximate location), and foreground-only rather than background access.
- Coarsen at the edge: round coordinates before they leave the device. Three decimal places of latitude are roughly 110 metres, two decimal places roughly 1.1 kilometres; choose the coarsest level that serves the purpose.
- IP geolocation: store region or city derived from the IP for fraud scoring rather than keeping a long-term table of IP to user mappings.
Separate, restrict and expire
- Separate identifiers from traces: keep precise traces in a store keyed by a rotating pseudonym, with the mapping table held separately under stricter access.
- Retention schedule: define retention per purpose (for example, live tracking deleted after delivery, fraud signals kept for a defined period) and enforce it with automated deletion. NIST SP 800-53 SI-12 Information Management and Retention covers this.
- Access control: limit query access to location stores to named roles, log every query and alert on bulk exports.
- Encryption: encrypt in transit (PR.DS-02) and at rest, with keys managed so that analytics roles cannot read raw traces.
Suppress sensitive locations
- Maintain a sensitive locations list (health care, reproductive health, places of worship, shelters, addiction and mental health services, and similar categories), as the FTC orders require of the companies concerned.
- Filter before storage or sharing: drop or coarsen points within a buffer of those places, and exclude them from audience segments.
Aggregate before sharing
- Share counts, not traces: heat maps and footfall counts with minimum cell sizes rather than device-level data.
- Contract and technical limits: where device-level data must be shared, prohibit re-identification and onward sale contractually and restrict fields technically.
Implementation approach
- Discover: inventory every system that collects or stores location (apps and their SDKs, Wi-Fi controllers, telematics, web logs, data warehouse tables with latitude and longitude columns, photo metadata). Record source, precision, purpose, retention and recipients (CSF 2.0 ID.AM-07 data inventories).
- Classify: tag precise location as sensitive data in the data catalogue so that access policies and data loss prevention apply.
- Assess: run a DPIA or privacy risk assessment for systematic tracking, for any sharing, and for any use involving employees.
- Fix collection first: reduce precision and background collection in apps, remove SDKs that collect location you do not need, and update notices and consent flows (PT-5 Privacy Notice).
- Apply retention and suppression: deploy deletion jobs and sensitive location filters, starting with the largest stores.
- Govern sharing: list every recipient, require consent evidence from suppliers, and stop flows that cannot be justified (AC-21 Information Sharing).
- Test and roll back: release precision and SDK changes behind feature flags to a pilot group, measure feature impact, and revert the flag if a critical function breaks while a coarser alternative is designed.
Small versus large organisations. A small business with one app and no data sales can often meet expectations by using approximate location, dropping unnecessary SDKs and setting short retention. A large organisation with data partnerships needs a supplier assessment programme, a maintained sensitive locations list, a privacy review for every new use, and regular audits of what partners receive.
Employees, devices and security tooling
Security teams handle location data too: mobile device management can locate devices, sign-in logs record IP-derived locations, and badge systems record movements. Use device location only for defined purposes such as finding a lost device, with notice to staff; do not repurpose it for performance monitoring without legal review. Sign-in location used for risk scoring (impossible travel, unfamiliar country) is generally proportionate, but retention and access should still be limited. See the Employee Monitoring and Privacy Guide.
Where identity fits
Location data becomes most dangerous when joined to an identifier: an advertising ID, an account ID or a device ID. Many of the flows that move it are machine to machine, through SDK keys, partner API credentials and data warehouse service accounts. Inventory those non-human identities, scope each to the minimum fields and time range, and rotate or revoke them when a partnership ends, so that a stale API key cannot keep exporting location data. See the API Key Management Guide.
Detection and measurement
- Monitor exports: alert on bulk queries or exports from location stores, and on new destinations for location fields in data pipelines.
- Scan apps: test each mobile release for location permissions, SDKs and outbound location fields, using the OWASP MASVS privacy category as a baseline.
- Audit partners: sample what each recipient actually receives against the contract.
Metrics:
- Percentage of location data stores with a documented purpose, retention period and owner.
- Percentage of location data deleted on schedule (from deletion job logs).
- Number of SDKs with location access per app (falling).
- Share of app sessions using approximate rather than precise location.
- Recipients of device-level location data with current consent evidence (target 100 per cent).
- Opt-out and limit requests completed within the legal deadline.
Framework mapping
| Framework | Reference | Relevance |
|---|---|---|
| NIST SP 800-53 Rev. 5 | PT-2, PT-3, PT-5, PT-7 | Authority to process, purposes, notice, specific categories of PII |
| NIST SP 800-53 Rev. 5 | SI-12, AC-21 | Retention and controlled information sharing |
| NIST CSF 2.0 | GV.OC-03, ID.AM-07, PR.DS-02 | Legal requirements, data inventories, protection in transit |
| ISO/IEC 27001:2022 Annex A | A.5.34 Privacy and protection of PII, A.8.11 Data masking | PII protection and de-identification |
| OWASP MASVS | MASVS-PRIVACY | App-level privacy baseline |
Common failure patterns
- Treating advertising IDs as anonymous: device-level traces keyed by advertising ID are sold or shared as "anonymous", which the FTC actions rejected.
- SDKs nobody owns: an analytics or monetisation SDK collects precise background location the product team never asked for.
- Indefinite retention: raw traces kept for years "for analytics", as in the five-year retention the FTC alleged against InMarket.
- Consent for one purpose, use for another: location granted for store finders reused for profiling or sale.
- No sensitive location handling: segments built that include visits to clinics or places of worship.
- Latitude and longitude in logs: application and debugging logs quietly capture precise coordinates and are kept and replicated without any of the controls on the main store.
Practitioner checklist
- Inventory every source and store of location data, with precision, purpose, retention and recipients.
- Classify precise location as sensitive data in the data catalogue.
- Default apps to approximate, foreground-only location unless a feature needs more.
- Remove SDKs that collect location without a documented need.
- Coarsen coordinates at collection and strip location from logs and file metadata.
- Enforce retention with automated deletion.
- Maintain a sensitive locations list and suppress those points before storage or sharing.
- Share aggregates with minimum cell sizes instead of device-level traces.
- Obtain consent evidence from suppliers and contract against re-identification and resale.
- Handle California limit requests and other state rights where they apply, and re-check legal status regularly.
Standards and references
- EUR-Lex: Directive 2002/58/EC (ePrivacy Directive), consolidated text
- California Privacy Protection Agency: California Consumer Privacy Act text
- Washington State Legislature: RCW 19.373.080 Geofence restrictions
- FTC: Order prohibits data broker X-Mode Social and Outlogic from selling sensitive location data
- FTC: Order will ban InMarket from selling precise consumer location data
- Scientific Reports: Unique in the Crowd: The privacy bounds of human mobility
- OWASP: Mobile Application Security Verification Standard
- NIST: SP 800-53 Rev. 5
Related NHI Mgmt Group resources: Anonymisation and Re-identification Risk Guide · Health Data Privacy Beyond HIPAA Guide · ePrivacy Directive Guide · Global Privacy Regulatory Map. This guide summarises location data privacy rules and enforcement as of 27 October 2026 and is not legal advice.