When you buy a company you buy its personal data, and with it every promise, consent gap and unreported breach attached to that data. Cyber due diligence asks whether the target is secure today. Privacy due diligence asks a different question: was the data collected lawfully, can the buyer use it the way the deal model assumes, and what regulatory exposure transfers on completion? This guide covers the privacy and data protection side of a transaction: data inventories, lawful basis and purpose, consent and notice, regulator history, cross-border transfers and the integration period. It is written for privacy counsel, DPOs, deal teams and security leaders who sit on the diligence workstream. For the security assessment itself, see the Cyber Due Diligence in Mergers and Acquisitions Guide.
Key takeaways
- Data value depends on lawful use. A customer database is only an asset if the buyer can lawfully use it for the purposes the valuation assumes; data collected under narrow notices or weak consent can be worth far less than its record count.
- Promises travel with the data. Regulators in the UK, EU and US expect the buyer to honour the privacy commitments the target made at collection, and to give notice or obtain consent before materially changing how the data is used.
- Liability for past failures transfers. The ICO fined Marriott £18.4m over a compromise of Starwood systems that began in 2014, two years before Marriott acquired Starwood, and that its due diligence did not uncover.
- Diligence itself is processing. Disclosing personal data in a data room needs a lawful basis and minimisation; most diligence questions can be answered with aggregates, samples and redacted extracts.
- Integration is where the risk lands. Merging databases, applying one retention rule and moving data across borders after completion create new processing that needs its own assessment, records and notices.
Why privacy diligence is different from cyber diligence
Security diligence looks for weaknesses an attacker could exploit. Privacy diligence looks for weaknesses a regulator, a court or a data subject could exploit, many of which exist even when the systems are well secured. A well-secured target can still hold marketing records with no valid consent or sell data its notice never mentioned.
The two workstreams overlap at breach history and security of processing (GDPR Article 32), and they should share a single request list and data room index. They diverge on what counts as a finding. The M&A Cyber Due Diligence Checklist covers privacy in four lines; this guide expands that section into a full workstream.
| Question | Cyber diligence answer | Privacy diligence answer |
|---|---|---|
| What data does the target hold? | Where it is stored and how it is protected | Why it was collected, under which notice and lawful basis, and for how long it may be kept |
| Has anything gone wrong? | Incidents, compromises, open vulnerabilities | Breaches notified or not notified, complaints, regulator inquiries, fines, undertakings, litigation |
| Can the buyer use the data? | Can it be migrated securely? | Is the planned use compatible with the original purpose, or does it need fresh notice or consent? |
| Who else touches the data? | Third-party access and integration risk | Processors, joint controllers, data sales and sharing, and whether contracts meet Article 28 |
| Where does it go? | Network paths and hosting locations | International transfers and the mechanism relied on for each |
What the regulators expect from buyers
Three regulatory positions set the baseline for most deals.
- UK ICO: The ICO's data sharing code has a section on due diligence after mergers and acquisitions. It says the buyer must consider data sharing as part of the due diligence carried out when taking on an organisation, and should establish what data is being transferred, the purposes for which it was originally obtained, whether those purposes have changed, and the lawful basis for sharing it. It also expects documentation, technical advice where different systems are involved, accurate records, a consistent retention policy and appropriate security. The section applies to insolvency asset sales as well as takeovers. The ICO notes the guidance is under review following the Data (Use and Access) Act 2025, so check the current version before relying on it.
- US FTC: In its 2014 letter on Facebook's acquisition of WhatsApp, FTC staff stated that an acquisition does not nullify privacy promises, and that using data in a way materially inconsistent with the promises made at collection, without the consumer's affirmative express consent, may be an unfair or deceptive practice under Section 5 of the FTC Act.
- California: Under the CCPA, transferring personal information as an asset in a merger, acquisition or bankruptcy where the third party assumes control of the business is not a "sale", provided the information is used or shared consistently with the statute. If the acquirer materially alters how it uses or shares the data in a way materially inconsistent with the promises made at collection, it must give consumers prior notice (Cal. Civ. Code 1798.140(ad)(2)(C)). See the CCPA and CPRA guide.
Under the GDPR and UK GDPR the same logic runs through the purpose limitation principle (Article 5(1)(b)) and the compatibility test for further processing (Article 6(4)), which weighs the link between purposes, the context of collection, the nature of the data, the consequences for individuals and the safeguards in place. In US healthcare, the HIPAA Privacy Rule definition of health care operations at 45 CFR 164.501 includes the sale, transfer, merger or consolidation of a covered entity with another covered entity, and the due diligence related to it, which is why covered entities can share protected health information for diligence under that heading. See the HIPAA Privacy Rule Guide.
The privacy request list
Send this list with the first information request, alongside the cyber questionnaire. Each line names the evidence that answers it, so the target cannot reply with a policy document alone.
| Area | Request | Evidence that answers it | Red flag |
|---|---|---|---|
| Data inventory | Categories of personal data, data subjects, volumes, systems and locations | Records of processing (GDPR Art. 30), data map, system list with record counts | No ROPA, or a ROPA that does not match the system list |
| Lawful basis | Lawful basis per processing purpose, with legitimate interests assessments where relied on | ROPA columns, LIAs, consent records | Consent claimed but no record of when, how and to what each person agreed |
| Notices | Every privacy notice version in force over the retention period | Versioned notices with dates, collection point screenshots | Current notice only; no history of what customers were told |
| Marketing | Basis for email, SMS and telephone marketing; suppression lists | Opt-in records, soft opt-in rules, preference centre logs | Purchased lists used for marketing |
| Special category and children's data | Health, biometric, children's and other sensitive data, and the condition relied on | Article 9 conditions, age assurance approach | Biometrics or health data with no documented condition |
| DPIAs | DPIAs for high-risk processing such as profiling, monitoring and large-scale sensitive data | DPIA register and signed assessments | Large-scale profiling with no DPIA |
| Data sharing and sales | Every disclosure to third parties, data sales, data brokers and advertising partners | Data sharing agreements, ad tech partner list | Revenue from data sales not reflected in notices |
| Processors | Processor list and contracts | Article 28 terms, sub-processor lists, DPAs | Key processors with no written terms |
| International transfers | Transfers out of the UK and EEA and the mechanism for each | Transfer register, SCCs, transfer impact assessments, adequacy reliance | Transfers to group companies or vendors with no mechanism |
| Rights requests | Volumes and handling times for access, erasure and objection requests | Request log and metrics | Backlogs or repeated complaints about access requests |
| Retention | Retention schedule and evidence it is applied | Schedule, deletion job logs | Indefinite retention of former customer data |
| Breaches | All personal data breaches, including those judged not notifiable, and the reasoning | Breach register (GDPR Art. 33(5)), notifications, regulator correspondence | An empty breach register in a business of any size |
| Regulators and litigation | Complaints, inquiries, audits, fines, undertakings and class actions in every jurisdiction | Correspondence, settlement terms, consent orders | Open inquiry or an order that binds future data use |
For how to build or test a ROPA, see the Data Mapping and ROPA Guide. For assessing high-risk processing found during diligence, see the DPIA Guide: Running a Data Protection Impact Assessment.
Testing lawful basis, consent and purpose
This is the part of privacy diligence that most often changes the price. The question is not whether the target has a privacy policy but whether the buyer's post-deal plan for the data is lawful.
- Write down the deal thesis for the data. Ask the deal team how the buyer intends to use the target's data: cross-selling to the buyer's customers, combining datasets for analytics, training models, or folding the target's users into the buyer's platform. Each use is a purpose to test.
- Trace each dataset to its collection notice. Take a sample of records from each major system and match each to the notice version shown when it was collected. A notice that promised "we will never share your data with third parties" limits what the buyer may do, even after completion.
- Run the compatibility test. For each planned use, apply the Article 6(4) factors. Serving the same customers under a new brand is usually compatible; marketing insurance from health app data is not.
- Check consent quality where consent is the basis. Look for granular, recorded, withdrawable consent with the wording the person saw. Bundled or pre-ticked consent, or consent records that cannot be produced, mean the dataset may not be usable for that purpose.
- Classify the outcome. Usable as planned; usable after notice; usable only after fresh consent; or not usable and to be deleted. Give the deal team record counts for each class, because that is what affects valuation.
Legitimate interests: Where the target relies on legitimate interests, ask for the legitimate interests assessments. A missing assessment is fixable; a balancing test that would plainly fail, such as covert tracking, is a liability.
Notice obligations after completion: When the buyer obtains data from the target, Article 14 generally requires the buyer to inform individuals within a reasonable period, at the latest within one month. Individuals may also have a right to object (Article 21), and an absolute right to object to direct marketing. Plan the customer communication before completion rather than after.
Regulatory and breach history
Ask for the full history, not just open matters. The Marriott case shows why. Attackers first exploited weaknesses in the Starwood guest reservation system in 2014; Marriott acquired Starwood in 2016; the attack was identified in September 2018. The ICO issued a notice of intent proposing £99.2m in July 2019 and a final penalty of £18.4m in 2020 for breaches of GDPR Articles 5(1)(f) and 32. The acquirer carried the liability for systems its diligence had not examined deeply enough.
Disclosure history matters too. Yahoo suffered a breach in December 2014 that became public in 2016, while it was closing the sale of its operating business to Verizon. The SEC later fined Altaba, the renamed Yahoo, $35m in 2018 for failing to disclose the breach to investors. For SEC-registered buyers, see the SEC Cybersecurity Disclosure Rules Guide.
- Ask for the breach register, not the notification list: GDPR Article 33(5) requires controllers to document all personal data breaches, including those not notified. The reasoning behind non-notification decisions is often the most revealing document in the data room.
- Read regulator orders for forward-looking terms: Undertakings, consent orders and settlements can restrict future data use, require audits for years, or bind successors. These constrain the integration plan.
- Look for complaint patterns: Repeated complaints about access requests or marketing often precede an inquiry.
- Check notification timeliness: Breaches notified well after the 72-hour GDPR window, without documented reasons for delay, are a signal of weak incident governance. See the Data Breach Notification Checklist.
Cross-border transfers and data localisation
Diligence has two tasks here: confirm that the target's existing transfers have a valid mechanism, and identify the new transfers the integration will create.
| Situation | What to check | Typical remediation |
|---|---|---|
| Target transfers EU or UK data to US vendors | Vendor certification under the EU-US Data Privacy Framework or UK extension, or SCCs with a transfer impact assessment | Update contracts; complete missing TIAs |
| Post-deal sharing with a non-EEA parent | Mechanism for intra-group transfers; whether the buyer has binding corporate rules that could extend to the target | Intra-group agreement with SCCs, or BCR extension |
| Target operates in China, Russia or other localisation regimes | Local storage requirements and export approval or filing obligations | Keep local instances; budget for regulatory filings |
See the Standard Contractual Clauses (SCCs) Guide, the Transfer Impact Assessment (TIA) Template and the China Cross-Border Data Transfers guide.
Running the data room without creating a new breach
Disclosing personal data to a potential buyer is itself processing by the seller, and receiving it is processing by the buyer. Both need a lawful basis, which in practice is usually legitimate interests, supported by minimisation and confidentiality terms.
- Answer with aggregates first: Record counts, category lists and retention statistics answer most questions without disclosing a single customer.
- Use redacted samples for testing: Where the buyer must test consent records or notices, provide pseudonymised extracts in a clean room or read-only viewer.
- Restrict employee data: Summarise or anonymise HR, health and disciplinary records until late-stage clean team review.
- Control access: Named accounts, MFA, watermarking, no bulk download for personal data folders, and an access log the seller reviews. See the Identity Integration in M&A Guide for identity controls across the deal.
- Delete on exit: The non-disclosure agreement should require losing bidders to return or destroy personal data and confirm it in writing.
Turning findings into deal terms
Privacy findings should reach the deal team in a form they can price. Grade each finding and propose a response.
| Finding | Severity | Typical deal response |
|---|---|---|
| Open regulator inquiry or unnotified breach of sensitive data | Critical | Specific indemnity, escrow or holdback, condition precedent to resolve or disclose; possibly walk away |
| Core dataset not usable for the planned purpose | High | Price adjustment reflecting unusable records; revised synergy case |
| Missing DPIAs, LIAs or Article 28 contracts | Medium | Pre-completion covenant to remediate, or a funded 100-day plan |
| Transfers without a mechanism | Medium to high | Remediate before completion if feasible; warranty and indemnity |
| Weak retention, stale data | Low to medium | Post-completion deletion programme; reduces breach exposure |
Ask for data protection warranties that are specific: compliance with applicable data protection law for a defined look-back period, no unnotified personal data breaches, no regulatory investigations, accuracy of the ROPA, and valid mechanisms for transfers. Warranty and indemnity insurers often exclude known issues and may exclude fines, so a known privacy problem usually needs a specific indemnity rather than reliance on general warranties.
Integration: the first 100 days
Most privacy harm in a deal happens after completion, when systems and teams are merged. Treat integration as a programme of new processing.
- Day one: Freeze any new use of the target's data beyond business as usual until purposes are confirmed. Confirm who the controller is for each dataset after completion, and whether any joint controller arrangements (Article 26) now exist.
- First 30 days: Send customer and employee notices where Article 14 or the target's promises require it. Update the buyer's ROPA to include the target's processing. Register or update DPO and regulator registrations.
- Before any database merge: Run a DPIA for combining datasets, profiling across brands or moving data across borders. Follow the ICO's point that combining systems risks loss, corruption or degradation of data: test migrations on copies and reconcile record counts.
- Days 30 to 90: Apply one retention schedule across both organisations and delete data that fails it. Consolidate rights request handling so customers of either business get one response route.
- Rollback plan: Keep the target's systems readable until reconciliation passes, so a failed migration can be reversed without losing records or rights request history.
Inherited credentials and service accounts also carry personal data access. See Inheriting NHIs in an Acquisition.
Decision guidance: depth by deal type and size
- Consumer-facing targets: Full lawful basis testing with record sampling. Customer data is usually central to the valuation.
- B2B software and data processors: Focus on the target's processor obligations, customer DPAs and sub-processor chains; a breach at a processor reaches many controllers.
- Health, children's and financial data: Treat special category and regulated data as high risk by default; involve specialist counsel per jurisdiction.
- Asset purchases and insolvency sales: Only the data transferred matters, but the buyer must still establish purpose and lawful basis for each dataset; the ICO code covers this case explicitly.
- Small deals: A short questionnaire, the ROPA, notices, breach register and regulator correspondence will surface most material issues in a few days. Large deals justify sampling and interviews with the DPO.
Measuring the workstream
- Request coverage: Percentage of privacy request lines answered with evidence rather than policy statements.
- Usable data ratio: Share of records in key datasets classed as usable for the planned purpose without fresh consent.
- Finding closure: Number of high and critical findings with an agreed deal response before signing.
- Notice completion: Percentage of affected individuals notified within the Article 14 window after completion.
- Integration DPIAs: Number of integration processing changes that went live with a completed DPIA, against the total.
Framework mapping
| Activity | NIST SP 800-53 Rev. 5 | NIST CSF 2.0 | ISO/IEC 27001:2022 Annex A |
|---|---|---|---|
| Data inventory of the target | PM-5(1) Inventory of Personally Identifiable Information | ID.AM-07 | A.5.9, A.5.34 |
| Lawful basis and purpose testing | PT-2 Authority to Process PII; PT-3 PII Processing Purposes | GV.OC-03 | A.5.31, A.5.34 |
| Notices to individuals after completion | PT-5 Privacy Notice | GV.OC-03 | A.5.34 |
| Assessing the target as a risk-bearing third party | SR-6 Supplier Assessments and Reviews | GV.SC-07 | A.5.19, A.5.20 |
Common failure patterns
- Treating privacy as a policy review: Reading the privacy notice and DPO letter, without tracing data to collection notices, misses the findings that affect value.
- Valuing records, not rights: Synergy cases built on cross-selling to the target's customers, without testing whether that use is compatible, collapse after completion.
- Accepting a clean breach list: Asking only for notified breaches hides the non-notification decisions that regulators later disagree with.
- Over-sharing in the data room: Uploading raw customer exports or HR files creates a new exposure for both sides.
- Merging databases first, assessing later: Integration that combines datasets before a DPIA and transfer assessment turns an inherited issue into the buyer's own new breach of the GDPR.
Practitioner checklist
- Agree the deal thesis for the target's data and list every planned use as a purpose to test.
- Send the privacy request list with the first information request and require evidence, not policies.
- Sample records from key systems and trace each to its collection notice and lawful basis.
- Obtain the full breach register, regulator correspondence and any orders binding successors.
- Map every international transfer and the post-deal transfers integration will create.
- Run the data room on aggregates and redacted samples, with access logging and deletion on exit.
- Grade findings and give the deal team record counts and a proposed deal response for each.
- Draft customer and employee notices before completion and send them within the Article 14 window.
- Complete DPIAs before any database merge, cross-brand profiling or cross-border migration.
- Apply one retention schedule across both businesses within 90 days.
Standards and references
- ICO: Data sharing code, due diligence when sharing data following mergers and acquisitions
- EUR-Lex: Regulation (EU) 2016/679 (General Data Protection Regulation)
- FTC: FTC staff to Facebook and WhatsApp: Privacy promises prevail
- California Privacy Protection Agency: California Consumer Privacy Act statute
- eCFR: 45 CFR 164.501 Definitions (HIPAA Privacy Rule)
- SEC: Altaba, formerly known as Yahoo, charged with failing to disclose massive cybersecurity breach
- Mishcon de Reya: ICO's £18.4m fine against Marriott, a reminder when conducting due diligence
Related NHI Mgmt Group resources: Cyber Due Diligence in Mergers and Acquisitions Guide · M&A Cyber Due Diligence Checklist · Identity Integration in M&A Guide · Mergers and Acquisitions Cyber Risk Executive Briefing. This guide summarises data protection considerations in transactions and is not legal advice.