Privacy by Design: design principles
1) Why it is needed (goal and area)
PbD ensures that privacy is built into the product by default, not "pasted" on top. For iGaming, it reduces regulatory risks (GDPR/ePrivacy/local laws), protects vulnerable users, increases trust and reduces the cost of incidents. Coverage: web/mobile, KYC/AML/RG, payments, marketing/CRM, analytics/DWH, logs/AWP, partners/vendors.
2) Seven principles (and how to land them in operations)
1. Proactivity, non-reactivity
Threat modeling (LINDDUN/STRIDE) at the discovery stage.
Privacy-acceptance criteria in Jira/PR templates.
2. Privacy by Default
All marketing/personalization toggle switches are off until there is no agreement.
Collect only "strictly necessary" default identifiers.
3. Privacy is built into the design
PII is stored in the regional circuit (data residency), control plane - without PII.
Tokenization/aliasing of keys in service events.
4. Full functionality (win-win)
Modes of "anonymous analytics" and "personalization with consent."
Equal UX without discrimination of those who refused tracking.
5. Security through the lifecycle
Encryption at rest/in transit; BYOK/HYOK; network segmentation; secret management.
WORM logs for evidence and audit.
6. Transparency
Short policies and a "summary box" of key conditions; privacy panel in the profile.
Reporting: who/what/when/why accessed the data.
7. User orientation
Simple texts, lack of dark patterns, availability of WCAG AA +.
Easy withdrawal of consent and convenient DSAR channels.
3) Roles and RACI
DPO/Head of Compliance - PbD policy, DPIA/TRA, risk control. (A)
Security/Infra Lead - cryptography, access, logs, vendors. (R)
Product/UX - privacy requirements in features, lack of dark patterns. (R)
Engineering/Architecture - tokenization, tenant/region isolation, API contracts. (R)
Data/Analytics - de-PII pipelines, PETs, aggregation. (R)
Legal - legal grounds, texts and locales. (C)
Marketing/CRM - consent/suppression, honest communications. (R)
Internal Audit - artifact samples, CAPA. (C)
4) Classification and taxonomy of data
PII basic: full name, e-mail, phone, address, date of birth, IP/ID of the device.
Sensitive PII: biometrics (selfies/liveliness), KYC documents, payment details, RG/SE statuses.
Operating rooms: game events, logs/trails (PII-free by default).
Marketing/Analytics: Cookies/SDKs (by consent).
Rules: minimization, separate storage, explicit purpose and shelf life.
5) Data Lifecycle
1. Collection - only required fields; CIW/consent; age checks.
2. Transmission - TLS 1. 2 +/mTLS, webhook signature, regional routing.
3. Storage - encryption, tokenization, key rotation, market isolation.
4. Usage - RBAC/ABAC, need-to-know, PETs for analytics.
5. Exchange - DPA/SCC, minimum sets, audited channels.
6. Retention/removal - term by category; Cascade delete jobs crypto delete archives.
7. Reporting/auditing - access and export logs, DPIA/DSAR artifacts.
6) DPIA/TRA (how to do it briefly)
Triggers: new PII categories, special categories, new vendors, cross-border transmissions, high RG/biometrics risks.
DPIA template: purpose → category of data → legal basis → flows/map → risks → measures (tech/org) → residual risk → decision.
Artifacts: flow diagram, field list, risk table, approval protocol.
7) Architectural patterns of PbD
Tenant/Region Isolation: physical/logical segregation of databases, keys and secrets.
Control vs Data Plane: global control - no PII; PII only locally.
De-PII Pipeline: before export to DWH - hash/salt, truncation, k-anonymity/cohortation.
Tokenization Gateway: tokens instead of primary identifiers on the service bus.
Edge without PII: CDN/edge cache - public content only.
Fail-Closed: Unknown 'player _ region' → PII operations are not allowed.
8) Technical measures and standards
Encryption: AES-256/GCM at rest; TLS 1. 2+/1. 3; PFS.
Keys: KMS, BYOK/HYOK, rotation, access by HSM roles, log of key operations.
Access: RBAC/ABAC, JIT accesses, separate admin and audit roles.
Logs: immutable (WORM), hash chains, storage in the region.
DevSecOps: secrets in Vault, SAST/DAST, PII field linter, privacy tests in CI.
Test data: default synthetics; if the re-data is de-identification and short retention.
9) PETs (Privacy-Enhancing Technologies)
Aliasing: replacing ID with tokens; key-map is stored separately.
Anonymization: aggregates, k- anonimnost/ℓ -diversity, bining/cohorts.
Differential privacy: noise on reports, "privacy budget."
Federated analytics: local models, exporting weights/aggregates only.
Masking/editing: delete EXIF, grind fields in KYC documents.
10) UX without dark patterns
Equal visibility "Reject All "/" Accept All "/" Customize. "
Clear target texts and examples of data use.
Not personalizing does not impair the basic experience.
Privacy panel in 1-2 clicks from everywhere; AA + availability.
11) Vendors and data transfer
Vendor registry: DC jurisdictions, sub-processors, certification, storage regions, DPA/SCC/IDTA.
"Minimum set" policy: only required fields, no free export.
Notification and revision when changing locations/sub-processors.
12) Data and events (minimum model)
data_asset{id, category{KYC PCI RG CRM LOG ANON}, region, owner, retention_days, lawful_basis, pii{yes/no}}
processing_event{id, actor, purpose, lawful_basis, started_at, ended_at, records_count, export{yes/no, basis_id}}
access_log{id, subject_id_hash, actor, action{read/write/export/delete}, ts_utc, reason, ticket_id}
erasure_job{id, subject_id_hash, scope, started_at, completed_at, evidence_id}
13) KPI/KRI and dashboard PbD
PII Minimization Index (average number of PII fields per feature).
Residency Coverage (% of records in correct region).
Export Justification Rate.
DSAR SLA (median performance/accuracy).
Tag Firing Violations.
Auditability Score (% of cases with full package of artifacts).
Incidents/Findings.
14) Checklists
A. Before Design
- The purposes and legal grounds of processing are defined.
- Data map and field list marked PII/sensitive.
- DPIA/TRA executed; residual risks are accepted.
- An "anonymous mode" or a mode with a minimum of data is thought out.
B. Build/Release
- Secrets in the manager, keys/encryption configured.
- Logs without PII; events and auditing are enabled.
- Regional routing and retention policy are active.
- Tests: consent-gates, deny-by-default for tags, erasure-path.
C. In operations
- Quarterly access and export reviews.
- Monitoring of firing violations and cross-border requests.
- DSARs/deletions are performed on time; artifacts are preserved.
15) Templates (quick inserts)
A) DPIA template (short)
Data categories: ____ (PII: yes/no)
Reason: ____
Streams/Locations: ____
Risk/Impact: ____
Measures: those (cipher/tokens/isolation), org (RBAC/training)
Residual Risk: ____ Decision: Approve/recycle
B) Field Minimization Policy
C) Clause with vendor (PbD obligation)
D) DSAR response (shutter speed)
16) Frequent mistakes and how to avoid them
Collection "just in case." → Minimization policy + code review of schemes.
Raw logs with PII in APM. → Masking/editing on the agent, local storages.
Global DWH with PII. → De-PII aggregates/aliases only.
No DPIA/consent artifacts. → WORM repository, auto-snapshots of UI/texts.
Unaccounted vendors/SDK. → Quarterly register, prohibition of "gray" connections.
17) 30-day implementation plan
Week 1
1. Approve PbD policy and DPIA/TRA templates.
2. Build a map of data/streams by key zones (KYC/PCI/RG/CRM/Logs).
3. Highlight regional perimeters (EU/UK/...); Define the key model (BYOK/HYOK).
Week 2
4) Enable tokenization/de-PII pipelines and deny-by-default for tags.
5) Configure WORM logs (accesses/exports/consent/deletes).
6) Update vendor contracts (DPA/SCC, locations, sub-processors).
Week 3
7) Implement privacy tests in CI (PII linter, CMP screen fixation, erasure-E2E).
8) Release of the privacy panel in the profile; improve texts and locales.
9) Train teams (Product/Eng/Data/CS/Legal).
Week 4
10) Conduct a DPIA review of the top feature, close the CAPA.
11) Start the KPI/KRI dashboard (Residency, Exports, DSAR SLA).
12) Plan v1. 1: diff. privacy for reports, federated pipelines.
18) Interrelated sections
GDPR: User Consent Management/Cookies and CMP Policy
Localization of data by jurisdictions
Age verification and age filters
AML/KYC and artifact storage
Compliance Dashboard and Monitoring/Regulatory Reports
Internal/external audit and audit checklists
BCP/DRP/At Rest & In Transit Encryption