Historical task source. The Eve task board holds current task state.
Phase 142: Sovereign Product Analytics & Experimentation Platform — Mixpanel / Amplitude / PostHog / Heap / Segment / Google Analytics Replacement#
Closes the product-analytics sovereignty gap. Phase 139 (Office Suite) delivers documents, Phase 98 (AI Platform Completion) covers agent / RAG observability, and Phase 137.15 delivers game-client crash telemetry — but nothing delivers generic product-analytics instrumentation (event capture, user / account identity, funnels, retention, cohorts, experiments, session replay, heatmaps, revenue attribution, warehouse export) for Oshun-built consumer and SaaS applications.
This phase ships a complete sovereign replacement for Mixpanel, Amplitude, PostHog, Heap, Pendo, Fullstory, Hotjar, LogRocket, Segment, RudderStack, Jitsu, Freshpaint, Google Analytics 4, Adobe Analytics, Kissmetrics, Snowplow, Rudderstack Cloud, June, Countly, Matomo, Plausible, Umami, and Adjust / AppsFlyer / Branch (mobile attribution).
142.1 @neith/metron-core — Event Schema & Ingestion#
142.1.1 Event Model#
- 142.1.1.1 Event envelope (id, time, user, session, context, properties)
- 142.1.1.2 Strongly-typed event schema registry (JSON Schema + codegen)
- 142.1.1.3 Semantic event taxonomy (track / identify / group / alias / page / screen)
- 142.1.1.4 Schema evolution with backfill / alias / rename rules
- 142.1.1.5 Deprecation workflow with usage heatmap
- 142.1.1.6 Required vs. recommended properties per event
- 142.1.1.7 PII tagging at property level
- 142.1.1.8 Unit / currency / timezone property typing
- 142.1.1.9 Nested / array / struct property support
- 142.1.1.10 Event lineage (which SDK / version / route shipped it)
142.1.2 SDKs#
- 142.1.2.1 TypeScript / JavaScript (browser + Node)
- 142.1.2.2 Swift / Objective-C (iOS / macOS / tvOS / watchOS / visionOS)
- 142.1.2.3 Kotlin / Java (Android / JVM server)
- 142.1.2.4 Rust
- 142.1.2.5 Go
- 142.1.2.6 Python
- 142.1.2.7 Ruby
- 142.1.2.8 PHP
- 142.1.2.9 C# / .NET / Unity
- 142.1.2.10 C++ (Neith Engine, Unreal, native)
- 142.1.2.11 Flutter / Dart
- 142.1.2.12 React Native
- 142.1.2.13 Elixir
- 142.1.2.14 CLI / curl-compatible HTTP endpoint
- 142.1.2.15 Snowplow / Segment / GA4 / Mixpanel drop-in compat shims
142.1.3 Ingestion#
- 142.1.3.1 Edge-terminated HTTPS with QUIC + HTTP/3
- 142.1.3.2 Client batching with offline queue
- 142.1.3.3 Gzip / Brotli / Zstd compression
- 142.1.3.4 Protocol buffers for server-to-server
- 142.1.3.5 Webhook / CDC ingestion from stores (Stripe, Shopify, custom)
- 142.1.3.6 S3 / file ingest for backfill
- 142.1.3.7 Server-side tag (reverse-proxy TAG manager)
- 142.1.3.8 Schema enforcement / dead-letter routing
- 142.1.3.9 Rate limit + quota per source key
- 142.1.3.10 Per-region ingest with data-residency routing
142.1.4 Pipeline#
- 142.1.4.1 Stream processor (Kafka / sovereign log / Redpanda / NATS JetStream)
- 142.1.4.2 Enrichment (IP → geo, UA → device, URL → referrer)
- 142.1.4.3 Bot / fraud / test-traffic filter
- 142.1.4.4 PII redact / hash / encrypt per rule
- 142.1.4.5 Consent enforcement (GDPR / CCPA / TCF 2.2)
- 142.1.4.6 Sampling (head / tail / stratified) per workspace
- 142.1.4.7 Dedup window with event-id or (user, event, time)
- 142.1.4.8 Late-arriving event reconciliation
- 142.1.4.9 Per-destination fan-out (warehouse, product views, 3rd-party)
- 142.1.4.10 Replay from durable log for reprocessing
142.2 @neith/metron-identity — Identity & Account Stitching#
- 142.2.1 Anonymous → authenticated identity merge
- 142.2.2 Device graph (device-id ↔ user-id ↔ session)
- 142.2.3 Cross-device identity via deterministic + probabilistic match
- 142.2.4 Group / account / organization / team entity
- 142.2.5 Account-level roll-up for B2B
- 142.2.6 Alias chain (legacy ids keep history)
- 142.2.7 Session definition (time-based, source-change, custom)
- 142.2.8 Identity resolution conflict detection
- 142.2.9 Identity-split UI (manual unmerge)
- 142.2.10 Identity retention / erasure (GDPR right to be forgotten)
142.3 @neith/metron-warehouse — Storage & Query#
142.3.1 Columnar Storage#
- 142.3.1.1 Sovereign ClickHouse-compatible column store
- 142.3.1.2 DuckDB / Parquet native backend for self-host
- 142.3.1.3 S3-compatible object store tier
- 142.3.1.4 Hot / warm / cold tiering with TTL
- 142.3.1.5 Per-event compression / encoding selection
- 142.3.1.6 Shard by workspace + time for elasticity
- 142.3.1.7 Materialized views for common aggregates
- 142.3.1.8 Secondary skip indexes (user-id, event-name, region)
- 142.3.1.9 Data-lake Iceberg / Delta Lake format option
- 142.3.1.10 On-disk encryption with customer-managed keys
142.3.2 Query Layer#
- 142.3.2.1 SQL (Postgres/ClickHouse-flavour) interface
- 142.3.2.2 Product-analytics DSL (funnel, retention, path, flow)
- 142.3.2.3 Query planner with caching + warm-tier routing
- 142.3.2.4 Approximate-count sketches (HyperLogLog, T-Digest, Theta)
- 142.3.2.5 Time-window partition pruning
- 142.3.2.6 Concurrency + cost manager (slot-based)
- 142.3.2.7 Query cancel / timeout / retry
- 142.3.2.8 Explain-plan UI
- 142.3.2.9 Saved query / workbook / snapshot
- 142.3.2.10 MCP / GraphQL / REST / gRPC API
142.4 @neith/metron-insights — Product-Analytics Features#
142.4.1 Core Reports#
- 142.4.1.1 Event count / trend / breakdown
- 142.4.1.2 Funnel (ordered + un-ordered, inline conversion, drop-off)
- 142.4.1.3 Retention (unbounded, bracket, rolling, custom-day)
- 142.4.1.4 Cohort definition language with save / subscribe
- 142.4.1.5 User-path flow (sankey / tree)
- 142.4.1.6 Lifecycle stages (new, current, resurrected, dormant, churned)
- 142.4.1.7 Power-user / engagement score (DAU / MAU / stickiness)
- 142.4.1.8 Sessions (duration, screens, depth, drop)
- 142.4.1.9 Revenue / LTV / MRR / ARR curves
- 142.4.1.10 Feature adoption / reach / stickiness matrix
142.4.2 Advanced Analytics#
- 142.4.2.1 Pattern / anomaly detection (change-point, seasonality)
- 142.4.2.2 Correlation search (property → metric)
- 142.4.2.3 Regression discontinuity / uplift analysis
- 142.4.2.4 Survival analysis with Kaplan-Meier
- 142.4.2.5 Markov / multi-touch attribution
- 142.4.2.6 Propensity scoring
- 142.4.2.7 Churn prediction model
- 142.4.2.8 Next-best-action recommender
- 142.4.2.9 RFM / segmentation clustering
- 142.4.2.10 Cohort-to-cohort comparison
142.4.3 Dashboards#
- 142.4.3.1 Drag-and-drop dashboard canvas
- 142.4.3.2 Embed into Neith Notes / Wiki / Sheets / Slides
- 142.4.3.3 Parameterized template
- 142.4.3.4 Scheduled email / chat digest
- 142.4.3.5 Tile refresh schedule with caching
- 142.4.3.6 Cross-filter / drill-down
- 142.4.3.7 Executive / team / personal space
- 142.4.3.8 Public share with token
- 142.4.3.9 Dashboard-as-code (YAML / TS) version controlled
- 142.4.3.10 North-star-metric hero tile
142.5 @neith/metron-replay — Session Replay & Heatmaps#
- 142.5.1 Canvas-safe DOM replay with mutation capture
- 142.5.2 Audio-off, video-off replay for privacy
- 142.5.3 Mobile screen-replay (iOS / Android) with system-view mask
- 142.5.4 Rage-click / dead-click / u-turn detection
- 142.5.5 Error correlation (console, network, crash, backend trace)
- 142.5.6 Scroll / heat / click / move heatmap
- 142.5.7 Funnel-driven replay (sample failed conversions)
- 142.5.8 Segment-defined replay collection
- 142.5.9 PII auto-mask (IBAN, SSN, card, password, custom regex)
- 142.5.10 Consent-gated recording
- 142.5.11 Shareable time-bookmarked clip
- 142.5.12 Cross-team annotation and comments
- 142.5.13 Replay → ticket (Neith PM) one-click
- 142.5.14 Retention per segment / plan tier
- 142.5.15 On-device edge replay with upload throttle
142.6 @neith/metron-experiment — Feature Flags & Experimentation#
- 142.6.1 Flag service with kill-switch, targeting, rollout %
- 142.6.2 SDK (client + server) with local evaluation
- 142.6.3 Targeting rules: user, account, geo, device, custom attribute
- 142.6.4 Segment-based rollout with percentage ramp
- 142.6.5 Sticky bucketing by user / account
- 142.6.6 A / B / n-arm / multi-variate experiment
- 142.6.7 Feature-flag dependencies and mutually exclusive flag sets
- 142.6.8 Guardrail metrics with automatic pause on regression
- 142.6.9 Statistical engine: frequentist + Bayesian + sequential
- 142.6.10 CUPED / stratification / variance reduction
- 142.6.11 Heterogeneous treatment effect (HTE) discovery
- 142.6.12 Switchback for marketplace experiments
- 142.6.13 Holdout group management
- 142.6.14 Audit log of every flag change
- 142.6.15 Approval + scheduled launch / ramp
142.7 @neith/metron-destinations — Warehouse & Reverse-ETL#
- 142.7.1 Warehouse sinks: BigQuery, Snowflake, Databricks, Redshift, ClickHouse, Postgres, sovereign Neith lake
- 142.7.2 Object-store sinks: S3, GCS, Azure Blob, sovereign
- 142.7.3 Kafka / Pub/Sub / Kinesis / NATS stream export
- 142.7.4 Webhook / HTTP destination
- 142.7.5 CRM destinations: Salesforce, HubSpot, Pipedrive, sovereign CRM
- 142.7.6 Email / SMS / push: Sendgrid, Twilio, OneSignal, sovereign
- 142.7.7 Ad platform destinations: Google Ads, Meta CAPI, TikTok, LinkedIn, Reddit
- 142.7.8 Reverse-ETL (warehouse → SaaS) with model catalog
- 142.7.9 Change-data-capture from app DB (Postgres logical decode)
- 142.7.10 Schema-projection per destination (column allow-list)
- 142.7.11 Exactly-once semantics via idempotency key
- 142.7.12 Back-pressure + retry + dead-letter
- 142.7.13 Cost telemetry per destination
- 142.7.14 Destination health probe
- 142.7.15 Version pinning + canary roll-forward
142.8 @neith/metron-privacy — Consent, Governance & Compliance#
- 142.8.1 Consent-management platform (CMP, TCF 2.2, GPP, US privacy string)
- 142.8.2 Per-purpose / per-vendor consent with audit
- 142.8.3 DSR workflows: access, portability, rectification, erasure
- 142.8.4 Data inventory / lineage map
- 142.8.5 PII classification with regex + ML scanner
- 142.8.6 Field-level encryption with customer-managed keys
- 142.8.7 Key-escrow / BYOK / HYOK
- 142.8.8 SOC 2 / ISO 27001 / HIPAA / PCI / GDPR mapping
- 142.8.9 Data-residency routing (EU, US, APAC, India, China)
- 142.8.10 Privacy-review workflow per new event
- 142.8.11 Automated retention + purge policies
- 142.8.12 Anonymization (k-anonymity, differential privacy)
- 142.8.13 Audit log with non-repudiable hash chain
- 142.8.14 Opt-out endpoints (Do Not Track, GPC)
- 142.8.15 Children's privacy (COPPA, age-gating, parental consent)
142.9 @neith/metron-mobile-attrib — Mobile Attribution & Marketing#
- 142.9.1 SKAdNetwork 4.x server
- 142.9.2 Privacy Sandbox / Attribution Reporting API integration
- 142.9.3 Install + post-install event mapping
- 142.9.4 Deferred deep-link / universal link
- 142.9.5 Custom audience upload + lookalike seeding
- 142.9.6 Cost ingestion from ad networks
- 142.9.7 Fraud / install-farm / click-injection detection
- 142.9.8 Probabilistic attribution (fingerprint-free)
- 142.9.9 View-through vs. click-through budgets
- 142.9.10 Multi-touch attribution modeling
142.10 @neith/metron-surveys — In-App Surveys, NPS & Feedback#
- 142.10.1 Survey builder (multi-type, branching, logic)
- 142.10.2 In-app / email / SMS / chat delivery
- 142.10.3 NPS / CSAT / CES scales with targeting
- 142.10.4 PMF (Superhuman PMF question) template
- 142.10.5 Trigger by event / segment / time
- 142.10.6 Rate-limit per user
- 142.10.7 Language / localization
- 142.10.8 Verbatim analysis (sentiment, theme, summarization)
- 142.10.9 Response → warehouse join to behavioural data
- 142.10.10 Incentive issuance + fraud detection
142.11 Operator Tooling#
- 142.11.1 Workspace admin with teams / roles / SCIM
- 142.11.2 Billing / plan / volume usage metering
- 142.11.3 Self-service quota + plan upgrade
- 142.11.4 Data-governance steward workflow
- 142.11.5 Per-event cost visibility
- 142.11.6 SQL lab with share + permission
- 142.11.7 Alert channel matrix (email, Slack/Chat, PagerDuty, webhook)
- 142.11.8 On-call rotations with paging
- 142.11.9 Per-source health dashboard
- 142.11.10 Status page + incident history
142.12 Migration & Interop#
- 142.12.1 Segment Sources / Destinations protocol compatibility
- 142.12.2 Mixpanel HTTP API compatibility layer
- 142.12.3 Amplitude HTTP API compatibility layer
- 142.12.4 PostHog capture-endpoint compatibility
- 142.12.5 GA4 Measurement Protocol compatibility
- 142.12.6 Snowplow / Enrich schema import
- 142.12.7 Bulk-import from warehouse with schema map
- 142.12.8 Dashboard importer (Mixpanel, Amplitude, Looker, Tableau)
- 142.12.9 Funnel / retention report semantic translator
- 142.12.10 Identity-table import with alias preservation