# SegmentStream AI

> Infrastructure that gives AI agents a marketing measurement brain.
> MCP server with built-in measurement skills for attribution, budget optimization,
> incrementality testing, and more across 14+ ad platforms.

## Pages

- [Home](https://segmentstream.com/index.md): AI-native marketing measurement infrastructure — connect your AI agent to SegmentStream's MCP server for cross-channel attribution, budget optimization, and incrementality testing.
- [AI Skills](https://segmentstream.com/ai-skills.md): Copy-pasteable prompts for SegmentStream onboarding, data quality, attribution diagnosis, measurement theory, and Measurement Engine capabilities.
- [Measurement Engine](https://segmentstream.com/measurement-engine.md): Nine whitepapers explaining every active algorithm in the SegmentStream Measurement Engine.
- [CLI](https://segmentstream.com/cli.md): Open-source CLI (private beta) for warehouse-native marketing attribution — connect BigQuery, Snowflake, or Databricks, add any data source, configure attribution in YAML, and run the Measurement Engine locally so data never leaves your warehouse.
- [Integrations](https://segmentstream.com/integrations.md): Ad platforms, analytics, CRM, commerce, AI clients, and warehouse integrations supported by SegmentStream.
- [Pricing](https://segmentstream.com/pricing.md): Three plans — Online (Service-as-a-Software, from $800/mo), Full Funnel (Service-as-a-Software, from $1,200/mo), and Enterprise (Human-Led, from $5,000/mo). Each plan has a Basic and Pro tier on the same engine.
- [For Agencies](https://segmentstream.com/agency.md): AI-native measurement infrastructure for agencies and independent consultants. Scale measurement expertise across every client.
- [About](https://segmentstream.com/about.md): SegmentStream is an independent marketing measurement company. The six principles we work by — independence from ad platforms, evidence over modeling, no black boxes, built for action, a system not a tool, and education over selling.
- [Course](https://segmentstream.com/course.md): On-demand course by Constantine Yurevich. 9.5 hours of video lessons across four modules on attribution, MMM, incrementality, and LTV-based measurement — from first principles. $299 one-time, lifetime access.
- [M&A Interest](https://segmentstream.com/ma-interest.md): Information for strategic acquirers and corporate development teams. SegmentStream is selectively open to acquisition conversations where the combined offering beats the sum of its parts, with strategic-fit rationale across 11 acquirer categories.
- [FAQ](https://segmentstream.com/faq.md): Answers to questions not covered on other pages — custom conversion values, data warehouse options, Meta restricted-category optimization, HIPAA/BAA position, retroactive conversion updates, and implementation timelines.

## Content Indexes

- [Blog](https://segmentstream.com/blog.md): Articles, product updates, and insights on marketing measurement, attribution, and budget optimization.
- [Blog Articles](https://segmentstream.com/blog/articles.md): Long-form marketing measurement articles.
- [Product Updates](https://segmentstream.com/blog/product-updates.md): SegmentStream product announcements and updates.
- [Company News](https://segmentstream.com/blog/company-news.md): SegmentStream company announcements.
- [Glossary](https://segmentstream.com/glossary.md): Plain-English definitions for marketing measurement concepts.
- [Podcasts](https://segmentstream.com/resources/podcasts.md): Podcast episodes and conversations about measurement and attribution.

## Trust

- [Trust & Security](https://segmentstream.com/trust.md): How SegmentStream handles your customer data — storage options, compliance framework, security practices, confidentiality obligations, and FAQ.

## Integrations

- [Google Analytics 4](https://segmentstream.com/integrations/google-analytics-4.md): How SegmentStream integrates with Google Analytics 4: reads raw event data from the GA4 → BigQuery export to power cross-channel attribution and budget optimization.
- [Adobe Analytics](https://segmentstream.com/integrations/adobe-analytics.md): How SegmentStream integrates with Adobe Analytics: reads raw hit-level data from Adobe Data Feeds delivered to a SegmentStream-provisioned Google Cloud Storage bucket, then layers attribution and budget optimization on top.
- [Heap by Contentsquare](https://segmentstream.com/integrations/heap-contentsquare.md): How SegmentStream integrates with Heap by Contentsquare: reads raw event data from Heap Connect — exported to your data warehouse — and layers identity graph and cross-channel attribution on top.
- [Amplitude](https://segmentstream.com/integrations/amplitude.md): How SegmentStream integrates with Amplitude: reads raw event data exported from Amplitude into your data warehouse, then layers identity graph and cross-channel attribution on top.
- [Mixpanel](https://segmentstream.com/integrations/mixpanel.md): How SegmentStream integrates with Mixpanel: reads raw event data exported via Mixpanel Data Pipelines into your data warehouse, layered with identity graph and cross-channel attribution.
- [PostHog](https://segmentstream.com/integrations/posthog.md): How SegmentStream integrates with PostHog: reads raw event data via PostHog Batch Exports into your data warehouse, layered with identity graph and cross-channel attribution.
- [Twilio Segment](https://segmentstream.com/integrations/twilio-segment.md): How SegmentStream integrates with Twilio Segment: reads raw event data routed by Segment into your data warehouse, layered with identity graph and cross-channel attribution.
- [Snowplow](https://segmentstream.com/integrations/snowplow.md): How SegmentStream integrates with Snowplow: reads raw enriched event data loaded by Snowplow into your data warehouse, layered with identity graph and cross-channel attribution.
- [RudderStack](https://segmentstream.com/integrations/rudderstack.md): How SegmentStream integrates with RudderStack: reads raw event data routed by RudderStack into your data warehouse, layered with identity graph and cross-channel attribution.

## Articles

- [Identity Graph](https://segmentstream.com/measurement-engine/identity-graph.md): Deterministic identity stitching across devices and touchpoints — login events, email matching, CRM joins — unified into a single user profile in BigQuery. The foundational data layer that cross-channel attribution and scoring depend on.
- [Cross-Channel Attribution](https://segmentstream.com/measurement-engine/cross-channel-attribution.md): First-click attribution and predictive conversion maturation — two methods that cross-validate to produce click-time revenue you can optimize against before all conversions are reported.
- [Predictive Cross-Channel Attribution](https://segmentstream.com/measurement-engine/predictive-attribution.md): ML-powered projection that fills the click-time attribution gap. See projected ROAS within days, not weeks — while cohorts are still maturing.
- [Self-Reported Reattribution](https://segmentstream.com/measurement-engine/self-reported-reattribution.md): LLM classification of free-text survey responses, stitched to sessions via BigQuery. Measures channels traditional tracking can't see — word of mouth, podcasts, offline.
- [CRM Funnel Attribution](https://segmentstream.com/measurement-engine/crm-funnel-attribution.md): Connect CRM, ERP, or any BigQuery table as a conversion source. Track every funnel stage from MQL to Closed/Won — the ground truth that calibrates predictive models and reconciles true campaign ROI.
- [Marginal Analytics](https://segmentstream.com/measurement-engine/marginal-analytics.md): Diminishing returns curve fitting that models marginal ROAS per channel at every spend level. See exactly where each dollar stops working, calculate the optimal budget mix, and get specific reallocation recommendations.
- [Automated Budget Allocation](https://segmentstream.com/measurement-engine/automated-budget-allocation.md): One-click execution of budget recommendations across every ad platform. Applies changes from Marginal Analytics, tracks prediction accuracy, and refines the model with reinforced learning.
- [Incrementality Testing](https://segmentstream.com/measurement-engine/incrementality.md): Geo holdout experiments with synthetic control groups. Three-phase market selection, A/A validation, and statistical rigor that measures what your ads actually add.
- [Server-Side Conversion Tracking](https://segmentstream.com/measurement-engine/server-side-tracking.md): Stop in-page pixels from broadcasting your audience to competitors. Forward only qualifying conversions to each ad platform server-to-server, with the hashed CRM identifiers that unlock cross-device advanced matching.