Ecosystem Surface Deep Dive

BI / MMM / Decision Intelligence.

Where data outputs become budget and growth decisions.

BI tells you what happened. The harder job is deciding what to do next — how to allocate budget, where growth actually comes from, what to stop doing. That is the work of marketing mix modeling, experimentation, and decision intelligence. The constraint is not chart-making. It is whether your metrics mean the same thing everywhere, and whether your models are calibrated against reality.

BI describes the past; MMM, experiments, and decision intelligence turn signals into budget decisions. The strategy is comparable metric definitions and calibration — not another dashboard.

ECOSYSTEM SURFACE Governed data foundation — resolved identity, consent, and first-party data: the inputs every surface draws from. GOVERNED DATA identity & match consent & policy first-party data BI / MMM / Decision Intelligence — Where signals become budget and growth decisions.. Operated by Analytics, data, finance, and marketing-effectiveness leaders.. SURFACE BI / MMM the activation surface operated in governed compute Decision & activation — the surface turns governed signal into a decision you can act on, then measure and feed back. DECISION + ACTION decide activate measure & learn Where governed data becomes a governed decision.
BI / MMM / Decision Intelligence — the lit node in the activation flow: governed data in, a governed decision out. Hover a node for detail.
Decision fit

Fast read.

Best when
You need to turn fragmented outputs into one comparable, calibrated view that drives budget.
Not when
You only need operational reporting, or you have no experiment / ground-truth to calibrate against.
Primary buyer
Analytics, data, finance, and marketing-effectiveness leaders.
Primary output
Calibrated, comparable decisions — budget allocation, growth drivers, what to stop.
Main risk
Confusing a dashboard or an uncalibrated model for a decision.
Best next step
Define metrics once in a semantic layer, then calibrate models with experiments.
Why now

Market context: from reporting to decisions.

  • BI matured from IT-built reports to self-serve and, now, natural-language and copilot-driven analytics.
  • Marketing mix modeling revived as privacy changes weakened user-level tracking — causal, aggregate, durable.
  • Experimentation and incrementality became the calibration layer that keeps models honest.
  • The semantic / metrics layer emerged so a metric means the same thing across every tool and agent.
  • Multi-touch attribution is losing primacy to causal and geo-based methods, though it is still in use.
  • Decision intelligence shifted toward “agentic analytics” — copilots are here; fully autonomous analysts are mostly not.
Evolution

Decision-layer evolution.

Decision-layer evolution A left-to-right path of 6 steps: Dashboards → Self-serve BI → MMM revival → Experiment calibration → Semantic metrics layer → Agentic decisioning. 1 Dashboards 2 Self-serve BI 3 MMM revival 4 Experimentcalibration 5 Semantic layermetrics 6 Agenticdecisioning
Decision-layer evolution
  1. 01

    Dashboards

    IT-built reports describing what happened.

  2. 02

    Self-serve BI

    Analysts and business users explore data directly.

  3. 03

    MMM revival

    Aggregate, causal budget allocation resilient to privacy change.

  4. 04

    Experiment calibration

    Geo tests and incrementality keep models honest.

  5. 05

    Semantic metrics layer

    One definition of each metric across tools and agents.

  6. 06

    Agentic decisioning

    NL copilots assist; autonomous, governed agents emerge.

Landscape

Who plays here — examples, not a ranking.

Named as examples, not a ranking. BI describes, MMM and experiments explain, the semantic layer keeps metrics consistent. Validate current names — AI-copilot features are largely GA, but fully autonomous analyst agents are mostly still emerging.

BI / analytics (with AI copilots)
  • Looker (Gemini in Looker)
  • Power BI (Copilot)
  • Tableau (Tableau Agent + Pulse)
  • ThoughtSpot (Spotter)
  • Snowflake (Cortex Analyst)
  • Databricks (AI/BI Genie)
  • Amazon Q in QuickSight
  • Sigma
  • Omni
Marketing mix modeling
  • Google Meridian
  • Meta Robyn
  • Recast
  • Mass Analytics
  • Analytic Partners
  • Nielsen
  • Mutinex
  • Keen
  • Cassandra
Incrementality & experimentation
  • Measured
  • Haus
  • INCRMNTAL
  • Meta GeoLift (open source)
  • Optimizely
  • Statsig
  • Northbeam / Rockerbox (MTA)
Semantic metrics & decision intelligence
  • dbt Semantic Layer (MetricFlow)
  • Cube
  • AtScale
  • Tellius
  • Pyramid Analytics
Capability map

What it does — and where it quietly fails.

What to weigh — and where it bites. Validate current support per platform.

CapabilityWhat it meansWhy it mattersWatch-out
Descriptive BIDashboards and reports.See what happened.Describing is not deciding.
Self-serve / NL queryAsk questions in language.Speed and access.Answers need a trusted semantic layer.
Semantic / metrics layerOne definition per metric.Comparability everywhere.Without it, every tool disagrees.
Marketing mix modelingAggregate causal allocation.Privacy-durable budget guidance.Useless if not calibrated.
Experiment calibrationGeo tests, holdouts.Ground-truth for models.Bad test design misleads.
Incrementality / liftCausal contribution.Real impact, not credit.Needs valid controls.
Multi-touch attributionUser-path credit.Granular, familiar.Losing trust; weak post-privacy.
Forecasting / scenarioPlan and simulate.Budget planning.Confidence intervals matter.
Data freshness / lineageTimeliness and provenance.Trust the number.Stale or untraceable inputs.
Agentic / copilotNL + assisted analysis.Faster exploration.Copilots GA; autonomous agents emerging.
Governance / accessWho sees / changes what.Trust and control.NL access can bypass governance.
Decision workflowFrom insight to action.Close the loop.Insight without action is shelfware.
First-party data

How the ecosystem feeds the decision layer.

This surface is where every other surface’s output is supposed to become comparable. That only works if the inputs share definitions.

  1. From clean rooms

    • Privacy-safe overlap and outcome measurement
    • Incrementality inputs for calibration
  2. From CDP / RMN / DSP

    • Spend, conversions, and audience outcomes
    • Channel-level results for the mix model
  3. From experiments

    • Geo tests and holdouts as ground truth
    • Calibration for MMM and attribution
  4. To the decision

    • One comparable view across channels
    • Budget allocation and growth drivers
How it connects

What feeds the decision layer — and the catch.

InputHow it feeds BI / MMMWatch-out
Clean-room outputsPrivacy-safe outcomes and overlapOutput policy and comparability
Media spend + resultsChannel inputs for the mix modelDefinitions must match across sources
ExperimentsGround-truth calibrationValid test design required
CDP / RMN signalsAudience and sales outcomesSelf-attributed metrics differ
Semantic layerOne definition per metricSkip it and tools disagree
NL / agent queriesFaster explorationNeeds lineage and governance
Agentic shift

What agentic analytics changes.

Agentic analytics promises an analyst that answers questions, monitors metrics, and proposes actions in language. Today the copilots — natural-language query and assisted analysis — are largely generally available and human-supervised. Fully autonomous analysts that plan multi-step investigations and act on governed data are mostly still emerging.

Agent use cases
  • Natural-language query and exploration
  • Automated metric monitoring and alerting
  • Assisted MMM and scenario planning
  • Draft insights and narratives
  • Anomaly detection
  • Proposed budget reallocation (human-approved)
Agent risks
  • Confident but wrong answers without lineage
  • MMM treated as truth without calibration
  • Dashboards / agents mistaken for decisions
  • Inconsistent metric definitions across tools
  • NL access bypassing governance
Governance needed
  • A semantic layer as the single source of metric truth
  • Calibration against experiments
  • Lineage and provenance on every number
  • Access and governance on NL / agent queries
  • Human approval before budget moves
Dashboards are not decisions

The most expensive mistake in this layer is mistaking description for decision. A dashboard shows what happened; a decision says what to do and why. Two disciplines bridge the gap: a semantic layer so every metric means one thing, and calibration so models are checked against real experiments. AI copilots make the description faster — but a faster wrong answer is still wrong. Calibration and comparability are the work; the chart is not.

Strategic read

SWOT.

Strengths
  • Causal, privacy-durable MMM
  • Semantic comparability
  • Faster NL exploration
  • Experiment-based ground truth
Weaknesses
  • Dashboards mistaken for decisions
  • Uncalibrated models
  • Inconsistent definitions
  • MTA losing reliability
  • Insight that never reaches action
Opportunities
  • Semantic layer as source of truth
  • Calibrated MMM + experiments
  • Agentic copilots on governed data
  • Unified cross-channel view
  • Decision workflows, not just reports
Threats
  • Confident-but-wrong AI output
  • Metric fragmentation
  • Over-trust in autonomous agents
  • Privacy limits on tracking
  • Governance bypass via NL access
Reference architecture

BI / MMM / Decision Intelligence.

BI / MMM / Decision Intelligence A vertical flow of 8 stages, top to bottom: Inputs: clean-room outputs · media spend + results · CDP / RMN / DSP signals · experiments → Semantic / metrics layer (one definition per metric) → BI (descriptive) · MMM (causal allocation) · experiments (calibration) → Decision-intelligence / agentic copilots → Calibration + lineage + governance checks → Decisions: budget allocation · growth drivers · what to stop → Action back into CDP / DSP / RMN → Outcomes feed the next round of calibration. 01 Inputs: clean-room outputs · media spend +results · CDP / RMN / DSP signals · experiments 02 Semantic / metrics layer (one definition permetric) 03 BI (descriptive) · MMM (causal allocation) ·experiments (calibration) 04 Decision-intelligence / agentic copilots 05 Calibration + lineage + governance checks 06 Decisions: budget allocation · growthdrivers · what to stop 07 Action back into CDP / DSP / RMN 08 Outcomes feed the next round of calibration
Running through
  • Semantic source of truth
  • Calibration
  • Lineage
  • Governance
  • Human approval
BI / MMM / Decision Intelligence
Output-led decision rules

Design backward from the output.

Output neededBetter-fit patternWatch-out
Budget allocationCalibrated MMM + experimentsMMM without calibration misleads.
Comparable cross-channel viewSemantic layer over all sourcesSkip it and tools disagree.
Fast explorationNL / copilot on governed dataCopilots GA; autonomous agents emerging.
Causal proofGeo tests + incrementalityValid control design required.
Decision, not reportInsight wired to a workflowInsight without action is shelfware.
First moves

What to build first.

  1. 01

    A semantic / metrics layer so every metric means one thing everywhere.

  2. 02

    A calibration habit — experiments and holdouts that keep MMM honest.

  3. 03

    Lineage on every number so NL / agent answers can be trusted.

  4. 04

    A decision workflow so insight actually changes budget, not just slides.

Anti-patterns

Where this goes wrong.

  • Mistaking a dashboard for a decision.
  • Running MMM without experiment calibration.
  • Letting each tool define metrics its own way.
  • Trusting NL / agent output without lineage or governance.
  • Treating multi-touch attribution as ground truth.
POC to production

12 questions before the POC becomes production.

  1. 01
    Business decision

    What single decision does this surface improve?

  2. 02
    Data inputs

    What data feeds it, who owns it, and where does it live?

  3. 03
    Identity logic

    How are people / accounts / SKUs resolved and matched?

  4. 04
    Consent / governance

    What is the consent basis and the output policy?

  5. 05
    Metric definition

    Are the metrics defined, owned, and comparable?

  6. 06
    Output policy

    What can leave — aggregate, score, segment, report, API?

  7. 07
    Activation rights

    Is the output eligible to activate, and where?

  8. 08
    Measurement method

    How is the result measured, and is the method defensible?

  9. 09
    Technical owner

    Who builds and runs the pipeline?

  10. 10
    Commercial owner

    Who owns the budget / commercial outcome?

  11. 11
    Feedback loop

    How do results flow back into the model and the decision?

  12. 12
    Production path

    What turns the POC into a governed, repeatable workflow?

Watch-outs

Practical caveats.

  1. 01

    A faster wrong answer is still wrong — copilots speed up description, not judgement.

  2. 02

    MMM is only as good as the experiments you calibrate it against.

  3. 03

    If metrics aren’t defined once, every tool will disagree — and so will every agent.

  4. 04

    Multi-touch attribution is losing reliability; lean on causal and geo methods.

  5. 05

    Autonomous analyst agents are mostly emerging — keep a human on budget moves.

Capability validation note

Product names, ownership, and availability across these surfaces change quickly. Treat this as an advisory fit guide, not procurement documentation — validate current capabilities and access against official sources before implementation.

Market references last validated: June 6, 2026. Revalidate before pitch use.

Need help connecting this surface to the operating model?

The surface only creates value when data, semantics, governance, activation, and measurement are designed together.