The new marketer-first interface that combines data, collaborators and conversations. Adobe marks several parts of it as in development or coming soon.
Last checked 2026-09-28
CX Enterprise atlas / Customer Engagement | Data & analytics
ProductGenerally available
Customer EngagementData & analytics
Also known as Mix Modeler, Adobe Mix Modeler, AMM
Adobe's marketing measurement and planning product in CX Analytics, and the new name for Adobe Mix Modeler (renamed at Summit, April 2026). It unifies marketing mix modeling with multi-touch attribution, plans and optimizes budgets with scenario planning, and adds a conversational, collaborative interface, prebuilt paid media connectors and sharing of results with Customer Journey Analytics and Content Analytics. Experience League still documents it as Mix Modeler.
The new marketer-first interface that combines data, collaborators and conversations. Adobe marks several parts of it as in development or coming soon.
Last checked 2026-09-28
Natural-language questions that return analysis, performance drivers and recommended actions from Adobe's AI. Not flagged as upcoming on the product page, but Experience League does not document it yet.
Last checked 2026-09-28
A shared space where teams comment on, share and assign insights, tag teammates to validate findings and track decisions, with role-based views for performance, growth and analytics teams. Marked "in development" and "coming soon".
Last checked 2026-09-28
A role-specific stream of insights, metrics and causal drivers surfaced without manual exploration. Marked "in development".
Last checked 2026-09-28
Opens an insight directly in another CX Analytics app, such as Customer Journey Analytics, with its context carried over. Part of the teamwork features marked "coming soon".
Last checked 2026-09-28
AI that flags risks and opportunities across campaigns (rising CPA, wasted spend, high-impact channels) and gives agentic recommendations for reallocating budget in flight. Marked "coming soon".
Last checked 2026-09-28
Consolidates and harmonizes paid, owned and earned marketing data with business factors and measures it with one framework that unifies top-down and bottom-up methods. Mix Modeler provides the underlying harmonization and models.
Last checked 2026-09-28
Ingestion from advertising, commerce, CRM and data warehouse platforms through integrations and APIs, plus external and internal business factors, harmonized into common fields, channels and touchpoints.
Last checked 2026-09-28
Marketing mix modeling on aggregate data and multi-touch attribution on event data, optionally calibrated with priors and combined (Adobe says with patent-pending technology) into one consistent set of results.
Last checked 2026-09-28
Experience Platform data usage labeling, policy enforcement with alerts, and role-based permissions applied to Marketing Campaign Analytics data.
Last checked 2026-09-28
Add-on with extra tools for data lifecycle and expiration, automated consent management and policy enforcement. In Mix Modeler it also enables customer managed keys.
Last checked 2026-09-28
Automated ingestion of paid media data into a model-ready system, for incrementality, delivery and pacing insights down to the ad level.
Last checked 2026-09-28
Prebuilt connectors to ad platforms, including Google and Meta, that handle authentication, schema mapping and scheduled sync of campaign, spend and performance data. They are Experience Platform advertising sources (Google Ads is in beta there).
Last checked 2026-09-28
A consistent, customizable paid media schema that keeps each publisher's campaign hierarchy and taxonomy down to the creative asset, shared across Adobe applications.
Last checked 2026-09-28
Reuses ingested paid media data and modeled results in other Adobe Analytics products, such as Customer Journey Analytics for journey analysis and Content Analytics for creative and content performance.
Last checked 2026-09-28
The landing screen of Mix Modeler, with configurable dashboards of harmonized data and plan performance.
Last checked 2026-09-28
Four KPI status cards and six configurable visualizations of harmonized data, filterable by brand, campaign, channel type, conversion type, dataset, media type, source type and traffic source.
Last checked 2026-09-28
A plan-tracking dashboard with budget, revenue, ROI and KPI status cards and six configurable actual-versus-planned visualizations, for spotting gaps and adjusting live plans. Marked beta in the documentation; extended as "Plans tracking" in November 2025.
Last checked 2026-09-28
Mix Modeler reads datasets in Experience Platform that are based on the XDM ExperienceEvent or XDM Summary Metrics class. Data arrives through any Platform ingestion method (SDKs, APIs, source connectors, batch or streaming), and the app links to Platform's Schemas and Datasets screens.
Last checked 2026-09-28
Event-level touchpoint and conversion data with a timestamp and person ID, for example from the Adobe Analytics source connector, Web or Mobile SDK, Edge Network Server API or other source connectors. It enables multi-touch attribution.
Last checked 2026-09-28
Aggregate data such as walled-garden spend and delivery (Facebook, YouTube), offline advertising and TV flight schedules, ingested with schemas based on the XDM Summary Metrics class. Date and granularity are mandatory, and channel is mandatory for marketing spend data.
Last checked 2026-09-28
Internal factors (promotions, price changes, holidays) and external factors (CPI, S&P 500, weather, competitor indexes). Datasets whose schema uses the Factor Standard Fields field group (factor type, value type, granularity) appear as factor datasets in dataset rules.
Last checked 2026-09-28
Experience Platform advertising sources that bring campaign, ad set and ad-level spend and delivery data into standardized datasets that Mix Modeler can use for modeling.
Last checked 2026-09-28
Ingests Facebook and Instagram performance data through the Meta Marketing (Insights) APIs on a schedule. Adobe names media mix modeling in Mix Modeler as a primary use case.
Last checked 2026-09-28
A paid media source connector that ingests Google Ads account, campaign, ad group, ad, asset and performance data through the Google Ads API into standard paid media XDM structures in Experience Platform, read-only, for cross-channel analysis and media mix modeling.
Last checked 2026-10-20
The harmonization service merges summary and event data into one consistent data view at the coarsest granularity found across the datasets. This view is the input for all models, and it can be browsed and downloaded as CSV.
Last checked 2026-09-28
Common dimensions and metrics that map differently named source fields to one concept. Global fields (for example channel, campaign, brand, clicks, impressions, spend, orders, net sales) are built in, and you can add your own.
Last checked 2026-09-28
Map fields of summary, factor and event datasets to harmonized fields, directly or with conditions. Each rule records the source (Adobe Analytics, Experience Events, Summary or Consumer Experience Events), granularity and start of week. Mappings are validated and Platform field-level access control applies.
Last checked 2026-09-28
Ad-hoc or scheduled sync that rebuilds harmonized data from the source datasets using the current dataset rules. Deleted source datasets drop out of harmonized data on the next sync.
Last checked 2026-09-28
Resolves conflicts when summary and event data update the same metric: use summary data, event data, or their sum by default, with overrides per metric and conversion type. Marked beta.
Last checked 2026-09-28
For aggregate datasets, sets whether restated data is added up (aggregation) or overwrites earlier values (replacement).
Last checked 2026-09-28
Internal and external factors are defined as harmonized fields in dataset rules for factor datasets, synced, and then selected in model configuration. This replaced the older dataset-based factor workflow in February 2026.
Last checked 2026-09-28
Definitions of the marketing events (by recipient, individual or cookie) whose impact is measured, each with a spend metric, a volume metric and a rule over harmonized fields. Each touchpoint has its own report.
Last checked 2026-09-28
Definitions of the business outcomes to measure (orders, in-store purchases, visits), each with a conversion metric, optional revenue metric and category. Each conversion has its own report.
Last checked 2026-09-28
Shows how many models use the harmonized data and how many conversions it contains over the last five years, by year or month.
Last checked 2026-09-28
Custom machine learning models, configured, trained and scored against harmonized data, that measure and predict a conversion outcome from marketing investment. A model needs one conversion, one or more touchpoints with at least one spend channel, a lookback window and a training window.
Last checked 2026-09-28
The modeling approaches Mix Modeler combines, described in Adobe's deep dive.
Last checked 2026-09-28
A top-down model on aggregate data that measures the incremental impact of online and offline marketing and business factors on conversions. It uses a multiplicative structure (baseline times media multipliers), which captures channel synergy and diminishing returns.
Last checked 2026-09-28
Optional bottom-up attribution of credit to event-level touchpoints (for example ad clicks, email opens, paid search clicks) on trackable journeys, enabled per model. It cannot measure offline channels or business factors.
Last checked 2026-09-28
Influenced scores (the share of a conversion each touchpoint is responsible for) and incremental scores (the influenced score minus the baseline that would have happened without marketing).
Last checked 2026-09-28
First touch, last touch, linear, U-shaped and time-decay attribution, shown alongside the algorithmic scores for comparison.
Last checked 2026-09-28
Feeds scored MTA results into the MMM as priors and feeds MMM results back to update MTA scores, so touchpoint-level and aggregate results stay consistent. The same mechanism blends in experimentation results and other prior knowledge.
Last checked 2026-09-28
A guided model canvas with Setup, Configure, Advanced and Set options steps. Models can be duplicated to create variants for different outcomes such as revenue, units sold or leads.
Last checked 2026-09-28
Filters, built from containers with nested AND/OR conditions, that restrict scoring to a subset of the harmonized data.
Last checked 2026-09-28
The lookback window (1 to 52 weeks) sets which touchpoints before a conversion get credit. The training window is chosen by Mix Modeler or set manually in years.
Last checked 2026-09-28
Internal and external factors from harmonized factor data, added to a model to separate baseline effects from marketing effects.
Last checked 2026-09-28
Optional priors for touchpoint contributions from past models, incrementality tests or stakeholder experience, applied through transfer learning.
Last checked 2026-09-28
Lag and minimum or maximum lookback (carryover) weeks per channel, capped at one eighth of the training window, so domain knowledge or test results shape how long each channel's effect lasts. Added March 2026.
Last checked 2026-09-28
Harmonized fields (for example campaign, media type, region, traffic source) chosen to break down conversion and touchpoint incrementality scores, instead of building separate models. Added February 2026.
Last checked 2026-09-28
A new model is trained and scored immediately. Afterwards it can be retrained on new data or scored incrementally or rescored for a date range, ad hoc or on a schedule. Rescoring does not update existing plans.
Last checked 2026-09-28
Warns when a model has drifted and prompts a retrain, from model insights and diagnostics.
Last checked 2026-09-28
Visualizations for trained and scored models, organized in tabs, that quantify incremental impact and show which channels perform or need optimization. Insights download as CSV or Excel.
Last checked 2026-09-28
The main tab: contribution by date split into base, non-spend and spend channels; contribution by channel (all or paid media); a marketing performance summary of ROI, ROAS or CPA per channel; and a weekly touchpoint breakdown.
Last checked 2026-09-28
Per-channel curves of marginal return against spend, marking the current spend point, the point of diminishing returns and the marginal break-even point.
Last checked 2026-09-28
A channel synergy heatmap (next-dollar synergy between pairs of spend channels, scaled 0–10) and per-channel adstock decay curves.
Last checked 2026-09-28
The Factors tab: the incremental effect of selected internal and external factors on the conversion baseline. Marked beta.
Last checked 2026-09-28
For MTA-enabled models: conversion overview and trends by attribution model, channel and touchpoint breakdowns, top campaigns, starter/player/closer position analysis and top conversion paths.
Last checked 2026-09-28
Model assessment charts of actual, predicted and residual values, training fit metrics per conversion (such as R², MAPE and RMSE), and a touchpoint effectiveness table.
Last checked 2026-09-28
Historical conversion and spend by fiscal quarter and product, spend by channel, and touchpoint spend and volume by quarter.
Last checked 2026-09-28
Scoring writes results back to Experience Platform datasets for use in other tools: aggregate scores (AMM AI Aggregate Scores) and, for MTA models, event-level scores (Attribution AI Scores), each with its own schema.
Last checked 2026-09-28
Incremental impact by channel, campaign and media type across the funnel, and touchpoint-level attribution when event data is available from Adobe Analytics or Customer Journey Analytics.
Last checked 2026-09-28
Model assessment charts, quality metrics (R², MAPE, RMSE), training fit metrics and automatic model drift notifications.
Last checked 2026-09-28
Self-service creation of model variants with their own touchpoints, date ranges, priors and factors, which can be retrained and rescored at any time.
Last checked 2026-09-28
Scenario plans that allocate a budget across channels and flight dates, optimized against a trained model. A plan needs a model, a date range and a budget or target. Plans can be duplicated, edited and rebased on another model.
Last checked 2026-09-28
Plans for "I have a budget to spend": enter budgets for one or more date ranges, optionally capped by a maximum budget, and get an AI-recommended allocation that maximizes the chosen conversion.
Last checked 2026-09-28
Plans for "I have a target to achieve": set a conversion, revenue, CPA or ROI target and Mix Modeler finds the spend to reach it, showing whether the target is achievable. Release notes list it as available since July 2025, but the option carries a beta badge.
Last checked 2026-09-28
Numerical optimization on the multiplicative MMM that finds the best budget mix under total and per-channel minimum and maximum constraints, including weighted optimization across several conversions. Marginal ROI or CPA can be balanced month by month.
Last checked 2026-09-28
Sets how budget is spread within each date range: automatically, or following a historical reference period of past spend.
Last checked 2026-09-28
Overrides for average revenue per conversion and for channel costs per date range, which Mix Modeler otherwise derives from recent seasonal history.
Last checked 2026-09-28
Forecasts for a plan: paid channel spend against ROI, forecasted revenue and conversions by channel, marginal channel return with break-even point, channel synergies, and a weekly or summary breakdown of ROI, CPA, revenue, conversions and spend.
Last checked 2026-09-28
Side-by-side view of plans with budget, channel budgeting mode (AI automated or custom plus AI), planned spend, and forecasted and paid-marketing ROI or CPA and revenue or conversions.
Last checked 2026-09-28
Plans defined by flight dates, total budget, optional channel constraints, value and cost assumptions and a base model, then created, duplicated and edited in a self-service interface.
Last checked 2026-09-28
Compares live performance with plan forecasts from continuously ingested spend and results, so budgets can shift while campaigns run. Its documented basis, Mix Modeler's Performance to plan dashboard, is in beta.
Last checked 2026-09-28
Mix Modeler inherits Experience Platform governance: access control, audit logs, data usage policy enforcement, encryption at rest and in transit, and data hygiene on input and output datasets.
Last checked 2026-09-28
Role-based permissions set in Experience Platform Permissions: roles scoped to sandboxes get the Adobe Mix Modeler resource with view or manage rights on harmonized data, models configuration and plans configuration, plus Platform dataset and schema permissions. Attribute-based access control hides restricted schema fields in dataset rules.
Last checked 2026-09-28
Create, update and delete actions on dataset rules, fields, touchpoints, conversions, models and plans, plus retraining, rescoring and harmonization syncs, recorded in Experience Platform audit logs.
Last checked 2026-09-28
Checks dataset rules and model configurations against data usage labels and policies (by default C8 analytics, C9 data science and C12 data export) and reports violations with lineage.
Last checked 2026-09-28
Bring-your-own encryption keys through Azure Key Vault for data processed in Mix Modeler models. Available when Mix Modeler is licensed with the Privacy and Security Shield add-on.
Last checked 2026-09-28