Moby recommendations, informed by Compass
Moby brings Compass context into the decision.
Compass brings first-party attribution, marketing mix modeling, and incrementality into one measurement system. When Moby 2 and Compass are both available, Moby may use that context to explain disagreements, support cross-channel analysis, and prepare more informed recommendations for review.
Attribution
Marketing mix
Lift tests
Example Moby recommendation
Do the broader signals support scaling paid social?
Moby may use platform performance, first-party attribution, MMM, and incrementality context to explain whether the next step is monitoring, deeper investigation, or a reviewed budget recommendation.
Investigate first
Review before
material action
material action
Illustrative motion — available signals and Moby capabilities depend on account setup and rollout
Compass organizes first-party attribution, MMM, and incrementality
Moby may use Compass context in recommendations
Act where measurement signals agree
Investigate where they diverge
Review material budget actions before approval
Compass organizes first-party attribution, MMM, and incrementality
Moby may use Compass context in recommendations
Act where measurement signals agree
Investigate where they diverge
Review material budget actions before approval
03 / The measurement system
Moby can use three distinct measurement signals.
First-party attribution provides a faster read on customer journeys and orders. MMM estimates how channels contribute to business outcomes over time. Incrementality tests whether marketing caused lift. Compass organizes the signals, confidence, and disagreement; Moby may use that context where supported.
Compass does not force every source to match; it helps teams use the right signal for the decision
04 / First-party attribution (MTA)
Moby can read the journeys shaping today.
First-party attribution uses Triple Whale’s connected customer journey and order data to show touchpoint-level paths across available attribution models. It is a fast operational signal, not causal proof. Where supported, Moby can use it alongside MMM and incrementality context.
Every signal, one readable sequenceMoby follows the path before diagnosing the outcome
M
Moby can reconstruct the path, then put it in measurement context.First-party attribution explains the journey. MMM estimates aggregate contribution. Incrementality tests causal lift.
Journey resolved
Klaviyo → Google → Meta → TikTok → AppLovin → Website → Shopify
05 / Marketing mix modeling
Moby can use MMM to understand where the next dollar may work.
MMM uses aggregate historical data and a Bayesian framework to estimate baseline revenue, channel contribution, carryover, efficiency, and diminishing returns across online, offline, and hard-to-track channels where data is available. Its weekly-updated results support cross-channel planning and scenarios, not real-time campaign management. Check model fit and historical spend range before acting.
Incremental contribution
Spend
Efficiency ratioDeduplicatedModeled incremental contribution per dollar across the historical window.
Marginal returnNext dollarThe expected incremental return from the next spend increment.
SaturationCurve slopeFlattening indicates diminishing returns and less efficient additional spend.
MMM is strongest within historical spend ranges; roughly 12 months of history is the minimum, with 18–24 recommended. It is not a week-by-week revenue forecast.
06 / Incrementality testing
Moby can recommend when the next answer requires a causal test.
Incrementality compares test and control groups to estimate results that would not have happened otherwise. Triple Whale supports GeoLift and Meta Conversion Lift workflows where available. Moby may use Compass context to recommend deeper investigation or validation, but test design and launch still depend on supported product capabilities and review.
Test regionsCampaign active
Observed resultActual KPI
Compare with counterfactualControl regionsSpend paused or reduced
Synthetic controlExpected KPI
Matched before launchIllustrative GeoLift design — feasibility and power depend on KPI history, regions, duration, expected iROAS, and budget
07 / Confidence and disagreement
Compass surfaces what to trust and what to investigate.
MMM provides the primary modeled read for channels that have not been incrementality tested. Incrementality results calibrate and cross-check that read, while first-party order data grounds the calculations. Compass does not silently average meaningful disagreement; it surfaces the gap so teams can act where signals agree and investigate where they diverge. Moby may help explain that context.
First-party groundingConnected order and customer journey data.
Grounding
Primary model readBaseline, contribution, efficiency, model fit, and saturation.
Modeling
Calibration signaliROAS, lift percentage, and incremental results from a controlled test.
Calibrating
Review confidenceSee whether the channel read is well-supported, directional, or model-based only.
Investigate disagreementReview timing, model history, and test design instead of silently averaging the signals.
Ask Moby to explainWhere supported, Moby may translate Compass context into analysis and a recommendation for review.
Business goalWhat decision matters now
Model qualityFit, history, and uncertainty
Validation historyTests completed or still needed
GovernanceHuman review for material actions
Review channel context, confidence, and disagreement in the Compass Command Center; availability varies by setup and rollout
08 / Budget scenarios
Moby can help explain a Compass budget scenario.
MMM supports Simulation for specific “what if” changes and Optimization for finding an allocation within a total budget and channel constraints. Results are projected model outputs across modeled categories or subcategories. Where supported, Moby may use Compass context to explain the scenario or prepare a recommendation for review.
Paid social+12%
Search-8%
Creator+6%
Current mix vs modeled scenario
Example constraintFixed budget
Projected impactModel estimate
Illustrative scenario — projected outcomes are model estimates, are most reliable within historical ranges, and are not guarantees or executed actions
09 / Guardrails and availability
More measurement context, not autonomous truth.
Moby 2 does not require Compass. Compass adds measurement context that may make certain cross-channel recommendations more informed and defensible. Available views, confidence scoring, Moby integration, and actions depend on account setup, connected data, product access, permissions, and rollout status. Material recommendations should still be reviewed.
What is supported
✓The Compass Command Center organizes channel performance, first-party attribution, MMM, incrementality, confidence, and signal disagreement.
✓Moby may use Compass context for analysis, explanation, recommendations, and supported workflows.
✓The Context Engine may add connected data, goals, preferences, workflow instructions, and prior context where available.
What this does not imply
−A silent average of platform reporting, attribution, MMM, and incrementality signals.
−A deterministic week-by-week revenue forecast or a user-level journey model from MMM.
−Automatic major budget changes or experiment launches without supported permissions and review.
01
Recommendations with broader measurement context.
Where available, Moby can consider Compass alongside connected performance data, goals, and saved business context.
02
Clearer investigation when signals disagree.
Compass surfaces confidence and disagreement; Moby may help explain what deserves monitoring or deeper review.
03
Reviewable recommendations before spend shifts.
Scenario results are projections. Business context and human review still determine whether a material action should happen.
10 / Put measurement context to work
Add Compass context to Moby.
When both products are available, Moby may use Compass measurement context to explain cross-channel performance, compare budget options, and prepare more informed recommendations for your team to review.