MMM Kick-off
GLOBAL +rescue
×
Seeda

MMM Kick-off Meeting

03rd April 2026

1. Overview

1.1 About Global Rescue

Global Rescue is a travel risk and crisis response company operating across two core functions:

The business operates three separate business units/tactics:

1.2 Current Measurement Limitations

Attribution Blind Spot

Especially between Direct to consumer vs Partner

Platforms/Partnerships claim full conversion credit in silos, inflating ROAS and making it impossible to verify true incremental impact across channels.

Partner Attribution Gap

No visibility into how paid digital and partner channels interact or influence each other.

Branded Search Inflation

Branded search captures demand generated elsewhere, not incremental growth.

Project Primary Goals

1
Resolve double-counting in attribution
2
Understand true incremental ROAS across channels
3
Optimise budget allocation across digital, partner, and offline channels
4
Increase chosen KPIs

A Year with Seeda — What Success Looks Like

A clear, unbiased view of which channels truly drive incremental revenue

Actionable budget reallocation recommendations

A clear strategy to leverage the right channels together

Amplify marketing impact and minimise the cannibalisations

Deep Dive — Model's Foundation

MMM Can Help You Achieve Your Goals!

Let's Define Them

📈

What Should We Measure?

Measurement

⚙

What Levers Do We Have?

Channels & Investment

ℹ

How Can We Get Context?

Market & Competition

Measurement

What are you trying to achieve with us? What effect to monitor? Which market are we investigating?

Channels & Investment

Channels currently invested in? How have you used them for goals? How is marketing managed?

Market & Competition

Overall market landscape? What about competition?

4. Timeline & Milestones

Timeline & Milestone for the 1st Model

Phase 1: Pre-Kick-off
  • Prepare Q&A section
  • Q&A file
    • Phase 2: Kick-off (Week 1)
  • Goal setting & KPIs
  • Measurement framework
  • Data onboarding
  • Tracker & workflow walkthrough
    • Phase 3: Data Process (Week 2)
  • Set up ingestion pipelines
  • Run initial data checks
  • Flag gaps/issues
  • Confirm readiness criteria
    • Phase 4: Data Review (Week 3) ⭐
  • Present initial data & trends from past 2 years
  • Highlight gaps, risks, assumptions
  • Align on metrics readiness
    • Phase 5: Modelling (Weeks 4-5)
  • Build MMM model
  • Measure channel contribution & ROI
  • Test scenarios & assumptions
  • Prepare initial insights
    • Phase 6: Model Output (Week 6)
  • Present model outputs & insights
  • Provide budget recommendations
  • Run scenario simulations
    • Phase 7: Finalising (Week 7)
  • Apply adjustments
  • Finalise model & budgets
  • Configure dashboards/platform views
    • Phase 8: Final Result (Week 8) ⭐
  • Present final model results & insights
  • Highlight improvement opportunities
  • Share possible next-step options

    • 5. Governance & Way of Working

    5.1 General

    AreaDetailRemark
    Meeting cadence & check-insBi-weekly, Quarterly
    Offline data ingestion meetingFrequently at the beginning, By demand
    Communication channelsEmails - for important info
    Slack - day to day    		Open to using GR's preferred messaging tool
    Document & asset storageSeeda's shared folder
    External Partners / Confidentiality & Data UsageData shared through Google and GCP
    Project Status ReportingWeekly update by Seeda team
    Acceptance milestonesGlobal Rescue: Data sign-off (week 3)
    Seeda: Model deliver (week 8)
    ▸

    5.2 Response Time SLA

    Request TypeSeedaGlobal Rescue
    General communicationWithin 1 business dayWithin 1 business day
    Data requestsSpecify requirements within 2 business daysDeliver data within 3 business days
    Feedback on deliverablesIncorporate within 3 business daysProvide within 5 business days
    Approval & sign-off—Within 3 business days
    Urgent / escalationWithin 4 business hoursWithin 4 business hours
    ▸

    5.3 Change Request Process

    1. Any change to scope, timeline, or data requirements must be submitted in writing.
    2. Seeda will assess impact on timeline and budget within 5 business days.
    3. Changes are only confirmed once both parties agree in writing.

    Verbal agreements do not constitute an approved change.

    ▸

    5.4 Escalation Process

    LevelTriggerOwnerResolution Target
    Level 1Minor delays or unclear requirementsProject leads/AM from both sidesWithin 2 business days
    Level 2Missed deadlines or scope disputesProject leads/AM from both sidesWithin 5 business days
    Level 3Unresolved conflicts or contract issuesExecutiveTBD

    6. Roles & Ownership

    Clear ownership across every stage — from kick-off to delivery.

    Area Seeda Owner Global Rescue Owner
    Project Lead Oversee project delivery, timeline & quality Elisa Tardy Own internal coordination, stakeholder alignment Michael Holms
    Data Provider Define data requirements & format Kai To Provide media spend, sales, channel data Wil Klass, Samantha Hosking, Knox Taylor
    Data Validation Review and flag data gaps Kai To Clarify and fill gaps Wil Klass, Samantha Hosking, Knox Taylor
    Stakeholder/Approver Present findings, get sign-off Sophia Cong Review, validate, approve deliverables Michael Holms
    Area Seeda Owner Global Rescue Owner
    Stakeholder Alignment Facilitate kick-off, check-in, result meetings Sophia Cong Facilitate meeting with decision-makers Wil Klass
    Methodology Design and build MMM model Elisa Tardy, Vincent Larrieu Review and sign off on approach Michael Holms
    Insights & Recommendations Deliver findings and budget recommendations Elisa Tardy Validate against business context Michael Holms
    Feedback & Revision Incorporate feedback, refine outputs Elisa Tardy, Vincent Larrieu Timely feedback at each milestone Michael Holms, Wil Klass
    External Partners/Vendors Coordinate with third-party tools Sophia Cong, Kai To Flag relevant external partners Wil Klass, Samantha Hosking, Knox Taylor

    7. Data Requirements

    Data Sources Needed

    Online

    Main: Google, Microsoft, Meta, StackAdapt, Mountain

    Other: TikTok, Spotify, X

    Offline

    • Partner list/program
    • Enterprise list/contract
    • Trade shows/events
    • Influencer marketing

    Transactions / Sales

    Transaction-level sales data across all business units.

    Format & Granularity

    Online

    Supermetrics (auto-pull via API)

    Offline

    • Excel, CSV, PDF
    • Daily or weekly time frame (by channel/campaign)
    • Region breakdown (optional)

    Transactions / Sales

    Custom API integration

    8. Next Steps

    Immediate Actions

    1. Separate meeting with Data Owner(s)
    2. Offline data to be shared
    3. Connect Supermetrics API

    Week 01

    • Data Onboarding
    • Measurement framework
    • Tracker & Workflow walkthrough

    Appendix

    ▸

    The Measurement Gap MMM Was Built To Solve

    The Attribution System Is Fundamentally Broken

    Platforms Claim Everything

    Each platform takes credit for conversions it touched, creating massive overlap and inflated totals.

    Privacy Killed Online Tracking

    iOS 14+, cookie deprecation, and consent regulations have eroded the accuracy of digital attribution.

    Offline Channels Are Invisible

    Partner programs, events, and enterprise sales have zero representation in digital attribution models.

    Too Much Data, No Clarity

    Dashboards overflow with metrics, yet the fundamental question remains unanswered: what actually works?

    ▸

    Methodology

    ▸

    Methodology 1: How Seeda Builds And Validates Your Marketing Mix Model

    Bayesian Engine

    Probabilistic framework that incorporates prior knowledge and updates with your data.

    →

    Data Transformations

    Adstock, saturation, and lag effects applied to reflect real-world media behaviour.

    →

    Model Validation

    In-sample fit, out-of-sample testing, and stability checks ensure reliability.

    ▸

    Methodology 2: How MMM Works — Bayesian Statistical Analysis

    The Rain Analogy

    • Prior: You believe it rains ~30% of days in spring.
    • Data: You check the sky — dark clouds, high humidity.
    • Simulation: You mentally simulate thousands of similar days.
    • Likelihood: Updated belief: ~70% chance of rain today.

    Applied To Your MMM

    • Prior: Industry benchmarks suggest Facebook ROAS is ~3x.
    • Data: Your spend, impressions, conversions over 2 years.
    • Simulation: Model runs thousands of scenarios to test relationships.
    • Likelihood: Updated: Facebook ROAS is 2.4x for your business specifically.
    ▸

    Methodology 3: 3 Checks Ensure The Model Is Accurate And Stable

    In-Sample Fit

    How well the model explains your historical data.

    R2 MAE MAPE

    Out-of-Sample Fit

    How well the model predicts unseen data (holdout period).

    R2 MAE MAPE

    Stability Checks

    Coefficients remain consistent across different time windows and data subsets.

    Coefficient Stability Convergence
    ▸

    Methodology 4: Raw Spend Is Transformed To Reflect Real-World Behaviour

    Lag

    Accounts for the delay between when a customer sees an ad and when they convert. Different channels have different lag periods.

    Adstock

    Captures the carry-over effect of advertising. A TV ad seen today still influences behaviour next week, but with decaying impact.

    Saturation

    Models diminishing returns. Doubling your spend on a channel does not double the output — every channel has a saturation curve.

    ▸

    MMM Refresher

    What MMM Tells You

    Incrementality

    What each channel truly adds above and beyond what would have happened anyway.

    Elasticity

    How sensitive your KPI is to changes in spend for each channel — where does an extra dollar go furthest?

    Synergies

    How channels amplify each other. Running paid social alongside search may produce more than either alone.

    All Effects

    The complete picture: base demand, seasonality, external factors, and every marketing channel's contribution.

    Budget Example

    Facebook
    $100
    +
    Google
    $300
    +
    OOH
    $600
    =
    Total
    $1,000

    The Output

    Channel Contribution

    How much revenue or conversions each channel is responsible for driving.

    Return On Investment

    The true incremental ROI for each dollar spent, per channel.

    Elasticity

    Where to increase or decrease spend for optimal returns.