Crafting an MMM Project Brief

Hello! Welcome to the final lesson in our module on Marketing Mix Modeling.

In our previous lessons, you've seen the strategic power of MMM. We've explored how it helps deconstruct performance through contribution analysis, how to use response curves to find saturation points, and how to leverage scenario planning to forecast the impact of budget changes. You now understand what you can get from a successful MMM project.

Today, we shift from interpreting the outputs to initiating the process. Our learning outcome is to develop a comprehensive brief for commissioning an MMM project from a vendor or internal team. A well-crafted brief is the single most important document for ensuring an MMM project succeeds. It aligns stakeholders, clarifies objectives, and ensures the final model delivers actionable business insights rather than just interesting data points. This is a critical leadership task that sets the foundation for the entire initiative.

1. The Anatomy of a Powerful MMM Brief

Think of an MMM brief or Request for Proposal (RFP) as the architectural blueprint for your measurement strategy. It tells your internal team or a potential vendor what you want to build, why you're building it, and how you'll define success. A weak brief leads to a model that doesn't answer your real questions. A strong brief leads to a strategic tool.

We can structure a comprehensive brief around four key areas: Strategic Foundations, Data Readiness, Vendor & Model Selection, and Actionability.

As the checklist highlights, everything starts with strategy. Let's break down how to translate these strategic foundations—and other key elements—into the sections of your brief.

2. Section by Section: Building Your Brief

A strong brief guides the vendor or your internal team through your thinking, ensuring they understand the business context as well as the technical requirements.

Section 1: Business Context & Objectives (The "Why")

This is the most important section. A technically perfect model that solves the wrong problem is useless. Your goal here is to anchor the project in clear business needs.

What to include:

• Primary Business Goals: Start high-level. Are you focused on revenue growth, increasing market share, improving profitability, or acquiring a new customer segment?
• Key Business Questions: This is where you get specific. The more precise your questions, the better. As a marketing leader with deep experience in Meta and Google Ads, you know that platform-reported ROAS can be misleading. MMM helps solve this. Good questions might include:
  • What is the true, incremental ROI of each of our major marketing channels (e.g., Paid Search, Paid Social, TV, SEO)?
  • What is the optimal total marketing budget to hit our annual revenue target of $X?
  • If we have to cut the budget by 20%, where should we reduce spend to minimize the impact on sales?
  • How long does it take for our brand-building activities (like TV) to impact sales (i.e., what is the adstock effect)?
• Current Measurement Gaps: Briefly explain why your current methods (e.g., last-click attribution) are no longer sufficient for strategic decision-making.

Section 2: Project Scope & Data Requirements (The "What")

Here, you define the boundaries of the model and list the data you can provide. Being upfront about data availability and quality is crucial.

What to include:

• Model Scope:
  • Timeframe: Specify the historical data period (e.g., "last 3 years of data from Jan 2021 - Dec 2023"). A 2-3 year lookback is typical.
  • Granularity: State the time granularity of your data (e.g., "all data will be provided at a weekly level").
  • Markets & Products: Which geographic markets and product lines should be included?
• Data Inventory: List all the data points you can provide. A thorough data discussion is essential.

Based on the video, your data inventory in the brief should cover:

1. Dependent Variable: Weekly sales revenue and/or units sold.
2. Marketing & Media Variables: Weekly spend and impression data for every channel in scope.
3. Promotions & Pricing: Data on discounts, sales promotions, etc.
4. External Factors: Mention known drivers like seasonality, holidays, competitor activities, and macroeconomic trends.
5. (Optional) Brand Tracking: If you have data on brand awareness or consideration, include it.

Don't worry if your data isn't perfect. Part of the vendor's job is to help with data cleaning and identify gaps. Your role is to facilitate access and provide the business context.

Section 3: Expected Deliverables & Usability (The "How")

This section defines what "done" looks like. What do you expect to receive, and how will you and your team interact with the results? Don't just ask for a model; ask for a decision-making tool.

What to include:

• Analytical Outputs: Based on our previous lessons, you should explicitly request:
  • Marketing Contribution: A breakdown of what drives sales (baseline vs. marketing channels vs. other factors).
  • Response Curves: For each major channel, showing diminishing returns.
  • mROI and ROAS: For each channel.
  • Adstock Effects: The carryover effect for each channel.
• Strategic Tools:
  • Budget Optimizer/Scenario Planner: The ability to simulate different budget levels and allocations to forecast outcomes (as we covered in the last lesson).
• Usability and Operations: How will you use the model on an ongoing basis?

The article brings up two vital points for your brief:

1. Model Cadence: Ask the vendor to define their process for keeping the model current. Be specific: "Please describe the frequency and methodology for model refreshes (incorporating new data) versus model rebuilds (re-estimating all parameters)."
2. Usability: How will your team access the insights? State your preference: "We require a self-service platform or dashboard that allows the marketing team to run budget scenarios without needing to file a ticket with an analyst."

Section 4: Vendor Evaluation Criteria (The "Who")

Finally, you need to tell potential partners how you will evaluate them. Being transparent about your criteria ensures you get comparable and relevant proposals.

Key criteria to list in your brief:

• Transparency and Validation: This is non-negotiable. As Recast highlights in their article, you must ask: "How will we verify and validate the model's results?" A good vendor should have a clear answer involving holdout data testing, comparing model predictions to actual results, and calibrating the model with incrementality experiments (like geo-lift tests).
• Methodology: Ask how their model handles key components like adstock and diminishing returns. Do they use modern Bayesian methods that provide probability distributions, or simpler regression models?
• Experience: Request industry-specific case studies and client references.
• Team & Support: Who will be on the project team? What level of training and support is included?
• Timeline and Cost: Request a detailed project plan with milestones and a clear pricing structure.

Conclusion

You are now equipped to create a comprehensive and strategic brief for an MMM project. Remember, this document is your primary tool for aligning internal teams and external partners around your business goals.

Key Takeaways:

• A great MMM brief is built on a foundation of clear business objectives and strategic questions (The "Why").
• It requires a detailed inventory of your scope and available data, being transparent about limitations (The "What").
• It must specify the desired deliverables, particularly strategic tools like scenario planners and the operational plan for usability (The "How").
• Finally, it should outline clear evaluation criteria for selecting a partner, with a strong emphasis on model validation and transparency (The "Who").

By focusing on these four areas, you transition from simply buying an analytics model to commissioning a strategic decision-making engine that can guide your company's growth.

Preview of the Next Lesson:
Commissioning an MMM is a huge step forward. But as we've hinted, MMM is not a silver bullet. It's one powerful tool in a larger measurement toolkit. In our next lesson, we will explore how to "Synthesize MMM findings with other measurement methods (e.g., incrementality) to create a holistic view." You'll learn how to use the strengths of different methodologies to build a more complete and robust understanding of your marketing performance.