Simple Guide on MMM Pricing: Balancing Complexity, Resources, and Value

In the world of Marketing Mix Modeling (MMM), effective pricing hinges on a project’s complexity and the intensity of resources required. The industry generally prices MMM projects between $125,000 and $500,000, depending on various factors. However, without a clear pricing framework, misunderstandings can arise — clients may feel they are overpaying, while providers might believe they are undercompensated. To avoid confusion and disagreements, it’s essential to establish a standardized, transparent pricing system that helps both sides understand the basis of the costs. In this guide, I will outline a structured approach to pricing, while acknowledging that these figures represent a range rather than exact estimates.

Disclaimer: The numbers provided here are estimates and for illustrative purposes only.

Background and Pricing Context

Marketing Mix Modeling (MMM) projects are comprehensive, data-intensive endeavors designed to deliver strategic insights into media investments, marketing efficiency, and budget optimization. These projects often involve a diverse array of data, such as sales, media spend, market conditions, and external factors like economic indicators, competitor data, and seasonality. The complexity and effort required to build each MMM can vary widely, making it essential to focus on both technical sophistication and resource needs when determining pricing.

To achieve this, we propose a pricing framework based on two key components:

1. Model Complexity

2. Resource Intensity

Each component is evaluated on a scale from 1 to 10, with 10 indicating the highest level of complexity or resource demand. The average of these scores forms the Composite Score, which directly influences the final price of the MMM.

1. Model Complexity The complexity of an MMM is determined by the techniques and features integrated into the model. This includes the number and variety of external factors, media transformations (e.g., decay effects, saturation), synergy analysis, seasonality adjustments, and halo impact assessments. The complexity matrix, as shown in the referenced figure, guides the classification of complexity into tiers, with higher tiers encompassing all features of the lower tiers along with more advanced capabilities:

• 1–3 Complexity Score: Basic regression models with simple seasonality adjustments and limited external factors that are easily accessible. No media variable transformations (e.g., lag, decay, saturation) are applied, and response curves are not provided.
• 4–6 Complexity Score: Incorporates a broader set of external factors (such as scraping reviews and obtaining consumer sentiments toward brands and products) and includes full media variable transformations. Response curves are provided.
• 7–8 Complexity Score: Integrates an extensive list of external factors (e.g., reviews and consumer sentiments from platforms like app stores, YouTube, Instagram, Amazon, and TripAdvisor). Also includes media synergies and halo effects.
• 9–10 Complexity Score: Calculates media synergies and halo effects, and measures the seasonality of media cost-efficiency by channel where applicable.

2. Resource Intensity

Resource intensity measures the number of customized sub-models needed to accommodate multiple KPIs and product segments. This component accounts for factors such as:

• Number of KPIs: For instance, separate sub-models may be required to measure sales, customer acquisition, and retention rates.
• Product Segments: Different product lines or segments often require distinct models, increasing the number of sub-models needed.
• Data Granularity: Collecting data at more frequent intervals (e.g., weekly, monthly, or daily) increases the time and resources needed for data preparation.

Resource Intensity Scores (Examples):

• 3–4 Resource Score: 1 KPI, a single national model, and no product segments.
• 5–6 Resource Score: 1 KPI and up to 2 product segments.
• 7–8 Resource Score: 1 KPI and up to 4 product segments, or up to 2 KPIs with up to 2 product segments.
• 9 Resource Score: Up to 2 KPIs and 3 product segments, or 1 KPI with up to 6 product segments.
• 10 Resource Score: Up to 3 KPIs and 3 product segments, or 1 KPI with up to 9 product segments.

Note: This refers to customized sub-models tailored for each KPI and segment, not hierarchical models. Customized models allow for the selection of relevant external factors and specific decay rates and saturation curves for each media channel within each sub-model. For instance, consider Adidas with two major products: sneakers and clothing. If a media channel like social media is more focused on sneakers, then the relevant competitors, reviews, media transformations, and Bayesian priors would differ for each segment. In such cases, customized sub-models provide a more precise solution.

Weighting Complexity and Resource Intensity

Both complexity and resource intensity are crucial in determining the price of an MMM project, and deciding on their relative weight is key. A 50/50 split offers a balanced approach, giving equal emphasis to model sophistication and project effort. However, different business contexts may call for alternative weightings. For example, a highly sophisticated model with fewer product segments might justify placing more weight on complexity.

Reference Pricing Ranges

To provide clarity and align pricing with industry standards, we suggest the following price ranges based on the Composite Score:

Example: Pricing an MMM for Client A Suppose Client A’s MMM project involves three product segments and incorporates advanced transformations such as media decay and synergy effects. The model includes a comprehensive set of external factors, such as GDP and consumer sentiment, and focuses on one KPI for sales. Given these requirements, the project’s complexity is rated at 8, while the resource intensity is rated at 7 due to the number of segments and KPIs.

Composite Score Calculation:

• Complexity Score: 8
• Resource Intensity Score: 7
• Composite Score: (8 + 7) / 2 = 7.5

Based on the reference pricing table, the estimated price for this MMM would range between $170,000 and $250,000.

Cost of Model Refreshes: Full vs. Partial

Once an MMM is built and deployed, the business environment continues to evolve, as do marketing strategies, competitor dynamics, and consumer behaviors. For this reason, model refreshes are essential to keep the MMM aligned with current market conditions and ensure it remains an accurate decision-making tool.

In our proposed pricing framework, we recommend two types of model refreshes: Full Model Refreshes and Partial Model Refreshes. Each serves a different purpose based on the level of recalibration needed. Below is a detailed explanation of both options and their associated costs.

1. Full Model Refresh: Cost = 85%-100% of the Original Cost

A Full Model Refresh essentially re-establishes the MMM as if it were being built from scratch. This approach ensures that the model captures significant changes in the market, external conditions, and media dynamics.

What does a Full Model Refresh involve?

• Reassessment of External Factors: Re-examining current market conditions to identify new or changing external factors.
• Re-estimation of Coefficients: Running the model again to recalibrate the coefficients for all included variables.
• Updated Media Transformation: Adjusting decay rates, carryover effects, and saturation points to reflect recent media behavior.
• Updated Response Curves: Generating new response curves for each media channel.
• Recalculated Halo Effects and Synergies: Reassessing cross-channel impacts, halo effects, and media synergies.

When to choose a Full Model Refresh?

• Significant Market Changes: If there have been substantial changes in market dynamics or consumer behavior.
• Shift in Marketing Strategy: When media strategies or investments have changed significantly.
• High Deviation in Partial Refresh Analysis: If a partial refresh reveals high deviations between actual and predicted outcomes.

2. Partial Model Refresh: Cost = 50% of the Original Cost

A Partial Model Refresh is a quicker, less comprehensive update that applies the existing model parameters to new data.

What does a Partial Model Refresh involve?

• Data Transformation Using Previous Parameters: Feeding new data into the model using existing decay rates, response curves, and coefficients.
• Return Measurement Based on Previous Coefficients: Assessing the return on new media investments using previously determined parameters.
• Comparison of Actual vs. Predicted Outcomes: Checking if the deviation between actual and predicted results is within an acceptable range.

When to choose a Partial Model Refresh?

• Regular Monitoring and Evaluation: For routine monitoring and periodic checks.
• Low Market Volatility: If there haven’t been significant changes in external conditions or media strategy.
• Budget Constraints: When cost-efficiency is a priority.

Key Decision-Making Criteria: If the deviation between actual and predicted performance is within an acceptable range (e.g., less than 5%), the partial refresh is considered successful. If there’s a larger deviation, a Full Model Refresh is recommended.

Final Takeaway

This standardized pricing framework based on model complexity and resource intensity provides an objective way to price MMM projects. It also allows clients to understand why a highly complex model with numerous product segments will cost more. This approach helps stakeholders make informed decisions based on their budget, needs, and desired level of insights. Additionally, incorporating the option of full or partial model refreshes ensures that MMMs remain relevant and accurate, adapting to changing market conditions efficiently.