## An introduction to Meridian

Meridian is an open-source MMM built by Google that provides innovative
solutions to key measurement challenges. Think of it as a flexible and powerful
statistical framework. Because Meridian is open-source, its underlying code is
fully visible and available for you to modify. Meridian is free for anyone to
use, giving you the ability to control your model, your data, and your results.

This framework is purpose-built to help you answer three core business
questions:

- What is the historical **ROI** and **contribution** from each of our **marketing channels**?
- What is the **response curve** for each channel? In other words, how does the marketing impact vary with spend?
- Based on these results, how should we allocate our future budget to maximize our business **outcome**?

## Core benefits of Meridian

Meridian aims to empower the industry to build best-in-class
MMMs and drive better business outcomes. It achieves this through several key
features that lead to faster, more granular, and actionable insights.

- **Designed for causal inference:** Meridian is designed to estimate the true causal impact of your marketing. The causal inference framework and assumptions inform the model's estimate of **incremental outcome** for each **treatment variable**.
- **Customized by your expertise:** You can inject your existing business knowledge into the model through **priors**. A prior gives the model a starting point based on your information. This process helps tailor the model to your unique business reality and allows the model to produce insights that reflect both your domain expertise (via a prior) and data.
- **Enriched with control variables:** You have the option to include various **control variables** . For example, including Google Query Volume data as a **confounding variable** helps the model better account for organic brand interest when estimating the causal effect of your advertising.
- **Geo-level modeling:** Meridian supports the use of geo-level data, which contains more statistical information. This improves media effect estimation and also provides more granular geo-level insights.
- **Reach and Frequency methodology:** Meridian provides the option to use reach and frequency data as model inputs to provide additional insights. Reach is the number of unique viewers within each time period, and frequency is the corresponding average number of impressions per viewer. This provides more granular insights into how a channel's impact varies as a function of both reach and frequency. For more information, see [Bayesian Hierarchical Media Mix Model Incorporating Reach and Frequency
  Data](https://research.google/pubs/bayesian-hierarchical-media-mix-model-incorporating-reach-and-frequency-data/).
- **Designed for action:** Meridian contains **optimization \& scenario planning abilities** which are built off of its advanced modeling foundation. Meridian is designed to drive actionable business outcomes through its outputs around key business metrics, such as ROI estimates and credible intervals, response curves, and budget optimization recommendations.

## Overview of the Meridian journey

Meridian provides a structured, end-to-end workflow. The user journey includes
the following main steps:

1. [**Pre-modeling**](https://developers.google.com/meridian/docs/pre-modeling/intro): This centers on
   collecting, organizing, and cleaning your
   data. You provide the model with key business data, including your target
   KPI (often sales data), media data for each marketing channel, and any
   relevant control variables that might also influence both your marketing
   and KPI.

2. [**Modeling**](https://developers.google.com/meridian/docs/advanced-modeling/intro): Meridian's core
   Bayesian model is applied to your data.
   The model builds in assumptions about complex, real-world
   effects like the **lagged effect** of advertising and
   **saturation** . The model's primary output is a **posterior distribution**
   of model parameters, which feeds into the post-modeling analysis.

3. [**Post-modeling**](https://developers.google.com/meridian/docs/post-modeling/intro): This stage involves
   assessing your model results and
   reviewing actionable insights. Meridian provides a standard set of
   visualizations that help you understand the model fit, contribution of each
   channel, ROI, and more. To help turn insights into action, Meridian provides
   a budget optimization report that recommends how to allocate your budget to
   get the most out of your marketing investment.