Macromill Group

Tokyo, Tokyo, Nov 22, 2024 - Macromill, Inc. (Headquarters: Minato-ku, Tokyo, Representative Executive Officer and CEO: Toru Sasaki; hereinafter, "Macromill"), has deployed "Robyn" ^*1, an open-source MMM package provided by Meta in its MMM (Marketing Mix Modeling) service. "Brand Dynamics Modeling" is specialized in verifying the return on investment (ROI) of branding initiatives. In addition, Macromill has begun utilizing MMM data export^*2 in "Business Portfolio", a business asset management tool on the Meta platform.

By integrating Macromill's proprietary algorithms with Robyn's analytical functions, Macromill is positioned to offer MMM services that incorporate branding indicators (awareness data such as customer recognition, recall, and brand-level purchase intention) in a more efficient and timely manner. Furthermore, by integrating Macromill's "Business Portfolio" with our clients Meta ad accounts through the "Partner" function, we can provide consistent support from the collection to the preparation of Meta ad placement data.

^*1   Robyn is an experimental, AI/ML-powered, open-sourced Marketing Mix Modeling (MMM) package from Meta Marketing Science. See more: https://facebookexperimental.github.io/Robyn/

^*2   The function generates MMM data from ad-serving performance data.

■ Overview of MMM service "Brand Dynamics Modeling"

Macromill's new service offers a comprehensive solution for data-driven branding insights. It provides a single point of access for acquiring data on branding indicators, visualizing the impact of branding measures, and optimizing investments. As branding indicators are evaluated as intermediate indicators, both short-term sales promotion and mid- to long-term branding measures can be assessed on an equal footing. This enables clients to visualize the financial impact of their medium- to long-term marketing initiatives, optimize their media strategies, and enhance their marketing ROI.

To access the related press release: https://www.macromill.com/press/release/20240930.html (Japanese)

■ Background

Although MMM is a widely recognized and effective method for assessing the impact of marketing initiatives, the substantial time investment required for model calibration and result visualization has been a significant obstacle to comprehensive analysis. Meta's open-source MMM "Robyn" development project is designed to make MMM analysis more accessible and efficient by using machine learning.

Moreover, the vision of project Robyn is to build a community of MMM methodologists, discussing innovation and contributing to open source code.

In our Macromill as well, as we work to support client companies in visualizing the impact of their medium- to long-term marketing investments, optimizing media strategies, and improving marketing ROI, we have been customizing the open-source code of "Robyn" to accelerate the delivery speed of our MMM services and enhance analytical accuracy. This research and development effort aims to bring greater efficiency and precision to our MMM offerings.

■ Effects of implementation of Robyn in MMM services

1. Using "Robyn" for MMM Analysis Including the Effects of Branding Initiatives

   Macromill has successfully integrated our proprietary MMM algorithm with "Robyn". This integration allows us to efficiently provide analysis results for key performance indicators while maintaining the ability to evaluate the impact of branding initiatives and their effect on sales on an equal footing.

2. Support data collection and preparation using "Business Portfolio"

   Macromill's "Business Portfolio" accounts can be linked to clients Meta ad accounts through the "Partner" feature. This collaboration will support the aggregation of Facebook, Instagram, and other ad placement data on the Meta platform using MMM data export functionality. This streamlines the data collection process, reducing the human hours required.

Yuya Tanaka, Marketing Science Partner, Facebook Japan G.K.
A member of Robyn development team and leading its deployment in Japan, commented

We are delighted that you are actively utilizing "Robyn", the open-source MMM that develops alongside the community, as well as MMM data export in "Business Portfolio", which has limited track records in Japan.

Data preparation in MMM is a crucial step, yet it often requires a substantial amount of work. We hope that this feature will help reduce the workload and enhance the analysis quality. We’ll keep working to strengthen the MMM community for the industry’s benefit, including Macromill.

Macromill will continue to actively pursue strategic partnerships with companies to further expand the data needed to implement MMM.

Ends,