Discover how Meridian, Google’s latest innovation in marketing mix models, is redefining the landscape of marketing analytics with a focus on privacy. 

POV By Victoria Stapleton, VP Digital Analytics, Acronym

Google’s Meridian Marketing Mix Model:  

Google’s announcement of Meridian, its latest foray into marketing mix models (MMMs), marks another milestone in its preparation for full retirement of third-party cookies on which marketing and web analytics tools have largely depended (learn more about dealing with that change in our prior article). Along with solutions like the Privacy Sandbox, Enhanced Conversions, more statistics-based reporting solutions in Google platform and GA4, Meridian is another important step to understanding marketing performance in a more privacy-focused way.  

Note: at the time of this article’s publishing, Meridian has limited availability and requires an approved application from Google.  

What is Meridian? 

Meridian is Google’s open-source MMM designed to answer critical marketing questions. It serves as a compass for marketers, guiding them through the complexities of advertising spend and its impact on sales and conversions. At its core, Meridian is a statistical tool that quantifies the effectiveness of each marketing channel, providing insights into how various elements of the marketing mix contribute to overall business goals. 

What are the key features of Meridian? 

Meridian stands out with its Bayesian hierarchical models, which offer a nuanced understanding of marketing effectiveness across different levels, such as geography or product categories. This approach allows for the incorporation of ROI priors, integrating existing knowledge into new analyses. 

Another key feature is its privacy-centric design. In a world increasingly conscious of data privacy, Meridian provides a framework that respects user privacy while still delivering actionable insights. It also supports scenario planning, enabling marketers to forecast the outcomes of various strategic decisions. 

How does Meridian compare to other self-service MMMs?  

When it comes to choosing the right MMM, marketers must weigh the pros and cons of each model against their specific needs:  

 1. Meridian: 

• • Complexity and Customization: While Meridian provides flexibility for customization, its complexity may be a limitation for analysts who are not well-versed in Bayesian modeling or hierarchical geo-level analysis. 
  • Data Requirements: Like any MMM, Meridian requires historical data on marketing spend, sales, and other relevant variables. Insufficient or poor-quality data can impact model performance. 
  • Resource Intensive: Training and maintaining a Meridian model can be resource-intensive due to the need for computational power and expertise. 
  • Privacy Concerns: Although Meridian emphasizes privacy, any data used for modeling should be handled carefully to avoid privacy breaches. 

2. Robyn: 

• • Black Box Nature: Robyn’s AI/ML-powered approach can be seen as a black box, making it challenging for analysts to fully understand the model’s inner workings. 
  • Dependency on Facebook’s Prophet: Robyn relies on Facebook’s Prophet library for time series decomposition. Analysts need to trust the accuracy and reliability of this external component. 
  • Limited Documentation: As an experimental package, Robyn’s documentation may be less comprehensive compared to more established models. 
  • Model Calibration: While Robyn aims to reduce bias, calibration still requires careful validation against ground-truth data. 

3. LightweightMMM: 

• • Simplicity vs. Complexity: While LightweightMMM’s simplicity is an advantage, it may lack some advanced features found in more complex MMMs. 
  • Limited Features: It focuses primarily on channel attribution and budget optimization. Analysts seeking more sophisticated features (e.g., seasonality, external factors) may find it lacking. 
  • Community Support: Being relatively new, LightweightMMM may have limited community support and fewer resources available for troubleshooting. 
  • Bayesian Approach: Bayesian models require prior knowledge or assumptions, which can be a limitation if accurate priors are not available. 

 Conclusion 

The introduction of Meridian by Google is a testament to the tech giant’s commitment to advancing marketing analytics. It offers a sophisticated, privacy-conscious tool that empowers marketers to make informed decisions.

However, the choice between Meridian, Robyn, LightweightMMM, and other open-source MMM tools depends on the specific needs of the marketing team, their expertise, and the complexity of the marketing challenges they face. Each model has its strengths and weaknesses, and the best choice will align with the organization’s strategic objectives and data capabilities.  

Regardless of which model is chosen, no open-source solution offers expert support, domain expertise, or robust user enablement. For marketing departments inexperienced with MMM or other advanced analyses, Acronym strongly recommends external support.  

Need help deploying your open-source MMM solution? Contact Us today!

By Victoria Stapleton, VP Digital Analytics, Acronym

In the digital marketing industry, there has been much lament about the impending demise of the cookie and how this will change the way marketers measure and optimize their campaigns. With digital ad spend projected to exceed $600 billion, the ongoing concern around cookies isn’t going anywhere soon as marketers brace for this shift and have concerns about how to adjust.

The Great Cookie Divide: First-Party vs Third-Party

It’s true that the longevity and, therefore, reliability of both first- and third-party cookies are changing. However, there is a distinct difference in the changes between third- and first-party cookies. Here are some essential differences to understand:

• Third-party cookies have been phased out by Google in the Chrome browser for 1% of users starting January 4, 2024, with the goal being a full phase-out by the end of this year.

However, Chrome is the only major browser left that hasn’t substantially phased out third-party cookies; so, for 34% of browsers worldwide the present-day scenario is currently ‘third-party cookie-less’.

• For both first- and third-party cookies, browsers that clear cookies automatically after a set number of days, such as Apple’s Safari browser, will influence long-range user tracking. No cookies are safe from this type of restriction.

The Winds of Change: Legislation, Litigation, and Consumer Pressure

These changes are primarily driven by increased legislation, numerous lawsuits, and consumer pressure along with some existing measurement challenges. Below are some of the most recent changes impacting the shifts in data analytics restrictions for marketers:

Legislation: Privacy laws around the globe that restrict the use of cookies without user consent, such as GDPR, CCPA, ePrivacy Directive, and a plethora of others. These laws require marketers to obtain explicit and informed consent from users before collecting and processing their personal data, which can reduce the reach and effectiveness of cookie-based advertising and analytics.

Industry: Browser policies that block or limit third-party cookies, such as Safari’s Intelligent Tracking Prevention (ITP), Firefox’s Enhanced Tracking Protection (ETP), Chrome’s Privacy Sandbox, etc. These policies aim to protect user privacy and security by preventing cross-site tracking and fingerprinting, which can undermine the accuracy and reliability of cookie-based measurement and targeting.

Consumer: User preferences and behaviors that favor more privacy and control over their online data, such as using ad blockers, deleting cookies, or opting out of tracking to various degrees. These preferences and behaviors reflect the growing awareness and concern of users about how their online data is collected and used, which can reduce the availability and quality of cookie-based data and insights.

Navigating Through Challenges: The Bright Side of the Post-Cookie Era

There’s plenty happening simultaneously, but the sky is certainly not falling. This transition provides marketers with an opportunity to become more sophisticated in the way they analyze and optimize marketing programs. Traditional last-touch attribution models are just not cutting it anymore.

Medium to large-size businesses with a multi-pronged strategy have a plethora of options to assess spend effectiveness outside of cookie-dependent platform data, including marketing mix modeling, conversion, brand lift studies, data clean rooms, multitouch attribution models, and many other methods.  Additionally, there are technologies to help target your most valuable audiences including contextual advertising, cohort-based advertising, and device graphs.

With these tools, marketers do not need the invasive, granular tracking that cookies have so notoriously enabled to make smart investment decisions. Embracing the increased restrictions on cookies as a clear signal from your customers for a less invasive digital experience shows them that you understand their concerns and want to earn their trust.

Strategies for Success: Best Practices in the Post-Cookie World

To successfully navigate the post-cookie era, marketers should consider the following best practices:

1. Utilize first-party data and consent management platforms to collect and store user information in a compliant, transparent manner.
2. Explore alternative methods and technologies, such as contextual advertising, cohort-based advertising, and device graphs, to effectively measure and target audiences.
3. Embrace a customer-centric and value-driven approach to marketing that focuses on building trust and loyalty with users, rather than relying on intrusive and irrelevant ads.
4. Don’t tackle it alone, seek expert help. Refine your strategy and update your implementation rapidly by getting assistance from a trusted digital marketing agency like Acronym, which specializes in digital intelligence. Contact Us today!