How Gina Tricot improved its ROI by +53% by leveraging Cassandra MMM SaaS

How Gina Tricot improved its ROI by +53% by leveraging Cassandra MMM SaaS

Marketing Mix Modeling and ROI Optimization Case Study



By Cristian Nozzi, CTO and co-founder of Cassandra

To gain a deeper practical understanding of the value that marketing mix modeling (MMM) SaaS solutions can drive for direct response-heavy advertisers, we at Cassandra collaborated closely with Meta Marketing Science team to capture the impact of Cassandra's MMM solution at Swedish fashion company Gina Tricot

The case study below will show you not only the significant improvement in return on ad investment (ROI) the brand saw, but also the benefits of MMM's forecasting capabilities when it comes to decision making based on predicted future outcomes. 

Huge thanks to everyone who helped with this effort, including Emil Garrote and Marcus Österdahl from the Gina Tricot marketing team and Alfonso Calatrava and Igor Skokan from the Marketing Science team at Meta.

MMM Context 

The availability of some of the data that helps to inform digital advertising, like cookies and user-level identifiers, has decreased as new privacy regulations and platform policies give people more options to limit how their data is shared with businesses. As a result, marketers are having  harder time assessing the true ROI of their media, as well as appropriately designating budgets to the right channels and planning media to maximize sales. 

This has been the backdrop of an upswing in MMM adoption by DR-heavy advertisers who need a measurement solution to inform their decision-making in an increasingly complex world. MMM is a statistical methodology that only uses aggregate spend-level data, meaning its usefulness is not affected by recent privacy-related changes. And with AI-supported automation and improved granularity, MMM can help marketers assess online and offline media ROI, detect opportunities for optimization, and improve media planning. 

At Cassandra, we collaborated with the Meta Marketing Science team and built an accessible Marketing Mix Modeling SaaS solution for non technical marketers leveraging  the research and development done with Project Robyn, Meta’s open source MMM code

Thanks to this, Cassandra is able to help all marketing teams improve their media mix ROI with data driven budget allocation prescriptions within few weeks of onboarding to our solution, allowing to effectively support fast-moving decision makers 

Gina Tricot: a company for women, by women

Gina Tricot, established in 1997 in Borås, Sweden, is a fashion brand committed to delivering the latest trends in women's fashion. 

With a strong emphasis on sustainable and ethical practices, Gina Tricot has become a go-to destination for conscious fashion consumers.

Today, the company operates approximately 150 stores in four European countries, as well as a growing e-commerce within Europe.


Up until the company adopted Cassandra's MMM SaaS solution, Gina Tricot relied heavily on Google Analytics and cookie-based last-click attribution for its advertising effectiveness measurement. 

However, with ongoing privacy-related changes such as the depreciation of cookies, as well as last-click attribution's lack of nuance, the team was struggling to understand the value of its media, making decisions about allocation challenging. 

Due to these prevailing uncertainties, the Gina Tricot team, including Emil Garrote, Head of Paid Media and Analytics, and Marcus österdahl, Digital Analytics Specialist, decided to work with Cassandra to implement its MMM software.


With the help of its machine learning-based MMM solution, Cassandra was able to provide the Gina Tricot team with a comprehensive analysis of their marketing data for the Netherlands market

The followings are the insights they unlocked:

  • Confidence interval of each Media ROI: In Marketing Mix Modeling, when we talk about the ROI (Return on Investment) for each media, we present a range of likely values. This method acknowledges that while our estimates are accurate, there's a small chance of variation due to external factors.From this data we unlocked revenue contributions for all their media, being able to quantify both the amount of incremental sales driven by each channel as well as their related ROI

ROI confidence interval MMM
  • Unlock historical contribution of Media, Organic and Seasonality factors: With more than 12 factors influencing sales in their media mix, Cassandra unlocked the contribution over time of each factor, diagnosing what caused historical changes in sales.
Contribution over time MMM case study

  • The Budget Allocated to Influencer Marketing exceeded its returns: Upon analyzing the diminishing return curves of Tiktok and Instagram influencer marketing, Gina Tricot saw that these approaches had a good ROI but saturation curves needed to be taken into account. . This information led the marketing team to optimize the  investments in Influencers in order to improve ROI. 

Instagram Influencer’s Diminishing Returns


 Achieving  +53% ROI MoM (Month over month) 

After diagnosing the past, the Gina Tricot team ran Cassandra’s budget allocator to plan its December 2023 marketing budget.

This feature takes into account: 

  • Diminishing returns, adstock 
  • Uncertainty, ROI confidence intervals 
  • Seasonality 
  • Historical spend variance 

To formulate the Media Plan that has the highest probability of generating a Lift in ROI for the brand. 

Cassandra’s budget allocator in its MMM solution enables marketers to receive calibrated media plans that will improve their likelihood of success.  

Because Gina Tricot's goal was to maximise sales by unit volume, Cassandra recommended scaling back the budget in generic search campaigns and Instagram and Tiktok influencer marketing and reallocating to direct response campaigns on Meta and Performance Max.

These optimal reallocations resulted in a lower overall budget by 26%.

MMM budget allocator validation

One month after reallocating the budget based on the initial MMM results, Cassandra and the Gina Tricot team refreshed the model with new data.

In this followup analysis, the incremental contribution measured by the Marketing Mix Model reported a +53% increase in ROI compared to the 52% that was previously forecasted by the budget allocator reconfirming the accuracy of the MMM analysis also to be able to predict the future outcomes.


Thanks to Cassandra's MMM solution, the Gina Tricot team was able to quickly build out their MMM capabilities without any internal data science resources -- and ultimately improve marketing ROI by 53% with optimal budget allocation. 

Going forward, the Gina Tricot team plans to continue running new MMM model refreshes on a frequent basis  , as well as to further validate key insights from the MMM analysis results through  incremental lift experiments

Foto del profilo di Marcus Österdahl

"In response to growing privacy restrictions, our company required a privacy-centric marketing measurement system that could be swiftly and easily implemented.
The implementation of Cassandra's MMM solution not only allowed us to gain a holistic view of how our media advertising drives business outcomes, but also provided impressive and precise results that helped us improve ROI in a consistent way across all our affected markets."

-Marcus Österdahl, Digital Analytics Specialist, Gina Tricot

Want to replicate these results with MMM?