Harnessing the Power of Multi-Touch Attribution and Marketing Mix Modeling

We are constantly searching for new ways to optimize our marketing strategies and maximize return on investment. One key area of focus that still evades many companies is understanding the effectiveness of different marketing channels and touchpoints in driving customer conversions. Enter Multi-Touch Attribution (MTA) and Marketing Mix Modeling (MMM), two powerful analytics techniques that help us uncover the true impact of our marketing efforts.

However, navigating the world of MTA and MMM can be daunting for professionals without a data science background. In this article, we will break down these complex methodologies in an easily digestible way for business professionals, regardless of their technical expertise or functional role.

Integrating concepts like market mix modeling and multi-touch attribution model into our marketing data analysis can enhance our understanding and provide significant insights into the customer journey and marketing budget allocations. These techniques offer a refined view of marketing channels, assisting teams in pinpointing where investments are most effective and where adjustments might be needed.

We'll explore the pros and cons of MTA and MMM, discuss when it makes sense to leverage each method, and examine how recent advancements in marketing analytics are democratizing access to these powerful tools.

My guest contributor is Gabriele Franco, CEO and co-founder of Cassandra. This recent marketing analytics startup has made headlines with its affordable and easy-to-use MMM platform. Join Gabriele and me as we unravel the mystery of MTA and MMM, empowering you to make more informed decisions and unlock the full potential of your marketing efforts.

What is Marketing Attribution (MTA vs. MMM)?

Marketing attribution is a crucial aspect of marketing strategy, as it helps us understand the effectiveness of our marketing efforts and allocate resources accordingly. We live in a data-driven world, and marketers must be able to quantify the impact of their campaigns and identify which channels or tactics deliver the best results. This understanding allows us to optimize our strategies, make informed decisions, and drive better return on investment (ROI).

Employing marketing attribution tools such as marketing analytics and data science helps leverage the full potential of marketing efforts. These tools aid in efficiently distributing marketing resources across various touchpoints, ensuring each marketing dollar is spent optimally to enhance customer segmentation and journey understanding.

There are two primary techniques used to measure marketing attribution:

1.       Multi-Touch Attribution (MTA)

2.       Mix Modeling (MMM)

While both approaches aim to quantify the impact of marketing activities, they do so with drastically different methodologies and levels of granularity.

MTA focuses on a time-stamped customer journey by attributing fractional value to each touchpoint, while MMM uses aggregated historical data to measure the impact of marketing activities on sales. Each method has unique advantages and challenges, which we will explore in the next couple of sections.

Multi-Touch Attribution

MTA is a marketing attribution technique that assigns credit to different marketing touchpoints in the customer journey, from the first interaction to the final conversion. MTA aims to provide a more detailed understanding of the customer journey, capturing the effectiveness of individual touchpoints across various channels and informing marketing professionals on which channels or tactics drive the best results.

Integrating machine learning algorithms into our multi-touch attribution model enables us to tap into advanced marketing analytics, providing a deeper dive into each touchpoint's effectiveness. This approach allows marketing teams to rapidly adapt strategies based on real-time data, ensuring optimal performance of marketing efforts and maximizing ROI

Granular insights into the customer journey

One of the critical advantages of MTA is its ability to provide granular insights into the customer journey. MTA is developed with user-level data to capture the effectiveness of individual touchpoints, such as email opens, paid search views, ad clicks, Youtube views, or Direct Mail exposure across various channels. This level of detail allows marketers to understand which specific interactions contribute to the final conversion and helps identify opportunities to optimize campaigns or target particular customer segments more effectively.

Real-time optimization

It offers near real-time data, which enables marketers to react faster and make campaign adjustments. With MTA, marketers can identify underperforming touchpoints and shift resources to higher-performing channels, tactics, or audience segments. This real-time feedback loop allows marketing leaders to be more agile in decision-making and improve overall marketing performance.

Enhanced personalization through user-level data

MTA's user-level data allows for better targeting and personalization of marketing efforts. By tracking individual customer interactions and understanding which touchpoints resonate with specific audience segments, marketers can tailor their messaging and offer to align with customer preferences and behavior. This level of personalization can improve engagement, conversion rates, and customer loyalty.

Marketing Mix Modeling

In contrast to MTA, Marketing Mix Modeling (MMM) is an attribution technique that uses aggregated historical data and statistical analysis to understand the impact of various marketing activities on sales. MMM is designed to help marketers predict how changes in marketing investments across channels will affect sales outcomes or how saturated a channel is. MMM helps marketers make informed decisions about their overall marketing strategy and budget allocation by providing a broader, more strategic view of marketing performance.

The strategic application of marketing mix modeling (MMM) combines historical data with data scientists' insights to craft a comprehensive view of the marketing landscape. This holistic approach is essential for achieving a balanced marketing strategy that addresses pain points in customer experiences and market basket analysis, guiding marketers toward more informed budgetary decisions.

It can often be difficult or cost-prohibitive for most companies to piece together omnichannel touchpoints at the customer level effectively. The lack of user-level impression data for traditional channels like TV and Radio further complicates it. Commercial ID graphing solutions that provide omnichannel identity resolution are a sound option, but they can easily reach $250K+ per year.

Google (through GA 360) and Facebook have digital attribution solutions, but they are typically biased and don't include impression data from others' digital platforms.

This is why MMM becomes so appealing: it doesn't care about user-level data. It relates weekly, aggregated changes in marketing channel level spend and impressions to changes in your sales. By optimizing a few things (such as decay and saturation rates), MMM does a decent job of articulating channel effectiveness and illustrating budget allocation opportunities. 

A holistic view of marketing effectiveness

One of the main advantages of MMM is its ability to provide a comprehensive understanding of marketing effectiveness. MMM considers both online and offline channels (along with earned, paid, and sometimes owned media) and external factors like seasonality, promotions, and economic indicators. This holistic approach allows marketers to understand the broader context of their marketing activities (typically over 2-3 years) and make more informed decisions about resource allocation and campaign planning.

Long-term insights into marketing impact

MMM is well-suited for capturing the long-term impact of marketing campaigns and brand-building efforts. While MTA focuses on individual touchpoints and short-term interactions, MMM looks at the bigger picture and considers the cumulative effects of marketing activities over time. This long-term perspective can help marketers understand the actual value of their marketing investments and inform strategic decisions about campaign duration, budget allocation, and channel mix.

Less prone to biases and data collection challenges

Unlike MTA, MMM is less affected by biases and issues related to cookie tracking, ad blockers, and other data collection challenges. Since MMM uses aggregated historical data rather than user-level data, it is less susceptible to the inaccuracies and gaps in data that can result from these challenges. It makes MMM a more reliable tool for understanding marketing activities' overall impact and informing long-term marketing strategy.

When to use MTA or MMM?

The choice between MTA and MMM largely depends on a business's specific needs and goals. MTA may be the more suitable option if a company primarily focuses on digital channels and requires real-time optimization. On the other hand, if a business wants a broader, more strategic view of its marketing performance across various channels, including offline and PR activities, MMM would be a better choice.

Incorporating marketing attribution, including MTA and MMM, into your overall strategy ensures a comprehensive analysis of marketing efforts and customer data. This dual approach allows businesses to tailor their strategies to meet specific marketing channels and customer segmentation, enhancing the overall effectiveness of the marketing spend.

Combining MTA and MMM for comprehensive insights

It would be best to combine MTA and MMM to gain the most comprehensive insights into your marketing effectiveness. By leveraging the granular, real-time insights provided by MTA and the holistic, long-term perspective offered by MMM, you can develop a complete understanding of your marketing performance. This combined approach allows for more accurate decision-making, improved campaign planning, and effective resource allocation.

The importance of data quality and accuracy

Regardless of whether you choose MTA, MMM, or a combination of the two, the quality and accuracy of your data play a crucial role in the success of your marketing attribution efforts. Ensuring that your data is clean, consistent, and up-to-date is essential for obtaining meaningful and actionable insights from your attribution models.

Regular data audits, proper data governance, and the implementation of data quality checks can help maintain your data's integrity and improve the overall effectiveness of your marketing attribution strategy.

As privacy regulations and tracking technologies continue to evolve, businesses must adapt their marketing attribution strategies accordingly. Staying informed about changes in privacy laws, such as GDPR and CCPA (and now the upcoming CPA), as well as advancements in tracking technologies, is essential for maintaining compliance and ensuring the continued effectiveness of your attribution models. By proactively addressing these challenges and researching new, innovative solutions, you can ensure that your marketing attribution efforts remain relevant and impactful in an ever-changing landscape.

The democratization of analytics opens doors (Robyn and Cassandra)

Historically, you could only get decent Marketing Effectiveness & Optimization studies done if you were ready to wait 12-16 weeks and shell out at least $150K. The advent of open-source software, the democratization of advanced analytics techniques, and data science research investments by tech giants have made sophisticated analytics solutions accessible to the masses.

While most small to mid-size companies (including many advertising agencies) still need more human or tech capital to do things like Marketing Mix Modeling, the cost and effort to perform actionable MMM in-house have decreased significantly.

Meta's Robyn project is a solid open-source toolkit I encourage everyone to check out. Robyn is a semi-automated MMM package that can handle model building, including hyperparameter tuning, descriptive analytics (e.g., the share of ad spend vs. impact), and marketing spend optimization. It is currently available for R, and the team is working on a Python wrapper. I've tried it and have been delighted with the results.

As the world of marketing analytics continues to evolve, innovative solutions like Cassandra are emerging to democratize access to MMM further. Cassandra is a cutting-edge platform designed to provide businesses of all sizes with the tools they need to effectively analyze and optimize their marketing efforts – at a fraction of the cost that most marketing consultancies charge.

To learn more about Cassandra's unique features and benefits, we are excited to provide an introduction by their co-founder and CEO, who will share his insights on how this powerful platform is transforming the marketing analytics landscape.

An Introduction to Cassandra: a Game-Changer for Marketing Analytics

Why does Cassandra exist?

Before starting Cassandra, I worked at my marketing agency, Hybrida.io, where we managed over $12 million in advertising investments for our clients. I realized that we needed a scientific way of allocating our clients' marketing investments to get the highest ROI out of them. We started testing MMM, and we realized that most of the companies that did not have a data science team could not access scientific methodology because they were costly ($100K+ per project) and extremely slow. Each project needs six months to obtain results.

This realization led us to develop Cassandra, an automated marketing analytics platform that provides fast, affordable, and actionable MMM insights.

How to Run Marketing Mix Modeling Service (Old vs. New Way)

The Old Way:

1. Data Collection: Gather historical marketing and sales data from clients, including advertising spend, promotions, pricing, distribution, and more.

2. Data Preparation: Clean and organize data accurately and consistently, addressing outliers, missing values, and format inconsistencies.

3. Model Selection: Choose a suitable statistical model, such as linear regression, time-series analysis, or machine-learning algorithms.

4. Model Calibration: Train the model with prepared data, adjusting parameters for minimal errors and optimal predictive accuracy.

5. Model Validation: Compare model predictions to actual sales data, ensuring an accurate representation of marketing channels' impact on sales.

6. Insights Generation: Analyze model results to identify insights and recommendations, such as effective channels, budget reallocation, or strategy adjustments.

7. Reporting: Create an easily digestible, comprehensive report with visual aids to present findings and recommendations from the MMM analysis.

8. Ongoing Support: Assist clients in implementing MMM recommendations, offering periodic check-ins, performance reviews, and model updates as new data arises.

The New Way:

With Cassandra, you can automate all these parts for a highly affordable price and start running your first MMM in 1 hour.

Our goal was to create a product that even non-technical marketers could use, making it simple and accessible for businesses of all sizes. With Cassandra, users can create their marketing mix, simulate marketing budget allocation decisions, and analyze various scenarios in just four easy steps:

Step 1: Connect your data and respond to the form

Step 2: Train your MMM in the cloud

Step 3: Unlock your MMM insights

Step 4: Simulate spend scenarios with your Marketing Budget allocator

An Opportunity for Brands: Cura of Sweden Case Study

Thanks to Cassandra, Brands can now optimize their marketing mix by measuring the real incremental impact of each marketing activity. They can allocate their marketing budget across various media channels, identify optimal investment levels for each channel, and determine whether they are overspending or underspending in specific areas. On average, our clients experience a 30%+ lift in ROI, helping them achieve better results and grow their businesses.

One of our successful collaborations has been with Cura of Sweden, a Swedish company that develops high-quality and innovative sleep-enhancing products for better sleep health. Cura Sweden faced challenges in internationalizing their brand across multiple countries while maintaining marginal costs in line with their financial projections. They found that Multi-Touch Attribution (MTA) using Google Analytics was insufficient due to the long time to purchase, which made it impossible to track performances accurately and understand the contribution of each investment on their sales.

To address these challenges, the company used Cassandra to analyze its media mix and develop a customized media plan to maximize its sales. Using the budget allocator, they could simulate investment scenarios and improve the effectiveness of their marketing channels.

Our collaboration with Cura of Sweden significantly improved their advertising budget, total paid orders, and cost per order. Specifically, they experienced an:

• 86% increase in orders from paid media

• 16% decrease in cost per conversion

• 52% increase in marketing budget invested

Cassandra identified which marketing channels were providing the best ROI and optimized their media mix accordingly, resulting in a substantial improvement in their sales and advertising budget allocation.

Incremental revenue-generation opportunity for Agencies

Marketing Mix Modeling is not only an opportunity for brands to increase their ROI but also for Agencies. In this recession climate, understanding the utility of every dollar invested in marketing is essential. Marketing Mix Modeling allows agencies to fulfill this need to their clients and can be repaid a lot.

The reasons why MMM represents an opportunity for agencies are:

1. Competitive Edge: Stand out from competitors by offering MMM services and positioning your agency as a data-driven, results-focused partner for businesses.

2. Enhanced Client Retention: MMM provides insights into clients' marketing performance, helping them optimize budgets and improve ROI, fostering long-term relationships.

3. Additional Revenue: Incorporate MMM into your portfolio to access a new revenue source, promoting overall profitability and growth.

4. Upselling Opportunities: Offer MMM as an add-on to existing clients, increasing revenue potential.

How to Price Marketing Mix Modeling Services

Pricing depends on project complexity, client size, and customization level. Consider this flexible pricing structure as a brief guide for agencies:

1. Basic Package: $5,000 – $10,000

• Small businesses with limited channels and budgets

• Fundamental MMM analysis and recommendations report

2. Intermediate Package: $10,000 – $25,000

• Medium-sized businesses with multiple channels and moderate budgets

• Comprehensive MMM analysis, recommendations report, and ongoing support

3. Advanced Package: $50,000 – $500,000+ (Used by large consultancy businesses)

• Large enterprises with complex strategies and significant budgets

• Customized MMM analysis, strategic recommendations report, ongoing support, and periodic reviews

Take Cassandra for a Test Drive

In conclusion, Cassandra is a cutting-edge marketing analytics platform that offers fast, affordable, and actionable MMM insights for businesses of all sizes. Its user-friendly product enables non-technical marketers to create their marketing mix, simulate marketing budget allocation decisions, and quickly analyze various scenarios.

If you want to learn more about Cassandra and how it can help you optimize your marketing efforts, visit their website at cassandra.app. You can start for free and create your first marketing mix model. If you need support or would like a demo, the customer success team is available to help you get started and work together to develop a marketing mix plan tailored to your brand.