Marketing Mix Modeling and Multi Touch Attribution

In today's digital age, where the constant flow of information is data and the interconnection of platforms has revolutionized the way companies interact with their customers. customersmarketing has evolved in unimaginable ways. While the emergence of digital marketing promised an era of accurate measurement and analysisThe reality has proven to be more complex. The initial promise of "measuring it all" faces challenges in a world that increasingly values the privacy and data autonomy.

In this context, organizations are at a crossroads. While the industry giants are adopting strategies that favor them, creating walled gardens (Walled Gardens) of data, companies are looking for tools and methodologies that allow them to navigate this changing landscape and gain valuable insights about the impact of your marketing strategies. This is where the combination of Multi Touch Attribution models (MTA) and Marketing Mix Modeling (MMM) is presented as a promising solution.

This paper seeks to explore the synergy between MTA and MMMWe will outline how their integration can offer a more holistic and accurate view of the effectiveness of marketing strategies and efforts in an ever-evolving digital world. Through a detailed walkthrough, we will investigate the characteristics, strengths and limitations of both modelsculminating in a proposed solution that integrates the best of both worlds.

Welcome to a journey towards a more integrated and efficient attribution in the contemporary digital landscape.

Glossary of terms

MMM (Marketing Mix Modeling)Originated in the era of the traditional advertisingIn the MMM approach, statistical techniques are used to evaluate the impact of different marketing channels on vital metrics as well as sales. It is designed to look at the big picture, considering long-term influences and branding strategies.

MTA (Multi Touch Attribution)This methodology traces and assigns value to each point of contact that a customer has with a brand before a conversion is made. With the rise of the digital world and the multiplicity of channels, the MTA allows you to analyze the customer journey in detail.

The MMM and MTA integration not only provides a holistic understanding of marketing efforts and actions, but it also optimizes the budget, allowing companies to allocate resources efficiently.

MTA and MMM integration

The scenario is becoming increasingly complex and considering the uncertainty at the international level, the professionals of marketing and sales need more than ever to validate the business impact of their strategies and actions.

In order to meet these challenges, based on NATEEVO's years of experience, as well as on the IAB's work on the subject, we propose to review the different solutions available and clarify when it will be more convenient to apply one, the other or one combination of them.

In the first part we will try to understand each of the models, together with the necessary experimentation.

Subsequently, we will further elaborate on our proposal for the unificationThe following is the key to choosing the most convenient option for measuring your marketing strategy.

Characteristics of both models

What the MTA models provide us with

While for a significant number of years the standard of measurement was the attribution Single Touch of last touch. Today's attribution models Multi Touch (MTA) with the different digital attribution models (Last Click, First Click,...) and above all with the emergence of digital attribution models (Last Click, First Click,...) and above all with the emergence of digital attribution models (Last Click, First Click,...). Data Drivenare the most commonly used models. They are characterized by the following features:

  • It uses the available digital interactions (those that have converted as well as those that have not) to calculate holistic attribution for different digital MK actions.
  • It includes the Raw Data of the Digital Customer Journey. Results can be obtained for all taxonomy and data levels (e.g. Campaign, Site, Media, Keywords...).
  • Analyze the impact of marketing investments, not only at the media level, it is also possible to contrast the results of different advertising campaigns and their creativities (Online).
  • Determine the effectiveness of advertising actions in the short and medium term and possible changes in effectiveness trends.
  • Detect possible synergies between media to enhance their effectiveness.
  • Allows the visibility of the user's available Path, allowing to confirm the performance of our Digital MK strategies (Prospecting, Remarketing....).
  • Understand user and media behaviors, as well as the impact of different variables.
  • It makes it possible to evaluate the frequency of impact on users.

What the MMM models provide us with

Marketing Mix Modeling (MMM) has a different methodology with respect to the characteristics and origin of its data, the different modeling techniques, as well as the results and use cases it solves.

Although modern MMM are being updated with new functionalities and analytical capabilities, its operation is based on the MMMs applied in the traditional marketing since the middle of the 20th century.

The MMM are characterized by:

  • It allows the analysis of the different levers that have an impact according to regressive statistical techniques (aggregated data).
  • The results are limited to the starting level of aggregation of the data.
  • It allows to understand how competitors' actions impact (if such information is included).
  • It includes the analysis of the impact of marketing investments, not only at the level of digital media, it is also possible to contrast the results of different advertising campaigns (Offline & Online).
  • It facilitates the inclusion of other levers, such as seasonality aspects as well as inventory availability.
  • Determine the effectiveness of advertising actions in the long term.
  • It facilitates the detection of possible synergies between media, which enhance their effectiveness.
  • It allows the analysis of the non-linear behavior of the media and from these, their saturation levels.
  • Based on saturation levels, scenarios can be generated to optimize the media strategy, depending on budgets.
  • The models allow predictions and the setting of business objectives.

Current limitations of MTAs

Goodbye to the complete view of the user paths on Adtech's platforms.

Traditionally, the Adtech industry has used Tag-based technology (in-house or from Adservers) for the measurement of impressions and clicks in our advertisements. This allowed duplicate conversions between the different media (Google, Meta and others) as well as to be able to analyze Path with the interactions both from click-like impressions analyzing the complete and unified user path up to the conversion and thus enabling MTA analyses, such as advanced data-driven attribution models.

This methodology, always limited by the real availability of data (multi-device, multi-environment...) was the main technique of advanced analysis in digital marketing. But the increase in limitationssuch as:

  • Measuring Goal impressions.
  • Multi-device measurement (Apple)
  • Third-party cookie blocking in browsers.

They have limited the value contribution of the methodology by making it difficult to global vision.

However, the MTAs maintain a large value contribution in the following areas:

  • Regarding its capacity for granular data analysis.
  • For its application value to different optimization options, within the same digital channel.
  • The consequent advantage for data activation and automation techniques.

The holistic unification solution MTA + MMM + Experimentation

The most complete solution requires the combination of MTA + MMM methodologiesThe company's strategy is based on a constant and diversified strategy of experimentation in order to have the right information for data-driven decision making.

MTA analyses provide us with direct information about the interactions with usersallowing us to know:

  • The frequencies and impacts of each media, as well as at the journeys level and the overlap between the different media.

The specific performance of each campaign, creative, advertisement. Allowing your analysis through different attribution models (Linear and probabilistic DD).

The MMM complements the results of the MTA for strategic decisions regarding investment in the different channels and helps to "improve the efficiency of the MTA".calibrate"DD models. It allows us to analyze and improve the media mix thanks to:

  • Incorporate a more global and practical view of marketing campaign performance that improves media mix planning.
  • Evaluate changes in specific products, incorporate external media such as TV and press, as well as other external factors that may impact results.

Together with the experimentationwill help us to confirm hypotheses and statistically validate the experiments. that can have the greatest impact on our marketing strategy. The main recommended experimentation methodologies are:

  • Incrementability analysis.
  • GEO analysis.
  • User media test A/ B

Considerations to be taken into account

Updated Technological Environment:

Innovations in artificial intelligence and machine learning are redefining the precision and applicability of these tools, offering insights more accurate and personalized.

Ethical and Privacy Challenges:

Detailed analysis involves large-scale data collection. It is imperative that organizations handle this information with care, respecting regulations such as GDPR and CCPA.

Consumer Perspective:

Through a proper combination of MMM and MTA, it is possible to design more personalized campaigns that improve the customer experience and, therefore, strengthen the relationship with the brand.

Cost and Return on Investment (ROI):

Although these methodologies may require considerable upfront investment, the potential return in terms of more informed strategic decisions and better optimized campaigns is significant.

Integration with other tools:

When deployed on Data Platform systems and integrated with marketing automation CRMs and other marketing platforms, MMM and MTAs can provide a more comprehensive analysisconnecting data points and revealing deep insights into customer behavior.

AUpgrades and Continuous Improvement:

The field of marketing is evolving rapidly. As a result, these tools must be regularly reviewed and adapted, ensuring that companies keep up with trends and continuously optimize their strategies based on current data.

Conclusions

The increasing importance of user privacy, its impact at the level of the company, and the Wallet Garden and gradual elimination of Cookies The third part, shifts attention to MMM models that are immune to these changes. However, the value of Touch Points The direct contact with the users makes MTAs long-lived, even if they require the consent of the user. useras well as with the support of new technologies (mainly new IDs) as it offers great value for marketing automation and direct activation, which MMMs can complement, but not replace.

As has been the case for some time, the departments of offline and online marketing remained separated in silos. Today, the breaking down of silos at the organizational level is bringing a great increase in value to organizations. The unification and combination of MTA and MMM methodologies y Experimentation is the technical answer that gives us an increasingly holistic view of the levers and actions that affect our marketing activity and therefore our customers in a more complete way.

It is critical that companies do not adopt a "one size fits all" approach to measurement and attribution. Depending on the nature of their business, marketing strategies, channels and budgetsSome may benefit more from MMM models, others from MTAs, and many from a combination of both. The key is to maintain flexibility and a willingness to adapt to an ever-changing digital environment.

In addition, the role of experimentation cannot be underestimated. Experimentation allows brands to validate hypotheses and strategiesensuring that marketing efforts and investments are targeted in the most efficient way.

Finally, companies must be aware of their budget and available resources. While it's true that technology and cloud solutions are making advanced analytics and attribution solutions more accessiblecompanies still need to carefully evaluate the profitability and the value of these tools depending on your particular situation.

In short, in a world post-cookie and with a renewed emphasis on user privacyFor brands, it is essential that they remain proactive, adaptive and well-informed to ensure success in their digital marketing strategies.

Stay up to date

Fill in your details to sign up for our CDP and CDA newsletter and we will send you all content updates.