By Phil Dubois, CEO and co-founder, AdAmplify
For years, multi-touch attribution (MTA), buoyed by the much-vaunted 360-degree view of the customer, has been the holy grail of marketing measurement. Marketers dreamed of understanding the many stages of the customer journey so that they could optimize it from beginning to end. But with privacy changes driving signal loss, marketing mix modeling (MMM), a top-down approach, is enjoying a resurgence. And the chorus of MTA detractors continues to grow louder.
Privacy changes will affect some multi-touch attribution models, especially those that rely on third-party data. And MTA isn’t perfect, nor can it capture the entirety of every single customer journey. But MTA is still a powerful way to understand customer behavior and optimize marketing and it would be a mistake to succumb to statements as simplistic as “MTA is dead.”
Here’s what MTA can and can’t do, how privacy changes affect it and how marketers can implement both MTA and MMM to capitalize fully on measurement investments.
What MTA can and can’t do
MTA can help provide a more detailed view of how customers interact with various touchpoints along their journey and it can help businesses make more informed decisions about how to allocate their marketing budgets.
Of course, MTA, like other forms of attribution, can’t perfectly tell the entire story of every single customer journey. Here are some things that MTA can and can’t do.
MTA can:
Provide granular insights by tracking and measuring the impact of individual touchpoints, including those that are often overlooked by other attribution models. Commonly overlooked channels include social, email and display.
Determine the effectiveness of marketing campaigns and individual ad units. By tracking touchpoints, MTA can help businesses determine which channels, campaigns and creative are driving the most conversions and which ones are not.
Help marketers understand how long the gap is between purchases to foster loyalty by reactivating prospects at the optimal time.
What MTA can’t do:
Determine exactly how much each touchpoint influenced conversion: MTA is focused on tracking and measuring individual touchpoints but it can’t indicate with complete certainty that the last touch, as opposed to say the first touch, played the biggest role in a purchase. MTA deals in probabilities, not omniscient certainties.
Account for external factors: MTA is focused on measuring the impact of marketing touchpoints but it can’t account for external factors that may impact a customer’s decision to convert, such as economic conditions, competitive offerings or changing consumer preferences.
Replace human judgment: While MTA can provide valuable insights into marketing performance, it can’t replace the human judgment and experience needed to make strategic decisions about marketing budget allocation and overall strategy.
How privacy changes affect MTA’s effectiveness
Privacy changes are making it harder to track consumers across digital properties. Regulations like Europe’s GDPR and the California Privacy Rights Act are demanding that brands get consumers’ consent to share their data. And private companies are following suit. Apple has downgraded its mobile identifier and Google plans to nix third-party cookies on Chrome.
These changes will have a significant impact on MTA for companies that depend on third-party data, which changes like the elimination of third-party cookies on Safari and Chrome are intended to crack down on. Brands and retailers will need to adapt their tracking methods to comply with privacy regulations while still capturing enough data to produce accurate attribution models.
While signal loss will affect ecommerce companies that rely on third-party data, many will be able to use first-party data from their online stores to weather the storm. For example, Shopify sellers can track visitors who were influenced by their marketing campaigns. This means signal loss doesn’t necessarily erase the transparency into customer behavior that enables MTA.
Why MTA is still worth pursuing and how it complements MMM
MMM has a long history and is coming back in favor because of the development of new statistical tools that allow better analysis. MTA has a much shorter history but is also becoming more sophisticated through machine learning and other developing technologies.
MMM uses statistical models to evaluate the impact of different marketing channels on business outcomes, such as sales, revenue or customer acquisition. MMM is ideally suited to track how different marketing channels, especially offline, work together and how changes in spend levels affect business outcomes. So, it is also more helpful for larger enterprises with omnichannel businesses, less so for, say, online-only ecommerce companies.
Using Machine Learning, MTA can provide a more accurate view of how individual touchpoints contribute to conversions, which can help identify which specific tactics are driving performance. So, while MMM provides a high-level view of the overall impact of marketing on business outcomes, MTA can provide more detailed insights into how individual touchpoints are contributing to those outcomes.
These two approaches can be used together to get a more complete picture of how marketing is impacting the business. Additionally, MTA can help inform the inputs to MMM models. For example, MTA can help identify specific tactics that are driving performance, which can be used to inform the variable selection and weighting in MMM models. Similarly, MTA can provide insights into the effectiveness of creative, which can be used to improve the quality of input data in MMM models. MTA also provides greater visibility into the performance of online channels, whereas MMM has an edge offline.
Marketing discourse often deals in attention-grabbing proclamations such as “SEO is dead!” or “MMM is the new MTA!” In reality, thousands of companies rely on these technologies and disciplines because each of them brings something valuable to the table. MTA is not dead. MMM is not a useless alternative. Together, they are more than the sum of their parts.
