Why Marketing Attribution Hasn’t Lived Up to the Hype—Yet
By Trevor Testwuide, co-founder and CEO at Measured
Cross-channel marketing measurement has been through a tremendous evolution over the last 10-plus years. Today more than ever, cross-channel measurement is evolving and adapting to modern data collection realities.
In the early and mid-2000s, marketing data and analytics for digital media was just a curiosity that no one knew how to capitalize on. The first ones to figure out how to capture and measure media assumed a significant competitive advantage for a while. Over the last decade, marketing data analysis has become table stakes for any brand looking to grow or evolve.
A decade ago, digital advertising was mainly limited to display and search, and those programs were tiny compared to offline marketing.
Today, the average digital marketing mix consists of a comprehensive social strategy, online video, programmatic display and search. These are complemented by a mix of non-addressable and offline tactics such as podcast, TV and direct mail to make for a broad paid media portfolio.
While the opportunities for a brand to reach and stimulate its audience have evolved, marketing measurement has not kept pace.
The media opportunities to target a person with the right message at the right time and place on the right device has exploded. Along with this, the data set and opportunities for analysis have also become overwhelming for most.
For as much change and progress as there has been, marketers still struggle to execute data-driven decisions that drive growth in revenue and profits for their businesses. Advances in data and measurement and the evolving rules of tracking and identity management have led to new challenges. Marketers still must seek a competitive advantage, particularly as more tools and data have flooded the market.
With all of these changes, it’s clear that some of the progress we thought would change the world has fallen flat, but the progress continues with some new promising possibilities.
Multi-touch attribution has fallen short
This one hits close to home because I spent six very meaningful years of my career living in the MTA category. The people I worked with were talented and had the best intentions.
We thought attribution via terabytes of user-level data, billons of cookies and millions of converting and non-converting sequences would unlock the secrets of what made customers buy something or not. MTA was supposed to spark a revolution in marketing decision-making and show us which ads to buy to ensure a sale.
It turns out that user-level data is messy and the politics of ad tech played a major role. Instead of a clear path through a dark forest, we were trying to follow a trail of breadcrumbs obscured by dirt and fallen leaves.
We learned that it is essentially impossible to track users accurately across multiple channels and devices over a reasonable time.
The biggest non-search recipients of digital budgets – Facebook, Instagram, YouTube and Google Display Network – represent more than 80% of spending, but their impressions can’t be tracked and mapped to one another at the user level.
All that is left is paid search, and you don’t need MTA to optimize search.
Long before Cambridge Analytica and GDPR, MTA was dead-on-arrival. Now GDPR and other privacy laws are making MTA deployments more expensive while the walled gardens shrink the impact MTA can have.
Marketing mix modeling is too much for most
Marketing Mix Modeling, when developed with a deep understanding of the business, can be a valuable strategic tool. Its strength in modeling both media and non-media data sources can make it a powerful tool for the enterprise marketer to inform annual and quarterly strategy planning.
The reality of MMM is there is quite a bit of art to the science. The modeler is most effective when she has a keen understanding for the nuances of the business such as competition, macroeconomic influences, momentum and seasonality.
The problem with MMM is that its success depends on several factors, including the size of non-addressable marketing budgets, the number of offline transactions and the availability of large steady-state multi-year historical data sets.
Those requirements mean any MMM benefits are minimal for marketers in rapidly changing markets.
Combine those requirements and limitations with the dependence on third-party consulting services, and you end up with a great tool that is too expensive for companies outside the Fortune 1000.
Incrementality testing is becoming a must-have tool
Every marketing team engages in some flavor of A/B testing with various messaging and tactics, but the quality of testing varies from ad hoc programs to those supported by advanced data science teams.
Uber and Netflix have spent millions building experimentation capabilities in-house. They have entire teams executing tests to inform the incrementality of various marketing programs.
But applying incrementality testing to cross-channel media decisioning is a cross-functional challenge fraught with complexity for any team.
Even for the best and brightest in-house teams, it would be steep challenge to assemble the required expertise across marketing analytics, adtech, data science, data engineering and product into an orchestrated practice to solve for cross-channel incrementality measurement.
Meanwhile, marketers at smaller organizations leverage vendor-provided testing tools to compare creative messaging or seek answers to media lift. The simple truth is there are powerful advanced incrementality measurement capabilities available, and deploying an in-house practice to seek parity is expensive, timely and high risk.
Ad hoc split-testing won’t lead to a meaningful competitive advantage or a winning customer acquisition strategy. Marketers need an always-on best-in-class design of experiments to inform their most meaningful high-value decisions for growth.
Over the last two decades, marketers have seen a major evolution in performance-driven media. In parallel, the industry has understandably been frustrated with the limitations of measurement, especially given the abundance of big data.
Incrementality measurement done right is the path forward for marketers to inform trusted high-value media investment decisions.