It’s an exciting time to work in our industry. Every day brings massive corporate mergers, rapid technological advancement, evolving buy and sell side business structures and processes to re-organize around the centrality of big-data at all levels of decision making.
The amount of information and computing power that marketers now have at their fingertips brings us ever closer to the elusive promise that digital media has long held: robust business intelligence about what consumers like and don’t like, deployed in real time within a media environment of always-on consumer addressability. Every fact a marketer needs to know, and millions she doesn’t, all collected, analyzed, and acted upon instantly via automated tools to create an ongoing consumer feedback loop and an intimate portrait of consumer preference.
As the industry acclimates to the unprecedented control and precision programmatic tools afford, we unfortunately have made little progress in one glaringly obvious, but important area—attribution. We need to evolve beyond simplistic last touch methodologies (which assign all credit for an action to the last point of contact) to optimize digital investment. The case against last touch is obvious. TV buyers have long developed campaigns based on the notion that message frequency against a target (or multiple views of the same ad) is a necessary condition for building consumer awareness and message recall. We’ve understood since the 60’s that each touch point contributes and adds value in driving towards a marketing objective.
And yet, despite this common-sense notion and many multi-touch attribution (MTA) platforms available in the marketplace, far too many advertisers and agencies still optimize their digital media investment against last touch. There are myriad reasons, as well as plenty of well-documented cross-screen measurement problems that have dissuaded marketers from thinking creatively about their measurement options, but this indifference nonetheless translates into billions of dollars of media investment every year into areas that have no substantial impact on bottom line business performance. This drastically distorts resource allocations across the marketplace, usually at the expense of publishers within high and mid funnel touch points who often invest heavily in quality content and brand experiences that add real value to digital media campaigns.
This is why the IAB Performance Committee has focused specifically on buy side attribution guidance in 2016. The first step in our attribution agenda was to establish a common language when discussing attribution tools, technologies, and methodologies. We completely rebuilt and modernized our Attribution Primer to account for the many developments in cross screen measurement and data collection over the past five years. This update fills six gaps:
- We incorporated user level device mapping as a necessary foundation for understanding cross screen path to conversion. This requires definitions of probabilistic and deterministic measurement.
- We focus on path to conversion analysis as central for practitioners to infer behavioral scenarios that are likely to produce desired outcomes.
- We outline a general consensus on the basic ways to apply credit to events in the path: simplistic approaches, fractional rules based approaches, and fractional algorithmic approaches.
- We discuss the need to account for online to offline attribution options, specifically store visits, purchases, and phone calls.
- We describe common challenges around attribution model scope, optimization lag, and managing optimization throughout purchase funnel.
- We review the need for practitioners to focus on better, not perfect solutions given the tools available to the industry. Similarly, we discuss the need to use an iterative, long-term process to refine methodologies over time based on data collected.
We hope this primer provides a helpful reference point to educate the marketplace around the limitations of and alternatives to last touch methodologies. The committee will continue to build upon this guidance over the next year and develop more acute material on buy side attribution best practices. If you’d like to get involved, please reach out to me directly at [email protected].