Conversion Attribution in Cross-Channel Marketing: Giving Credit Where Credit is Due

Inplace #2

ADOTAS — We roll our eyes and leave the room when the Grammy Award-winning rapper spends his 90 seconds of mic time at the Kodak Theater thanking everyone from his parents to his Little League baseball coach to Jesus. Digital marketers, however, could take a cue from 50 Cent.

In the advertising world, many factors contribute to the final outcome and all touch points along the entire interaction spectrum must be considered when evaluating the impact of the marketing strategy.

At the end of the day, no matter what the advertising strategy is, P&G wants to move Mach3 razors off the shelf. It’s imperative that all aspects that contribute to the sales receive a pat on the back. The reality is that with the most popular analysis strategies, channels like video or addressable TV often gets the short end of the stick when it comes to receiving credit for conversions.

We’re Doing It Wrong Today

The economics of video versus other digital advertising puts video at a disadvantage when the all-too-common “last view/click” attribution model is used. With one video ad for every 50 display ads and an even smaller fraction when compared to search, last interaction modeling is analogous to the NBA lottery, where display and search are more likely to win, given their number of balls in the hopper.

While last interaction is often the default model conversion management tools and it might make sense with an intent goal in a single-channel strategy, rarely is it the appropriate model for the growing channels of video for the upper funnel goals of marketers. Other models exist which allow for doling the other touch points varying levels of credit, but even this functionality in the existing tools is not fool-proof, as every marketing plan and its outcomes are different.

Multiple Channels

Appropriately crediting performance across a marketing plan of search, display and video is hard enough. How does the marketer do the same when we fold in social or connected TV? Did the fact that 25 tweens on Facebook liked my brand generate any purchases? Did the fact that I bought 25-34 Male Auto-Intenders through an addressable television ad exchange help with last month’s auto sales?

These newer channels force marketers to chase customers from PC to phone to TV to Facebook. We are held accountable for every penny we spray into each of these pools.

Not A Simple Task

On top of aggregating audiences across the various channels, heavy data analysis is required in order to understand the impact of each. A 2011 Experian Marketing Services study found that more than 1/3 of marketers considered understanding the efficacy of each component of multi-channel marketing strategy to be one of their biggest challenges.

The lack of a consistent cookie footprint, closed-box social media environments, the aversion of sharing user data and the fact that digital channels do not operate autonomously yields a non-trivial challenge.  This is not a task for the weary and certainly not for dataphobes, but is not impossible. Using first- and third-party data to link users across the marketing flow and advanced data platforms that improve with each passing month are the high-level antidotes that we need to take to attack this beast.

Imperfect But Elegant Solution

In contrast to traditional marketing initiatives, the digital world allows marketers to be more strategic with their details through measurement. This is crucial in the proper attribution of performance and eventual conversions in a media plan that spans numerous channels, including some new and highly sought-after, but very nascent mediums like addressable television.

Each and every form of interaction that a customer has with the brand can contribute to long-term marketing goals. It would be remiss of a marketer to ignore any initiative. Considering all marketing touch points in the funnel requires inordinate amounts of Big Data analysis — tracking users across the digital highway from device to device — and none of this comes easy or without measurable costs.

Data-driven distributed and algorithmic attribution across all touch points is something that has yet to be solved, despite the marketing brain power spent investigating this.

The long-standing basic attribution models imply causality between any (or many) of these touch-points and the marketing goal, but rarely can this be inferred when applied to unique marketing strategies, which have distinctive goals. Marketers must still get their hands dirty with tried-and-true A/B testing across the different customer interactions to understand which were correlated to performance (or lack thereof) to help drive strategy going forward.

Thanking Everyone

Marketers do realize that their customers are spending more time consuming content across a range of devices and, as such, are augmenting their marketing strategies to address this. One of digital’s biggest promises is accountability, but we’re failing to provide that with respect to proper attribution.

As the ad technology platforms and marketers make strides towards addressing some of the issues with commonly used attribution strategies, it will likely result in changes to marketing budget allocation, tearing down of agency political silos and setting advertisers on the correct path towards success.

Having spent a lot of time evangelizing analytics products for digital marketing, it has become clear that the demands of the marketing world are often eons ahead of the speed at which technology moves.

No one starts training for a marathon with a 22-mile run. In training to finish this conversion road race, we need to start with small jogs and mark off the miles one-by-one. Start by using different models and, in comparing the outcomes, we will start to understand which is most appropriate given our current problem. Once you’re sure that the compass is pointing in the right direction, let the technology lend a hand.

The fragmentation that is omnipresent in today’s marketing environment and the non-trivial analysis required to tell a story with the nearly infinite amount of data might deliver an educated, but imperfect conclusion, however it gets us closer to where we want to be. With time, intelligent advances in attribution will allow us to properly credit all of the avenues that bring us the marketing Grammy. Stay tuned…