• by Ray Schultz , Columnist, October 22, 2018

When asked about attribution, an email specialist I know joked: “That’s above my pay scale.” 

It better not remain so if he expects to continue getting paid. Many firms are wrestling with cross-channel measurement, from email to social. And they are pouring money into it. But too few are getting it right, judging by The Business Case For Digital Attribution, a white paper by i-com. 

To hear i-com tell it, there are two ways to attribute. The one known as marketing-mix modeling (MMM), is from the top down.

That’s where you measure the contribution to sales of each channel, be it email, TV, direct mail or digital. The flaw in this process?

“MMM is longitudinal in nature, typically requiring 12-24 months of granular data inputs, and is performed annually, semi-annually or quarterly,” the study notes.

Then there is the bottoms-up approach — multi-touch attribution, or MTA. This process doesn’t merely look at disparate data sets — it provides a unified view of each customer’s conversation path. It is seen as complementary to MMM.

MTA has been around for awhile, but a small subset of brands have fully embraced it. The study cites a DMA surveying showing that only 13% of U.S. firms are using advanced attribution.  

What is preventing firms from using this purportedly superior approach? One hurdle is organizational silos that prevent brands from gaining a holistic cross-channel view of activity. Another is budgeting.

“In many instances, budgets and responsibilities are split between departments and agencies, without a designated person or team to lead cross-channel coordination,”

That leads us to the next problem: That brands sometimes invest in solutions without bringing key stakeholders on board and emphasizing the need for change. They often leave their agencies out of this process too.

Few companies now pursue last- or first-click measurement. This dated practice results in false positives and negatives, and it rewards bad behavior such as retargeting and “cookie-bobbing.” Similar problems accompany u-shaped, even weighted and time-decay allocation.

So what does MTA, which i-com apparently is peddling, offer to end users? According to i-com, MTA:

Includes all paid and non-paid interactions for both converting and non-converting users

Uses machine-learning, algorithmic models to determine “the relative contribution of each impression click visit and other interaction.

Allocates credit to each paid and non-paid interaction based on the modeled results

Reports the attribution conversion of each channel vendor, strategy, placement and creative.

Finally, it allocates credit where it’s due, and provides better integration of “anonymous user data and campaign data by ad serving, search, email, social and affiliate platforms.”