PHASE 2:

OPTIMIZE THE ROI OF EACH TOUCHPOINT

Customer journeys have grown increasingly complex over the last decade due to the number of devices people use and the wealth of information available online. This has led to a huge increase in the number of touchpoints along the path to purchase.

Understanding attribution modeling

Because consumers no longer follow a straight path from awareness to consideration to purchase, the practice of attribution modeling, or determining how much credit for sales or conversions should be assigned to various touchpoints, is increasingly important. For example, marketers need to know how much ROI to credit to Facebook, Google AdWords, and other media channels to determine whether they should increase or decrease spend with each partner.

Marketers can use many different approaches to attribution, from models that give full credit to the first or last touch to ones that apportion credit among various touchpoints. The optimal choice depends on the marketer’s sales cycle and mix of the touchpoints involved. Many marketing organizations will start with a static allocation for each touchpoint (for example, two points for a website visit or one point for a Facebook ad impression) and then use data-driven modeling to update the allocations in real time based on how much each channel contributes to conversions. But before organizations can use any attribution modeling, customer data from every marketing channel has to be kept in one unified location.

Driving ROI from attribution models

Marketing organizations can build their own attribution models or select an off-the-shelf solution from companies such as Tableau and Power BI. From there, they can make spend decisions based on actual revenue impact (specifically, how much each customer touchpoint is contributing to conversions). In practice, this involves monitoring campaign performance on a weekly or even daily basis and reallocating spend to ensure the channels receive a share commensurate with their impact on sales and conversions. The marketing mix should then be rigorously evaluated and adjusted every month or quarter.

“Centralizing data in Snowflake allowed us to do the impossible and prototype in a matter of days.”

JACOB MULLIGAN,

Head of Analytics, Firefly Health

Case Study

HOW FIREFLY HEALTH BUILT A POWERFUL ATTRIBUTION MODEL IN JUST DAYS

Trying to obtain data-driven insights into patient health, virtual healthcare company Firefly Health relied on a combination of MySQL Workbench and Google Sheets. Seeking a more scalable approach to data analytics, the company’s analytics team began exploring data architecture enhancements, but siloed data sets prevented them from developing marketing attribution models and inhibited them from effectively tracking clinician productivity.

Realizing the need for a single source of truth for all of their analytics use cases, the team turned to Snowflake. Snowflake’s easy-to-navigate interface, comprehensive documentation, and integration with Fivetran and Looker enabled Firefly to consolidate the Snowflake implementation into a one-day project. Ingesting data from AppsFlyer (a mobile attribution platform) into Snowflake via Fivetran enabled the rapid development of Firefly Health’s marketing attribution model.

Now successfully deployed, Firefly Health’s marketing attribution model provides campaign-level insights, enabling visibility into customer journeys and maximizing return on advertising spend.