Attribution is a major influencing factor in the shift to programmatic.
Attribution is a major influencing factor in the shift to programmatic.
Things to Consider when Building an Attribution Model
As brands, agencies, and publishers continuously seek new and more efficient methods of buying and selling digital inventory, programmatic advertising has gained rapid adoption and acceptance worldwide. We define programmatic as media exchanged through automated software: DSPs (Demand Side Platforms) on the buy side, and SSPs (Supply Side Platforms) on the sell side. Often this exchange of inventory and audiences involves RTB (Real-time Bidding). The latest studies indicate that over 60% of digital inventory is currently being bought and sold through programmatic pipes. Many agencies and brands are starting to mandate that programmatic buys be the first option over direct insertion orders with publishers. Along with control, attribution is a major influencing factor in this shift to programmatic.
Attribution is the process of identifying touch points along the consumer’s journey to conversion and assigning a credit model or value to each of these events. While traditional media buying ROI can be measured through offline redemption or directionally through lift analysis and research, digital advertising can be measured one-to-one at the user level using an ad server and attribution model. In either case, the ultimate goal is to learn which marketing dollars are working hardest and to develop media plans that maximize business outcomes.
In the case of programmatic advertising this generally coincides with ad exposure. It provides a model for determining the effectiveness of digital advertising and ROI across display, social, and search engine marketing. Two keys to successful attribution modelling are cross-device and cross-channel user matching. In other words, the ability to recognize the same user across their different devices and across different channels. Inherently there are some blind spots created by walled gardens (like Facebook) and user-tracking limitations stemming from things like privacy regulations (think GDPR). It is important to understand these limitations and develop a model that’s best suited for your business.
Steps to consider when building an attribution model:
Attribution models take time to master and it may take some trial and error to determine what works best for your business. The reward is data-driven media buying that can have a positive impact on sales. My advice – be patient and stay the course!
Angus Baldocchi