What’s the special sauce behind ShareThis campaigns? How does ShareThis work for advertisers? What data goes into your optimization techniques? These are questions we here at ShareThis get all the time. The answers are both simple and complex. Simple and easy solutions for advertisers and marketers and complex data science and mining that powers our platform.
ShareThis has a built a comprehensive predictive optimization platform to deliver the best campaign performance. Every campaign has certain marketing objectives – direct response or “DR” campaigns in particular are often measured against a KPI such as Cost Per Action (CPA) or Cost per Click (CPC). There are many different and complex KPI metrics to measure and track, but for simplicity we can stick to CPA or CPC for this discussion.
Once an Insertion Order (IO) is signed, the first phase for planning a campaign begins.
And the question of how and where to best deliver a marketers ads, which audiences to target and reach so on are presented. These questions become the initial input to the “Predictive Campaign Optimization” platform. The optimization platform continuously readjusts and tunes various optimization knobs along the lifecycle of the campaign the plan to ensure these goals are achieved.
The ShareThis predictive optimization platform utilizes two level of optimization: An outer loop optimization and an inner loop optimization. In the outer loop optimization, the delivery and allocation across inventory is readjusted hourly and daily utilizing pricing and performance data but keeping the KPI goal in the view. The actual delivery is done via RTB (Real-Time Bidding) to ensure we are always securing the best, highest performing inventory for our customers, in real-time. The RTB does an inner-loop optimization to achieve the delivery goals. The delivery goal is readjusted frequently not only taking the KPI goal such as conversions and clicks but also factors like seasonality and day of week. A predictive model uses past performance (e.g., domains, segments…), pricing, and several factors to determine the plan. This process is continuously performed to optimize campaigns at various time boundaries. This data is provided to the RTB (Real-Time Bidder) to do the inner loop optimization and delivery.
The outer loop provides allocation and pricing guidance on multiple dimensions such as domain and segment to RTB. The outer loop has a broader context to optimize the campaigns; many variables are considered for the optimizations such as networks, domains, segments, time, day, price, etc.
Modeling in order to make campaigns the most effective and drive the highest ROI, is a fascinating and complex problem that requires understanding of data, building great predictive algorithms and efficient systems to make it happen.
We’re thrilled at ShareThis to be able to leverage sharing data, SQI (Sharing Quality Index) predictor and campaign performance data to optimize each and every campaign to meet the goals of our customers efficiently and economically.
By Prasanta Behera, Chief Architect, ShareThis
