As an agricultural supplies client has a very specific target audience. Their products span from ridable lawn mowers to commercial mowing, golf, and turf equipment. Most of the US population does not have the space or need for the latter, and those that do own rideable equipment often have a strong affinity to a particular brand. Not only is the audience limited, but it is extremely difficult to attribute their offline sales to their digital media spend as their products are sold in person at a third-party location. Historically, they had optimized their programmatic media to drive activity to their website in an effort to generate brand affinity.
The Xaxis team worked closely with Mindshare to thoroughly review all available data and metrics of media success. The most influential insights were optimized to directly with Copilot, Xaxis' proprietary AI technology. Copilot strategies work on top of DSP optimization to make more specific bidding decisions. The machine learning solution was able to identify even more granular performance trends with its clustering algorithm that resulted in an evolution of the goal metric. Weather data, and ultimately third-party offline foot traffic data were implemented into a custom blended KPI for the advertiser.
Human and artificial intelligence are used in concert for a bespoke strategy that gives the brand a competitive edge.
The Xaxis team started with a deep dive of the client’s data to identify a custom outcome that was closer to equipment sales. The solution evolved over three tests.
The first test used artificial intelligence technology to improve CPA optimization and increase the number of clicks through to the client’s site. A clustering algorithm was applied to optimize towards a client-set CPA goal.
A later campaign used cost per landing page as a proxy to purchases in store. The Xaxis analytics team found a strong correlation between conversions and average temperature in a region over a ten-day period. The team hypothesized that including temperature data into their Copilot strategy and optimizing away from extreme temperatures would be more efficient in driving landing page visits.
In the third activation, Xaxis partnered with a foot traffic data provider to link digital activities with store visitation. The goal was to understand how website behavior can predict dealership visits. Xaxis provided a file of dealership locations by latitude and longitude to the third party provider to measure in-store visits. Website data, including each visitor’s timestamp and URL details, was collected over the same time period. A custom probability score was developed for each page that calculated a user's likelihood to visit a dealership based on their site browsing behavior. Copilot is currently using these scores to maximize in-store visits.
Over the last nine months, data-driven insights have been key to the strategy for this advertiser. Every execution running with Copilot optimization using unique KPIs and data is supported by a 35+ member team of Data Scientists, Engineers, and Analysts to drive stronger outcomes.