Xaxis Norway / Netherlands, 2019

Converting on cruise control via Copilot for Ford Europe


The Ford Norway team looked to drive conversions on their site. An extensive analysis showed that visitors to Ford.no had vast differences in user behavior. 13.4 million historical data points were analyzed, and it became apparent that variables such as geography, car model, weekday and time of day have a significant impact on conversion rate and cost. Our hypothesis was that these holistic insights could be utilized to improve the custom-built programmatic buying algorithms, which historically were optimized based on self-generated campaign conversion data. Setting the goal to both increase campaign cost effectiveness and performance.

We used a full year worth of Adobe Analytics data from Ford to be as thorough as possible. All programmatic purchases were analyzed, along with website traffic and conversions from all digital sources – IO display, search, social, content, and organic. We then built nameplate-specific bidding algorithms based on the overall behavioral patterns observed on the Ford website and utilized Copilot, Xaxis’ proprietary AI, to apply these unique algorithms for optimization.
Xaxis and Mindshare developed tailored and automated bid strategies for each of Fords’ nameplates based on 8.4 million historical data points. The deep analysis of Ford’s data paid off and led to significantly improved results on all KPIs. This strategy was more cost-effective and better performing than all other channels on the media plan and led to significantly improved results on all KPIs with an average CPA decline of 44,6%, and a 55% decrease in CPA on harder conversions.** *Conversion points that show the highest statistical correlation to actual car sales. **Results are based on all programmatic Ford activity prior to project start (Q1/Q2 2018) versus the case time-period, Q3/Q4 2018. This includes all Ford advertised nameplates. From analysis of almost 5 million historical data points on Ford.nl, Xaxis and Mindshare developed unique automated bidding model per car type. During the campaign we A/B tested the custom strategy next the best Q3 copilot strategies. The custom model outperformed the other strategies on CPA, CR (click) and on ROAS*.This has led to significantly better overall campaign results with an average increase of 47% on ROAS, and an average CPA decrease of 43%.* *Results are based on all programmatic Ford activity from October 2018 to EOD
Decrease in overall CPA
Decrease in overall CPA

Want to learn more?

Get in touch to find out how outcome-driven media can deliver measurable results.