Skip links

Mitsubishi Motors Canada uses propensity modeling to increase the conversion rate

The Goals
  • Harness first-party data to predict consumer behavior.
  • Reach customers that are likely to make a car purchase.
  • Drive valuable traffic to the car pricing tool on the Mitsubishi website.
The Approach
  • Used Analytics 360 data to create a propensity model in BigQuery.
  • Connected Display & Video 360 with Analytics 360 to engage with custom segments of likely-to-convert visitors.
The Results
  • 1.8x decrease in cost-per-action.
  • 14x higher conversion rate.
  • Modeled audiences drive 70% of total conversions.

Mitsubishi Motors has pioneered automotive innovation for more than a century. The global brand established itself in Canada in 2002 and has continued championing leading automotive technologies. Its latest car, the Outlander, has virtual assistance to help drivers with lane centering, speed control, traffic jam stop-and-go, and much more. With this new vehicle driving Mitsubishi further into the future, the automaker wanted to channel that innovation into its marketing strategy.

The marketing team at Mitsubishi realized that consumers were increasing their time researching product information online before purchasing. With this knowledge, the automaker decided to shift to a digital-first strategy. The “build-and-price” tool on their website took center in their approach. This interface allows consumers to compare features and costs before visiting a local dealership. Mitsubishi used this tool as a “proxy conversion” because it’s the clearest digital signal of interest in purchasing the vehicle. The carmaker wanted to acquire more “proxy conversions,” which would eventually help the brand drive more car sales at dealerships.

Mitsubishi’s ad tech solutions partner, Dandelion, proposed using Google Marketing Platform and Google Cloud to harness the power of the brand’s first-party data and predict which consumers were more likely to use the build-and-price tool. Mitsubishi would then be able to build digital marketing campaigns to reach those customers.

Predicting the future with Google Analytics 360

Dandelion first exported Mitsubishi’s Google Analytics 360 data to BigQuery, part of Google Cloud. With BigQuery ML, they created and executed a machine learning model that predicted consumers’ likelihood or propensity to complete the build-and-price tool. This enabled Dandelion to develop a custom audience that could reach Mitsubishi’s ideal customer.

The next step was to utilize these models. By importing these audiences into Analytics 360, Mitsubishi could make them available for activation across multiple Google advertising solutions.

Making the right connections with Display & Video 360

With those audiences automatically available for activation in Display & Video 360, Mitsubishi easily reached its ideal audience. The brand’s creative digital agency Jan Kelley built the media campaign using these audiences, determining the best channels to activate. Over the six months that the campaign lasted, Mitsubishi saw an 81% decrease in their cost per acquisition and a 14x increase in their conversion rate.

After seeing these convincing results, Mitsubishi decided to use propensity modeling for its future campaigns. They plan to activate it in other solutions like Search Ads 360 and continue channeling their innovative spirit.