AccuWeather is a global leader with over 55 years of global forecasting experience. They are the most accurate and most trusted weather brand worldwide, serving over 30 billion data requests each day and providing forecasts in more than 100 languages and dialects.
With the primary revenue on their website coming from Google ad placements, AccuWeather’s goals were to:
- Drive sales through better pricing of advertisements.
- Increase in ads revenue by better pricing.
- Increase in Google cloud consumption by use of BigQuery, Storage and Cloud ML.
During a 12-week period, SpringML and AccuWeather data scientists focused on programmatic ads through the Google Ads platform as this platform provided the most data to work with and made up 60% of Accuweather’s programmatic. The two questions analyzed were:
- Does precipitation change the advertising conversion rate?
- Does high pollen counts change the advertising conversion rate?
The following questions and factors were also examined in order to model the correlation between weather events and advertising success:
- Which ads succeed and with whom?
- What is the user’s geography and what geography are they looking at?
- What categories/industries enjoy success?
- Where in the customer journey is the customer when they decide to click through?
SpringML’s Machine Learning model leveraged forecasting and outlier detection techniques and delivered meaningful insights directly to Management.
The project outcomes included:
- Meaningful insights for advertising customers
- Better pricing of advertisements
A better understanding of how meteorological events drive click-through rates as well as which factors make advertisements more effective.