ChinaManufacturing2023 – ongoingHONOR
Predictive machine learning models for ad audience targeting in Mexico
HONOR raised return on ad spend by 63% in Mexico after predictive models picked its ad audiences
Problem
HONOR used Google Analytics to pick ad audiences from user behavior on its ecommerce sites. Analytics could only find seed audiences among existing customers, so the audience range was limited. The company wanted a way to predict who was likely to buy.
Approach
In 2023 HONOR added BigQuery to support its Google Ads work. With help from the Google partner TrueMetrics, it built predictive machine learning models in BigQuery for its Mexican online campaigns in about two months. The models used behavior data from the Mexican ecommerce site to find users with a higher purchase probability. Those seed audiences were then used to find similar audiences in Google Analytics for ads in Google Ads. HONOR planned to roll the method out to other countries and to automate the transfer of results to Google Analytics, which it expected to cut manual work by 50%.
Results
Takeaway
HONOR started in one important market and measured the effect on ad results before extending the method. The company did not say what baseline the gains were measured against.
Categories
- Size
- Large enterprise
- Industry
- Manufacturing
- Business function
- Marketing and sales
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Google Cloud's customer story describes how HONOR used predictive machine learning models in BigQuery to find ad seed audiences for its Mexican online campaigns. It reports a 63% rise in return on ad spend, a 33.5% rise in ad conversion rate and a 40.6% fall in cost per click. The page carries no date; the Internet Archive first saved it on December 19, 2024.
Full details are in the original sources.
