ASUS Achieves Record Conversion Rate with Machine Learning on First-Party Data

Challenge

With the substantial volume of daily users visiting the ASUS e-commerce site, determining an effective online media strategy that is capable of distinguishing users at different stages of the purchase funnel has always been a challenge.

‍

Approach

With the data-driven solution, we are able to rate the individual audience with a variety of in-depth signals, like Session depth, Engagement rate & Conversion probability which is highly complex if only with manual work.

“User with high conversion probability should be communicated differently and aggressively.” is the core concept of this methodology. We prioritize media budget on audiences with the highest potential to enroll or purchase, by identifying high-value clients early in the process using digital signals. For example as the image shows, given the signal of high possibility to purchase, we should have a higher bidding price and a more aggressive communication strategy for visitor #2.

‍

‍

Results

Accurate user scoring has helped ASUS to better understand the profiles of their high-value visitors, as well as optimize budget planning, concentrate more resources on high-value users, and ultimately improve the ROAS of their media campaigns. Huge performance uplifts have been seen across media channels and the program is currently being rolled out across new markets.

‍