Online Customer Acquisition • Quality Customers Through Machine Learning
Results
-
82% recall of cases within the desired audience.
-
Reduced timelines from 120 days to 24 hours.
-
BBVA was able to make real-time decisions.
A journey of quality and speed.
In the financial industry, banks are accustomed to attracting new customers through credit card promotions and other benefits. But although the bank was investing in digital efforts to attract new clients, its usual audience failed to match their desired target. Faced with this challenge, we set a clear goal: to generate higher quality traffic without excluding any real users.
Optimizing the online customer acquisition process.
Across the entire consumer journey, the multiple interactions between the bank and its clients provide valuable information that can be leveraged to optimize processes. In order to improve efficiency in the use of this data, we carried out an analysis and refinement procedure—to then dynamically send the data to Google Analytics 360.
Measuring user behavior within the site—from how much time people spent on it, to the kind of information they provided while registering—became an essential part of the process. Then, we proceeded to combine online and offline information to obtain a behavior per user data chart, ready to be treated in a machine learning model developed through AutoML.
A predictive model that’s always on target.
The model allowed us to recover 82% of the cases that didn’t align with the bank's target and, through the Cloud functions, send this score to advertising platforms. Now with the necessary tools to make decisions in real-time, we reduced timelines from 120 days (the days it used to take the bank to draw conclusions about new clients) to 24 hours—achieving high quality at record speed.
