Using Data-Driven Approaches to Grow a Company User Base

At the recent Real Machine summit in Las Vegas, Rui Chen Head of Samsung Pay Data Analytics presented a talk on how Samsung is leveraging robust data analytics to grow and retain their user base. Along the way, Chen discussed ways machine learning is enhancing the company’s ability to generate usable, actionable insights from masses of consumer data.

Watch the full presentation here:


3 Main Data Analysis Goals

Chen started with an overview of the product. Samsung Pay is a mobile payment solution (a.k.a. “digital wallet”) that lets users consolidate payment cards on their phones. It also supports membership, reward and gift cards.
Because Samsung Pay’s popularity is growing fast, the company wants to extract the most meaning from the large amounts of consumer data it generates. Data-driven insights support better marketing ROI, reduce resource waste, and help Samsung deliver more relevant user experiences.

Chen outlined three main goals of the data analytics function for Samsung Pay:

1. Attract more first-time users
2. Improve engagement of people who have the app but aren’t using it
3. Retain more customers who are regularly using Samsung Pay

For the first goal, Samsung is using data analytics to identify people most likely to accept offers, open emails, and register for the solution, and then using this information to support targeted email campaigns.

The second goal is aimed at users who have added at least one card to the app but aren’t using it regularly. To increase engagement with this group, rigorous analytics of consumer data helps the company build focused, specific messaging and deliver it at the right time in the user cycle. This, in turn, minimizes resource drain that often goes along with less granular marketing approaches.
The last goal—retaining people who are already using Samsung Pay regularly—garnered a lot of attention during Chen’s talk, in part because it’s here that machine learning is proving advantageous for Samsung.

Customer Lifetime Value & Feature Analysis

The company needs, first of all, a deep understanding the consumer lifecycle. For example, knowing when users enter the system and make their first transactions, and the nature of each user’s transaction pattern, are crucial to understanding their lifetime relationship with the company.

It’s also critical to analyze lifetime user value and the probability of churn. User value is addressed through retention analysis in combination with cohort analysis. By analyzing hundreds of factors potentially affecting retention, Samsung has been able to generate very valuable insights, such as the knowledge that users with membership cards are retained at a rate that’s at least twice as high as those without.

This feature analysis has another really outstanding benefit: highlighting factors that can be engineered into future models of the app.

According to Chen, it’s in the area of churn probability, or user reliability, that machine learning is especially helpful. Among the many predictive models Samsung is using, the most useful are logistic regression and random forests. Chen noted the power of the machine to balance different models, each with different strengths, to get the results the company needs.

All these tactics support Samsung’s three goals around growing and retaining its user base. The company is also using data analytics to measure campaign effectiveness from a range of indicators.

Rui Chen’s presentation was a great illustration of the potential and power of data-driven approaches to understanding and influencing user behavior in enterprises.

Learn more about machine learning and machine learning use cases here.

Subscribe to stay infomed