My name is Lori Hood, and I'm the Chief Marketing Officer at Mobilewalla. I want to introduce you to our new video series on using alternative data to better predict churn in the telecommunications industry. Over three brief episodes, we're going to examine how effectively using data, particularly a combination of first and third-party data, can give you an edge at predicting and preventing churn.

The telecommunications industry is highly competitive. Data tells us that acquiring a new subscriber can cost 5 to 25 times more than retaining an existing one, and that globally, the churn rate for Telos is in the 20 to 30% range. You are modeling churn today, but what can you do to improve your results?

Machine learning is nothing without data, and advanced modeling requires rich, diverse data sets. Your first-party data is great, but it's limited in breadth, depth, and scale, and subscriber context. Third-party data adds the scale that's critical to accurate predictive modeling and is especially helpful in providing the key missing external attributes that can drive churn.

Your first-party data is a given as a key input into your churn models, but as we've discussed, the lack of breadth, depth, and external context can impact the performance of those models. Second-party data, which is often sourced through some sort of partnership, while high quality, is typically very difficult to come by and very expensive. Third-party data can help add the context and fill in the gaps left by your first-party data.

Third-party data can provide demographic and psychographic data, but it can also add external factors that influence churn, such as tenure, device type, mobility, and consumer behavior. Third-party data can also give you additional insights and views, such as household makeup or whether someone in that household uses one of your competitors. It can show whether subscribers with certain behaviors, such as long commutes, frequent travels, or those who reside in specific areas where your service is not as strong, tend to churn at a higher rate. Third-party data can also help you understand age, gender, and affluence, helping you identify both your best subscribers and those more likely to churn. And most importantly, third-party data helps you understand the competitive landscape and movement, both to competitive providers and even across other technology solutions, such as mobile broadband or fixed wireless access.

So, while your first-party data is incredibly valuable, it needs the context and a much bigger picture to identify churn threats more accurately and successfully. Improving your machine learning modeling to predict churn requires robust third-party data sets that build this more holistic picture of the consumer. So now, with some ideas about what types of data can help you predict churn, in our next episode, we'll talk about putting that data to work.