Traditional Business Intelligence systems have reached their limit
Gone are those days when telecom companies used to have an easy ride and make money with little effort. Most markets are now becoming highly saturated and telecom companies today, are currently suffering from huge competition .A Telecom Company spends averagely above $100 to acquire a new subscriber. It is a huge loss when those subscribers leave the franchise to other competitors within a couple of months of their acquisition. The customer lifetime value is bound to drop significantly, as competitors engage in price wars, smart and attractive Value Added Services (VAS) are offered, Apps using innovative Voice-over-IP technology are being developed.
It costs nearly 50 times less to keep an existing telecom subscriber than to acquire a new one. Meaning there is an urgent need for telecom decision makers to enhance their CRM efforts in order to focus on retaining their “hard earned” subscribers.
Traditional Business Intelligence systems are not well equipped to fully understand and profile the modern telecom customer, and so telecom companies relying on traditional Business Intelligence systems are losing money to their competitors who have fully embraced the new order of the fourth industrial revolution, where cloud, automation and data science is the key for any first mover in any industry.
As introduced in the previous article “introducing the application of data science per industry” the data we are producing is doubling every 2 years and so we are already sitting on a huge oil reserve,waiting to be explored . And this is where data science fits in perfectly-the perfect mining engineer for the perfect mining job. This is because Data Science is about modelling this massive, rich and freely available data, in order to come up with unprecedented insights about the customer’s wants, needs and preferences. The current telecom decision maker needs to understand that there is a more useful information that could be gotten from the company’s current warehouse data and even more so from external data sources (social media, smartphone data, CCTV footage, website analytics, etc) using techniques out of the reach of the traditional business intelligence and business analysts skills set . Data scientists are well trained to run models on both structured and unstructured data to come up with very valuable insight on what customers TRULY WANT. And with such insights , marketing action plans can easily be drawn and implemented.
A sample Data Scientist’s Workflow to reduce churn in a telecom company
A data scientist will typically follow the steps below to come up with propositions on how to reduce churn.
- Collecting the data:
- The data scientist will create a data base which aggregates data from as multiple sources.
- Cleaning the data
- About 90% of the time will be spent here, extracting key features/ words from the Social media and website data, removing empty cells and finally matching and merging all these data to come up with one big customer data base with hundreds of variables describing a customer.
- Exploratory Data Analysis:
- Descriptive statistics and Visualization will be applied to discover hidden patterns and interesting correlations, which is gives an idea on what to expect from the model in the next step below.
- Modelling and predicting churn
- First,a training set is used to train the model on the key features which determine who will easily become churn and a test set is used test predictions made by this newly trained model on and compare with actual churn figures to establish the accuracy of the model.
- Conclusions and Recommendations
- The conclusions would vary greatly.This is because it depends on the reality of each company,which is reflected in the realities discovered in the data provided.
|Possible Model Results
|1) 90% of subscribers churn after contacting call center more than 2 times in a month
|– Call center software show number of calls made by a customer since the beginning of the month, so that the agent will be aware more cautious when talking to customers who have called more than twice in the month.
|2) 80% of customers churn after buying a specific offer
|– The offer quota should be reviewed or even removed from the catalog
|3) Customers churn more when someone in his circle of friend on Facebook churns
|– Influences online should be identified and monitored more closely in order to anticipate and prevent them from churning.
|4) Customers spend much time on the website before clicking on the “airtime recharge” link
|– Website should be optimized to make the “airtime recharge” link more readily accessible.
Practical advice to telecom decision makers:
- Accept that data science is very essential for modern businesses
- Create a data / information culture and organigram
- Improve on Data warehouse and ERP systems to collect lots of data about the customers (what pack the customer subscribed to, at what time, at which location, with which sales agent,and linked to which device,etc)
- Recruit a talented data scientist with the right mix of :Programming, Statistics and telecom domain knowledge
- Work on applying the recommendations coming from the results of the churn prediction model
- It it works fine,the model should be integrated in the live system, so it becomes the ways of working of your telecom company.