Telecom Customer Segmentation and churn Prediction with Messaging Gateway using AI
Customer churn is one of the most critical issues faced by the telecommunications industry. Customer churn prediction is currently the main mechanism employed by the industry in order to prevent customers from churning. The objective of churn prediction is to identify customers that are going to lea
2025-06-28 16:29:41 - Adil Khan
Telecom Customer Segmentation and churn Prediction with Messaging Gateway using AI
Project Area of Specialization Artificial IntelligenceProject SummaryCustomer churn is one of the most critical issues faced by the telecommunications industry. Customer churn prediction is currently the main mechanism employed by the industry in order to prevent customers from churning. The objective of churn prediction is to identify customers that are going to leave the telecommunications service provider in advance. Customer churn prediction would allow the telecommunications service provider to plan their customer retention strategy. The high volume of data generated by the industry, with the help of data mining techniques implementation, becomes the main asset for predicting customer churn. Due to this reason, recent advancement in Artificial Intelligence and Machine Learning has made it easy to create such predictors with high complexity and huge amount of data.
Pakistan has one of the largest cellular service user bases in the region with total number of users approaching 154 million and growing at the rate of 10% per annum, almost 80% of the population is a consumer of cellular services. The industry was able to gross a total revenue of 446 million dollar in 2015 which was 6% more than the previous years (PTA, 2015).
Telecom industry in Pakistan is also going rapid transformation, Telenor has catch up with Mobilink in term of markets share, Zong has emerged as the 3rdlargest and fastest growing network in the country, Ufone has lost its pace and Warid has stumbled further and is face with ever worsening circumstances as the company has lost almost half of its customers base in the past 5 years to competitors due to consumer churn. (PTA, 2015).
Keeping in view the significance of customer retention, the damage caused by attrition and current situation and trends, it is eminent that the industry is approaching its maturity stage, hence generic growth rate will drop and the only way to increase the customer base will be to entice and attract competition’s customers. Considering these circumstances I have choose to investigate the factors which contribute to churn rate.
The Conclusion from different studies shows that Network Quality, Coverage, Rates and spam Messages were a major reason which contributed to churn whereas network effect, Value added services and international roaming were vital in some individual cases.
Previous studies shows that in Pakistan’s Telecom industry there has been a great battle between some major Giants and they are coming up with new offers and promotions every day.
Project Objectives1. Concentrate on your loyal customers.
Comparatively focusing on getting new customers, it could be even more constructive to gather your resources into your trustworthy, current customers.
2. Investigate churn as it exists.
Analyze your churned customers as a means of concluding why customers are unsubscribing or stop using your services. Analyze when and why churn occurs in a customer's lifespan with your organization, and make that data valuable to put into place appropriate volumes.
3. Show that you care for your customers.
Rather than waiting to associate with your customers until they collaborate with you, seek a forethoughtful way. Contact them with all the dividends you provide and show them you care about them, and they'll be implicit to abide.
Project Implementation MethodAs the demand for customer churn predictors are progressively increasing in almost every industry, especially in telecommunications, so by using state of the art classification and clustering algorithms we will be able to develop an AI model that will predict and give us the details of the customers which are willing to churn in the near future, and when the details are fetched they will be fed into our messaging system will have some predefined offers which will be selected with best interest of the customer and then an SMS offer will be sent to all the customers based on their interest.
Benefits of the ProjectThe key benefits are:
- Rather than waiting to associate with your customers until they collaborate with you, seek a forethoughtful way. Contact them with all the dividends you provide and show them you care about them, and they'll be implicit to abide.
- Analyze your churned customers as a means of concluding why customers are unsubscribing or stop using your services. Analyze when and why churn occurs in a customer's lifespan with your organization, and make that data valuable to put into place appropriate volumes.
- Comparatively focusing on getting new customers, it could be even more constructive to gather your resources into your trustworthy, current customers
- AI Model that predicts customer churns in telecom CRM data
- Pairing Algorithm that matches the predicted churners with best matched offer
- SMS Gateway which will send that offer to the predicted churners.
First, we will study about the literature and feasibility of this project then we will use the best possible way to get data and all the requirements and start our software part. We will start by doing exploratory data analysis and creating a pipeline for our data which will insure that data is well cleaned and all the values are according to their data types and according to standards. Then after performing all the analysis we will perform SMOTE analysis if required and or if our data is imbalanced. Then we will work on creating our AI model we will try both techniques classification and clustering and if required we will do semi-supervised machine learning techniques to create our model. We will try different algorithms such as Neural Networks, Random Forest, Gradient Boost, etc. from classification and K-Means, DB-Scan, Affinity-Propagation, etc. from Clustering. We can then use Voting technique to see which algorithm did the best. After that our model will give us the list with details of those customers who are willing to churn in the near future so we will feed this list to our pairing algorithm which already has some pre-fed offers which will be paired with them using AI based pairing algorithm. After that we will send those offers to those customers through our Arduino-GSM based SMS Gateway.
Final Deliverable of the Project HW/SW integrated systemCore Industry TelecommunicationOther Industries IT Core Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Industry, Innovation and Infrastructure, Responsible Consumption and ProductionRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 16000 | |||
| GSM Module sim900a | Equipment | 1 | 6000 | 6000 |
| Arduino UNO SMD | Equipment | 1 | 800 | 800 |
| Connecting wires | Equipment | 40 | 5 | 200 |
| Printing | Miscellaneous | 4 | 2000 | 8000 |
| Hardware Power Sysytem | Equipment | 1 | 1000 | 1000 |