IBM Telco Customer Churn
2026
Telecom Customer Churn & Retention Analysis
Identifying churn patterns, high-risk customer segments, and targeted retention opportunities using BigQuery, SQL, and Looker Studio.
ASK
Business Problem
A telecommunications company is experiencing customer churn, creating recurring revenue loss and increasing the need for targeted retention strategies. The objective of this analysis was to identify the customer characteristics most associated with churn, determine which segments face the highest churn risk, and translate those findings into actionable retention opportunities. This matches the business problem and questions we established at the beginning of the project.
Key Business Questions
What is the overall churn and retention performance?
How do contract type and customer tenure relate to churn?
Which services and payment methods are associated with higher churn?
How do monthly charges differ between churned and retained customers?
Which combination of characteristics identifies the highest-priority retention segment?
PREPARE
Data & Tools
The analysis used the IBM Telco Customer Churn sample dataset, containing 7,043 customers and 21 original attributes covering demographics, account information, services, contracts, billing, payment behavior, and churn status. The dataset was imported into Google BigQuery for preparation and SQL analysis.
Data Source: IBM Telco Customer Churn
Records: 7,043 customers
Original Attributes: 21
Target Variable: Churn
Environment: Google BigQuery / Google Cloud
PROCESS
Data Cleaning & Preparation
Before analysis, I validated the dataset for record integrity, missing values, blank fields, duplicate customer IDs, and data-type issues. The initial validation confirmed 7,043 unique customers with no duplicate customer IDs or missing values in the primary analysis fields. A secondary check identified 11 blank TotalCharges records, corresponding to customers with zero tenure.
An analysis-ready table, customer_churn_clean, was then created in BigQuery. TotalCharges was safely converted to a numeric field, and additional analytical variables were created:
TenureGroup • ChurnFlag • RetainedFlag • MonthlyRevenueAtRisk
These transformations allowed churn, retention, customer tenure, and revenue exposure to be analyzed consistently.
ANALYZE
SQL Analysis & Key Findings
SQL was used to establish executive KPIs and investigate churn across contract type, tenure, internet service, support/security services, monthly charges, payment method, and multidimensional customer segments.
Overall Churn — 26.54%
1,869 of 7,043 customers churned, while 5,174 were retained. Approximately $139.1K in monthly charges was associated with churned customers.
Contract Type — 42.71%
Month-to-month customers had substantially higher churn than one-year (11.27%) and two-year customers (2.83%). PROJECT 3 DATA ANALYTICS
Customer Tenure — 47.44%
Customers within their first 12 months showed the highest churn rate. Churn progressively decreased across longer-tenure groups, reaching only 9.51% among customers with 49+ months of tenure.
Internet Service — 41.89%
Fiber-optic customers showed substantially higher churn than DSL customers (18.96%) and customers without internet service (7.40%).
Support & Security — 48.96%
Among internet customers, those with neither Online Security nor Tech Support showed the highest churn rate, while customers with both services had substantially lower churn.
Payment Method — 45.29%
Electronic-check customers had the highest churn rate, compared with mailed check and automatic payment methods.
SHARE
Analytical Report
I developed a tabular, column-profiling analytical report in Looker Studio to communicate both the underlying customer data and the major churn patterns.
The report combines a customer data profile, cleaned-record preview, and comparative churn analysis across contract type, tenure, internet service, support/security, and payment method. This makes the analysis traceable from the underlying data through to the final business conclusion.
ACT
Priority Segment & Recommendations
The final synthesis combined contract, tenure, internet service, payment method, Online Security, and Tech Support to identify customer groups where retention efforts could have the greatest business relevance. The segmentation query deliberately focused on groups containing at least 50 customers rather than prioritizing extremely small segments.
Priority Retention Segment
Month-to-month • 0–12 months tenure • Fiber optic • Electronic check • No Online Security • No Tech Support
544 customers | 74.45% churn | 405 churned | ~$33.3K monthly revenue at risk
This segment combines high churn with meaningful customer volume and revenue exposure, making it a stronger retention priority than simply selecting the smallest segment with the highest percentage churn. Your documentation confirms the 405 of 544 customers and ~$33.27K monthly revenue exposure.
Recommended Action
Target this segment with a focused retention pilot during the first year of the customer relationship. Test incentives for longer-term contracts and automatic payments, bundle Online Security and Tech Support where appropriate, and investigate the fiber-optic customer experience. Measure the pilot using churn reduction and retained monthly revenue before scaling the intervention.
Final Outcome / Impact
The analysis transformed raw telecom customer data into an actionable retention strategy. Rather than treating all churned customers equally, it identified a specific high-priority segment where retention resources could be concentrated and quantified the monthly revenue associated with that risk.
VIEW DASHBOARD →
VIEW GITHUB REPOSITORY →
