Project Overview
CS Associates is a consulting firm in the customer services industry. They conducted a banking customer satisfaction survey to understand the key reasons customers might stop using their banks — and needed a comprehensive, interactive report that their existing survey tool simply couldn't produce.
The deliverable was a fully interactive Power BI dashboard that transformed raw survey responses into ranked churn driver analysis, demographic segmentation, and communication channel preferences — giving stakeholders the ability to slice and explore the data themselves rather than receiving a static snapshot.
Central finding: Security concerns are the dominant churn driver — 56% of respondents range from "Likely" to "Most Likely" to leave over security issues. Poor customer service and process inefficiencies follow closely, while a surprising 73% unemployed respondent base signals a significant untapped product opportunity.
Client Challenges
CS Associates faced two specific limitations with their existing survey tool that prevented them from extracting meaningful intelligence:
Non-Interactive Demographics Dashboard
The built-in demographic dashboard was static — CS Associates couldn't filter, drill down, or cross-reference demographic segments against specific survey responses. Insights were locked behind a non-interactive interface.
No Ranking Analysis on Core Survey Question
The tool provided no meaningful report on the primary survey question: "Rank the possible reasons you may stop using your bank." This ranked, multi-response question required custom data modeling that the tool couldn't handle.
Methodology
The solution required two Power BI capabilities working together: Power Query for transforming the raw survey data into an analyzable structure, and Data Modeling (One-to-Many) to connect the transformed tables for cross-filtering in the dashboard.
Power Query Transformation Steps
Headers: Promoted first row to column headers to create a properly structured table.
ResponseID: Added an index column renamed to ResponseID — the unique key linking respondent demographics to their survey answers.
Responses query: Created a new query called "Responses" isolating ResponseID and the ranking column — separating answer data from demographic data for clean modeling.
Split by delimiter: The multi-value ranking column was split by comma delimiter, transforming it from a single column of concatenated values into multiple discrete columns.
Unpivot: The resulting split columns were unpivoted — converting from wide format to long format with "Trigger" and "Ranking" columns. This is the critical step that enables ranked analysis across all responses.
Scale column: A conditional column "Scale" was added mapping numeric rankings (1–5) to human-readable labels: Most Likely, Highly Likely, Likely, Fairly Likely, Least Likely.
Data modeling: A one-to-many relationship was established between the Demographics table (one ResponseID) and the Responses table (many ranked answers per respondent), enabling cross-filtering between demographics and churn drivers.
Key technical decision: Unpivoting the ranking columns was the pivotal transformation. Without it, the five ranking options would remain separate columns with no way to aggregate or compare across triggers. The unpivot creates a normalized structure that Power BI's visuals can slice by any dimension.
Data Model
The final data model uses a one-to-many relationship between two tables — a clean, performant structure that enables every demographic filter to instantly update all churn driver visuals:
- Demographics table (one side): One row per respondent — ResponseID, age, gender, employment status, communication channel preferences
- Responses table (many side): Multiple rows per respondent — one row for each ranked churn trigger, with Trigger, Ranking (1–5), and Scale (Most Likely → Least Likely) columns
- Relationship key: ResponseID links both tables, enabling any demographic filter to cascade into the ranking analysis
Why this matters for the client: Before this model, CS Associates could only see demographics separately from opinions. With the one-to-many model, they can now ask questions like "Among unemployed male respondents aged 25–40, what is the top churn trigger?" — and get an answer instantly.
Live Dashboard
The interactive Power BI dashboard below lets you explore all survey findings. Use the filters to segment by age group, gender, employment status, or communication channel:
Insight 1: Demographics
The survey's demographic profile reveals both the bank's core customer base and a significant underserved segment:
The 73% unemployed figure is the most striking demographic finding. While this age group (25–40) represents the prime working years, the survey population skews heavily toward non-traditional income earners. This signals either a product-market fit issue for employed customers — or a genuine opportunity to build financial products specifically for freelancers, entrepreneurs, and informal workers who are currently underserved by standard banking products.
Product opportunity: No-fee accounts, micro-loan products, and financial literacy programs specifically designed for non-traditional income earners could convert this underserved 73% into engaged, loyal customers.
Insight 2: Security Concerns
Security is the most pervasive churn driver — the broadest distribution of "Likely or above" responses of any trigger in the survey:
Security Concerns — Likelihood to Leave
With 56.11% of respondents ranging from "Likely" to "Most Likely" to leave over security concerns, this is not a communication problem — it's a trust problem. Customers are not confident the bank is protecting their money and data. Addressing this requires both genuine security improvements and transparent, proactive communication about what those improvements are.
Insight 3: Poor Customer Service
Customer service quality is the second-most cited churn trigger — and uniquely concentrated at the "Highly Likely" level:
Poor Customer Service — Likelihood to Leave
35.88% cite poor customer service as "Highly Likely" to make them leave — the highest concentration of any single likelihood category across all churn triggers. This signals that service quality failures are felt as urgent, not just inconvenient. AI-driven support, faster response times, and better-trained front-line staff are the most direct interventions.
Insight 4: ATM Cash-Outs & the Digital Shift
A striking 72.14% of respondents say they are "Least Likely" to leave over ATM cash-out issues — meaning ATM availability is barely a concern. This is strong evidence that this customer base has already transitioned toward cashless and digital banking methods.
Rather than investing in ATM infrastructure, the data points toward expanding digital wallet services and incentivizing mobile banking usage. The customers have already made the behavioral shift — the bank's product roadmap needs to catch up.
Strategic implication: Resources allocated to ATM maintenance and expansion should be evaluated against the evidence that 72% of surveyed customers don't consider ATM access a significant factor in their banking relationship. Digital infrastructure investment has a stronger ROI signal.
Insight 5: Transaction Delays in Branches
26.72% of respondents cite in-branch transaction delays as "Most Likely" to drive them to leave — a significant operational pain point that is both measurable and fixable.
Long wait times in branches represent a solvable operational problem. Appointment scheduling systems, digital queue management, and streamlined teller processes are proven interventions. For a population that has already demonstrated digital readiness (see ATM finding), promoting self-service digital alternatives for common in-branch transactions could simultaneously reduce branch congestion and improve customer satisfaction.
Quick win: Introducing appointment booking for branch visits — even via WhatsApp or a simple web form — could immediately address the "Most Likely" churn risk from 26.72% of respondents with minimal development investment.
Insight 6: Concerns Over Bad Processes
30.15% of respondents are concerned about inefficient banking processes — the most common moderate frustration in the survey. Unlike dramatic failures (app crashing, transaction failing), process inefficiency is the slow erosion of trust: unnecessarily complex forms, unclear procedures, redundant verification steps, or unclear escalation paths.
Process reengineering — auditing the end-to-end customer journey for unnecessary friction and streamlining the most frequently used workflows — addresses this at the root. The goal is to make banking feel intuitive and frictionless, not bureaucratic and opaque.
Insight 7: Communication Channel Preferences
Channel usage varies significantly by age group — revealing a digital divide that requires different engagement strategies for different customer segments:
| Age Group | Branch | Internet Banking | Mobile App | Chatbot | |
|---|---|---|---|---|---|
| 25–40 | High | High | High | High | Medium |
| 41–56 | High | Medium | Medium | Medium | Low |
| 57+ | Medium | Low | Low | Low | Minimal |
The 25–40 group is channel-agnostic and digitally comfortable — they use everything. The 57+ group shows minimal digital channel engagement across all categories, indicating either access barriers, digital literacy gaps, or preference for human interaction. The 41–56 middle tier is the targeted intervention opportunity: they use branches heavily but have modest digital adoption — suggesting they're reachable through a combination of in-branch digital demonstrations and targeted mobile app onboarding campaigns.
Recommendations
Product Development — Serve the Unemployed 73%
Develop financial products explicitly designed for non-traditional income earners: no-fee accounts, micro-loan products with flexible repayment, savings plans with low minimums, and financial literacy programs. This demographic represents 73% of the survey base and is currently underserved by standard banking products.
Customer Support — Address the 36% Highly Likely to Leave
Implement AI-driven customer support (chatbots for common queries, ML-based ticket routing) to reduce response times and improve first-contact resolution. 35.88% of respondents rate poor service as "Highly Likely" to trigger churn — faster, more consistent support is the most direct intervention.
Digital Adoption — Meet Customers Where They Already Are
72% of respondents have already moved away from ATM dependency. Promote cashless and mobile banking solutions actively, particularly for the 57+ group through in-branch demonstrations and guided onboarding. Redirect ATM investment to mobile and internet banking infrastructure improvements.
Operational Efficiency — Reduce the 27% "Most Likely" Branch Delay Churn
Introduce branch appointment booking, digital queue management, and self-service kiosks for routine transactions. For a digitally ready population, promoting app-based alternatives to common in-branch tasks (document submission, account queries) reduces congestion while serving customers on their preferred channel.
Security Messaging — Convert 56% "Likely to Leave" into Trust
Security concerns affect over half the respondent base. Beyond genuine security improvements, implement proactive communication campaigns: regular security update emails, in-app security tips, transparency reports on fraud prevention performance, and two-factor authentication promotion. Trust is built through visibility, not just through actual security quality.