Predicting Customer Relationship Using Transactional Behavior for a B2B Telecommunications Company

Introduction

For a leading telecommunications provider, customer experience was a critical driver of business success. The company had a well-established CX strategy that included both transactional surveys (sent after a completed service interaction) and relationship surveys (sent to measure overall customer sentiment toward the brand). However, there was a fundamental flaw in the existing methodology—Transactional NPS (TNPS) was being used to measure single transactions, even though brand loyalty is built over time.

TNPS asked customers if they would recommend the brand based on a single transaction, which did not accurately reflect customer satisfaction with the customer journey as a whole. The company partnered with Numr CXM to redefine how transactional experience was measured, ensure more meaningful insights, and develop a predictive model that could forecast future NPS trends based on transactional behavior data.

Callout:
"A single transaction does not define brand loyalty—customer effort plays a bigger role."

The Challenge

Despite having a structured customer experience measurement system, the company faced three key challenges that hindered its ability to drive customer retention and improve customer lifetime value (CLTV):

1. Transactional NPS Was Not an Effective Measure of Single-Transaction Experiences
   
   
   • NPS is designed to measure brand advocacy, not the ease or effort involved in completing a single interaction.
   • Customers might be loyal to the brand but still have difficulty with a specific transaction, leading to skewed CX data analytics.
2. Survey Length and Engagement Issues
   
   
   • The transactional surveys included TNPS, CSAT, and Customer Effort Score, making them lengthy and leading to survey fatigue.
   • Response rates declined, reducing the reliability of customer experience metrics.
3. Lack of Connection Between Transactional Experience and Long-Term Loyalty
   
   
   • There was no established model linking single-transaction experiences to long-term NPS trends.
   • Without a clear connection between transactional feedback and future relationship NPS, it was difficult to optimize the customer journey for better outcomes.

Callout:
"TNPS wasn’t telling the full story—customers needed an easier way to express their experience."

‍

The Turning Point: Why NES Was the Right Metric

Numr CXM conducted an in-depth analysis to determine which customer experience metric was best suited for transactional surveys. The team applied machine learning algorithms to compare how well each of the three metrics—TNPS, CSAT, and NES—correlated with service drivers.

• Regression models were applied to two years of survey data, testing how well each metric explained service-related customer feedback.
• NES (Net Easy Score) emerged as the best predictor, as it showed the strongest correlation with customer behavior across all four business areas:
  
  • Sales
  • Delivery
  • Network Operations
  • Customer Relationship Management (CRM)

The results confirmed that ease of transaction had a greater impact on immediate customer satisfaction than brand advocacy. This made NES the ideal metric for transactional surveys, replacing TNPS.

Callout:
"Customers value ease and efficiency—NES proved to be the best measure of transaction quality."

The Solution

Numr CXM designed a comprehensive CX strategy to enhance customer satisfaction and predict future NPS trends based on transactional behavior:

1. Transitioning to NES for Transactional Surveys
   
   
   
   • Replaced TNPS with NES across all business areas to measure ease of transaction.
   • This simplified surveys, increased response rates, and ensured more accurate CX data analytics.
2. Establishing the Link Between NES and Relationship NPS
   
   
   
   • Created a predictive model that analyzed how NES trends influenced long-term NPS outcomes.
   • Mapped transactional behaviors to relationship survey results to uncover which service areas had the greatest impact on customer retention.
3. Developing a Predictive Model for Future NPS Trends
   
   
   
   • Transactional survey responses were used to predict which service areas had the strongest influence on NPS trends.
   • Query resolution had the highest impact, while sales experience had the least influence on short-term NPS changes.

Callout:
"Understanding the connection between NES and NPS allowed for proactive customer retention strategies."

Implementation: From Insights to Execution

Numr CXM executed the transition from TNPS to NES and developed the NPS prediction model in structured phases:

1. Analyzing 13 Quarters of Historic Data
   
   
   • Since previous surveys contained ease-of-transaction questions, Numr was able to extract NES trends from historical data.
   • This allowed for a seamless transition without losing past insights.
2. Determining the Impact of NES on NPS
   
   
   • Regression analysis measured the impact of different service areas on overall NPS trends.
   • The company discovered that query resolution had the strongest correlation to NPS, while sales interactions had a longer lag time.
3. Establishing the Time Lag Between Transactional Experience and Relationship NPS
   
   
   • Some service areas had an immediate impact on NPS, while others took longer.
   • Customer support issues impacted NPS within one quarter, whereas sales experiences influenced NPS after four quarters.
4. Building a Predictive NPS Model
   
   
   • A weighted NES formula was developed to forecast future relationship NPS based on transactional behavior trends.
   • A linear regression model was used to predict NPS one quarter in advance, allowing the company to take proactive action to improve customer satisfaction.

Callout:
"Instead of reacting to NPS trends, the telecommunications provider could now predict and prevent churn."

Results

The initiative delivered significant improvements in customer experience measurement and customer retention:

1. Higher Survey Engagement – NES simplified the survey process, leading to higher response rates and better data quality.
2. More Accurate NPS Predictions – The predictive model allowed the company to forecast NPS trends one quarter in advance.
3. Increased Customer Retention & CLTV – By addressing issues before they impacted NPS, the company improved customer lifetime value (CLTV) and reduced customer acquisition costs (CAC).

Callout:
"Switching to NES and predictive analytics made customer experience insights more actionable than ever."

Conclusion

By replacing TNPS with NES, the telecommunications provider modernized its CX strategy and made transactional feedback more meaningful. Through advanced CX data analytics, the company established a direct correlation between transactional experiences and long-term loyalty, allowing it to predict and improve future NPS trends.

With NES-powered insights, the company achieved measurable ROI of CX, improved customer satisfaction, and reduced churn and operational inefficiencies.

Callout:
"With predictive analytics, the telecommunications provider moved from measuring NPS to proactively shaping it."