A research summary of Laurence Vervaet’s master’s thesis at Ghent University

Customer churn is one of the most pressing challenges in B2B sales and account management. While much of the academic literature focuses on predictive models and data-driven churn analytics, what do key account managers (KAMs) actually do in practice to detect when a customer is about to leave?

In her master’s dissertation at Ghent University, Laurence Vervaet explores this question through in-depth interviews with 15 experienced sales professionals. The result is a rich, qualitative study that sheds light on how churn signals are perceived, interpreted, and acted upon in real-world B2B environments.

📘 Source: Vervaet, L. (2022). Detection of Customer Churn Signals: How Key Account Managers Detect Signals of Churn in Practice. Ghent University. Available via lib.ugent.be

Why This Research Matters

In B2B markets, losing a key account can be devastating. According to the Pareto principle, 80% of revenue often comes from just 20% of customers. Replacing a lost key account is not only expensive—it’s risky and time-consuming.

While churn prediction models and AI-driven analytics have gained traction in recent years, Vervaet’s research reveals that most KAMs still rely heavily on relational and intuitive methods to detect churn. This raises important questions about the gap between academic theory and business practice—and how SaaS providers can help bridge it.

Research Objectives

The study focused on three core questions:

1. What churn reasons do KAMs perceive within their client portfolio?
2. What signals signify potential churn?
3. What methods do KAMs use to detect these signals?

To answer these, Vervaet conducted 15 semi-structured interviews with KAMs and sales leaders across various industries in Belgium, including telecom, IT services, HR, and automotive.

Key Findings

1. Churn Detection Is Still Largely Relational

Despite the availability of CRM systems and churn prediction tools, most KAMs rely on direct communication, gut feeling, and informal networks to detect churn. They build strong personal relationships with multiple stakeholders in the client organization and use these connections to sense dissatisfaction or shifting priorities.

“It’s still people working with people. You detect churn by talking to them, not by looking at a dashboard.” — Respondent

2. Churn Signals Are Often Subtle and Contextual

The study identified two main categories of churn signals:

• Relational signals: complaints, reduced contact, changes in tone, unpaid invoices, or reluctance to renew contracts.
• Competitive signals: clients referencing competitor offerings, attending competitor events, or issuing tenders tailored to competitors.

These signals are often subtle and subjective, requiring KAMs to interpret them based on experience and context.

3. CRM and Surveys Are Underused or Fragmented

While some organizations use CRM systems and satisfaction surveys, these tools are often fragmented, underutilized, or not trusted by KAMs. Many respondents expressed skepticism about the accuracy of survey data or the effort required to maintain CRM records.

“A system is only as valuable as the data you put in.” — Respondent

4. Prediction Models and Uplift Analytics Are Rarely Used

Surprisingly, none of the respondents reported using churn prediction models or uplift modeling in their daily work. This is despite the fact that many of them work in sectors—like telecom and IT—where such models are common in academic literature.

This suggests a disconnect between research and practice, and a potential opportunity for SaaS providers to offer more accessible, actionable churn analytics.

5. Environmental and Organizational Changes Are Key Triggers

KAMs pay close attention to changes in the client’s organization—such as new leadership, mergers, or budget shifts—as well as macroeconomic factors like regulation or market disruption. These changes often precede churn and are picked up through news monitoring, social media, and informal conversations.

Practical Implications for B2B Sales Teams

Vervaet’s research offers several takeaways for sales leaders and SaaS providers:

• Invest in relationship-building: Strong interpersonal networks remain the most effective churn detection tool.
• Enable better internal communication: Signals are often detected by different people across the organization. CRM systems should facilitate—not hinder—this flow of information.
• Support intuitive detection with smart tools: Tools like sentiment analysis, LinkedIn Sales Navigator, and AI-powered alerts can complement human intuition.
• Train KAMs to interpret signals: Not all complaints or price objections are signs of churn. Experience and context matter.
• Bridge the gap between data and action: SaaS providers should focus on making churn analytics more transparent, interpretable, and actionable for frontline sales teams.

Final Thoughts

Laurence Vervaet’s thesis is a valuable contribution to the conversation around customer retention in B2B markets. It reminds us that while data and AI are powerful, human relationships and judgment still play a central role in detecting and preventing churn.

For SaaS companies, this presents both a challenge and an opportunity: how can we design tools that support—not replace—the nuanced, relational work of key account managers?

📘 Read the full thesis: Ghent University Library via this link