[Paper Review] Silent Abandonment in Contact Centers: Estimating Customer Patience from Uncertain Data

Antonio Castellanos, Johan S. H. van Leeuwaarden+1 more|arXiv (Cornell University)|Apr 23, 2023

Advanced Queuing Theory AnalysisBusiness, Management and Accounting3 citations

TL;DR

This paper proposes a novel method to estimate customer patience in contact centers by identifying silent abandonment—where customers leave without signaling—using text analysis and an EM algorithm. It finds that 30%–67% of abandonments are silent, reducing system efficiency by 5%–15%, and shows that accounting for silent abandonment significantly improves queueing model accuracy.

ABSTRACT

In the quest to improve services, companies offer customers the opportunity to interact with agents through contact centers, where the communication is mainly text-based. This has become one of the favorite channels of communication with companies in recent years. However, contact centers face operational challenges, since the measurement of common proxies for customer experience, such as knowledge of whether customers have abandoned the queue and their willingness to wait for service (patience), are subject to information uncertainty. We focus this research on the impact of a main source of such uncertainty: silent abandonment by customers. These customers leave the system while waiting for a reply to their inquiry, but give no indication of doing so, such as closing the mobile app of the interaction. As a result, the system is unaware that they have left and waste agent time and capacity until this fact is realized. In this paper, we show that 30%-67% of the abandoning customers abandon the system silently, and that such customer behavior reduces system efficiency by 5%-15%. To do so, we develop methodologies to identify silent-abandonment customers in two types of contact centers: chat and messaging systems. We first use text analysis and an SVM model to estimate the actual abandonment level. We then use a parametric estimator and develop an expectation-maximization algorithm to estimate customer patience accurately, as customer patience is an important parameter for fitting queueing models to the data. We show how accounting for silent abandonment in a queueing model improves dramatically the estimation accuracy of key measures of performance. Finally, we suggest strategies to operationally cope with the phenomenon of silent abandonment.

Motivation & Objective

• To address information uncertainty in contact centers caused by silent abandonment, where customers leave without signaling their departure.
• To estimate customer patience accurately despite missing and censored data due to silent abandonment.
• To improve the accuracy of queueing models by incorporating silent abandonment into performance estimation.
• To propose operational strategies that mitigate capacity loss and inefficiency caused by silent abandonment.
• To compare customer patience across different contact center channels (chat vs. messaging) and analyze behavioral differences.

Proposed method

• Use text analysis and a Support Vector Machine (SVM) model to classify whether a customer has abandoned the queue based on message content and patterns.
• Develop a parametric estimator for customer patience that accounts for both observed and unobserved (silent) abandonments.
• Apply an Expectation-Maximization (EM) algorithm to estimate customer patience distribution under missing and censored data conditions.
• Construct a modified queueing model that incorporates silent abandonment to better reflect real system dynamics.
• Validate model performance by comparing fit accuracy between models with and without silent abandonment.
• Use survival analysis techniques to handle censored data on abandonment times.

Experimental results

Research questions

• RQ1 What proportion of customer abandonments in contact centers are silent, and how does this vary across different service channels?
• RQ2 How does silent abandonment affect system efficiency, and what is its measurable impact on agent utilization and capacity loss?
• RQ3 To what extent does ignoring silent abandonment bias the estimation of customer patience in queueing models?
• RQ4 How do customer patience levels differ between chat and messaging systems, and what factors may explain these differences?
• RQ5 What operational strategies can reduce the negative impact of silent abandonment on service performance and agent workload?

Key findings

• Between 30% and 67% of abandoning customers in contact centers silently abandon the queue without signaling their departure.
• Silent abandonment reduces system efficiency by 5% to 15%, primarily due to agent idleness and wasted work on non-existent customers.
• The EM algorithm estimated customer patience at 81.1 minutes in messaging systems and only 2 minutes in chat systems, reflecting strong channel-dependent differences.
• Queueing models that account for silent abandonment show dramatically improved fit to real data compared to models that ignore it.
• The percentage of silent abandonments is higher in messaging systems (where wait times are longer), suggesting a correlation between wait duration and silent abandonment.
• Silent abandonment leads to inaccurate queue length measurements, which can mislead delay announcements and other real-time service metrics.