Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Silent Abandonment in Text-Based Contact Centers: Identifying, Quantifying, and Mitigating its Operational Impacts

Antonio Castellanos, Galit B. Yom-Tov, Yair Goldberg, Jaeyoung Park

TL;DR

The study investigates silent abandonment in text-based contact centers, where customers leave while waiting without signaling abandonment, creating uncertainty about service quality and resource use. The authors develop two complementary tools: (i) classification models that quantify the scope of silent abandonment using in-queue texts and metadata, and (ii) an Expectation-Maximization (EM) algorithm that accurately estimates customer patience under left- and right-censoring caused by Sab and uSab, with extensions to include covariates such as words written in queue. The XYZ dataset reveals Sab accounts for a substantial share of abandonment (71.3% of abandoned conversations in the detailed sample) and imposes measurable efficiency and cost penalties ($5,457 per agent annually). Extensions show that permitting in-queue writing triples patience and reduces service time, suggesting that design choices like enabling write-in-queue and adaptive concurrency can significantly mitigate Sab’s impact. The findings offer actionable strategies, including predictive targeting of Sab, updated concurrency rules, and potential use of AI bots to handle suspected Sab conversations, while highlighting the need for online experiments to validate causal effects.

Abstract

In the quest to improve services, companies offer customers the option to interact with agents via texting. Such contact centers face unique challenges compared to traditional call centers, as measuring customer experience proxies like abandonment and patience involves uncertainty. A key source of this uncertainty is silent abandonment, where customers leave without notifying the system, wasting agent time and leaving their status unclear. Silent abandonment also obscures whether a customer was served or left. Our goals are to measure the magnitude of silent abandonment and mitigate its effects. Classification models show that 3%-70% of customers across 17 companies abandon silently. In one study, 71.3% of abandoning customers did so silently, reducing agent efficiency by 3.2% and system capacity by 15.3%, incurring $5,457 in annual costs per agent. We develop an expectation-maximization (EM) algorithm to estimate customer patience under uncertainty and identify influencing covariates. We find that companies should use classification models to estimate abandonment scope and our EM algorithm to assess patience. We suggest strategies to operationally mitigate the impact of silent abandonment by predicting suspected silent-abandonment behavior or changing service design. Specifically, we show that while allowing customers to write while waiting in the queue creates a missing data challenge, it also significantly increases patience and reduces service time, leading to reduced abandonment and lower staffing requirements.

Silent Abandonment in Text-Based Contact Centers: Identifying, Quantifying, and Mitigating its Operational Impacts

TL;DR

The study investigates silent abandonment in text-based contact centers, where customers leave while waiting without signaling abandonment, creating uncertainty about service quality and resource use. The authors develop two complementary tools: (i) classification models that quantify the scope of silent abandonment using in-queue texts and metadata, and (ii) an Expectation-Maximization (EM) algorithm that accurately estimates customer patience under left- and right-censoring caused by Sab and uSab, with extensions to include covariates such as words written in queue. The XYZ dataset reveals Sab accounts for a substantial share of abandonment (71.3% of abandoned conversations in the detailed sample) and imposes measurable efficiency and cost penalties ($5,457 per agent annually). Extensions show that permitting in-queue writing triples patience and reduces service time, suggesting that design choices like enabling write-in-queue and adaptive concurrency can significantly mitigate Sab’s impact. The findings offer actionable strategies, including predictive targeting of Sab, updated concurrency rules, and potential use of AI bots to handle suspected Sab conversations, while highlighting the need for online experiments to validate causal effects.

Abstract

In the quest to improve services, companies offer customers the option to interact with agents via texting. Such contact centers face unique challenges compared to traditional call centers, as measuring customer experience proxies like abandonment and patience involves uncertainty. A key source of this uncertainty is silent abandonment, where customers leave without notifying the system, wasting agent time and leaving their status unclear. Silent abandonment also obscures whether a customer was served or left. Our goals are to measure the magnitude of silent abandonment and mitigate its effects. Classification models show that 3%-70% of customers across 17 companies abandon silently. In one study, 71.3% of abandoning customers did so silently, reducing agent efficiency by 3.2% and system capacity by 15.3%, incurring $5,457 in annual costs per agent. We develop an expectation-maximization (EM) algorithm to estimate customer patience under uncertainty and identify influencing covariates. We find that companies should use classification models to estimate abandonment scope and our EM algorithm to assess patience. We suggest strategies to operationally mitigate the impact of silent abandonment by predicting suspected silent-abandonment behavior or changing service design. Specifically, we show that while allowing customers to write while waiting in the queue creates a missing data challenge, it also significantly increases patience and reduces service time, leading to reduced abandonment and lower staffing requirements.
Paper Structure (38 sections, 2 theorems, 30 equations, 14 figures, 10 tables, 2 algorithms)

This paper contains 38 sections, 2 theorems, 30 equations, 14 figures, 10 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Paper Structure and Summary of Contributions
  3. Literature Review
  4. Context
  5. Scope of Silent Abandonment in Different Contexts
  6. Estimating Customer Patience
  7. The Scope of Silent Abandonment in Contact Centers and Its Impact on Efficiency
  8. The Service Process and the Data Describing It
  9. Estimating the Scope of Silent Abandonment
  10. Discussion: Reducing the Operational Impact of Silent-Abandonment Customers
  11. Estimating Customer Patience with Silent Abandonment
  12. The EM Algorithm: Model Assumption and Formulation
  13. Customer Classes with Complete Data.
  14. Customer Classes with Missing Data.
  15. The EM Algorithm Formulation.
  16. ...and 23 more sections

Key Result

Theorem 1

Under Assumption assumption:1, $\widehat{C_{1,t}^{i}}$, $\widehat{C_{2,t}^{i}}$, and $\widehat{C_{3,t}^{i}}$ are given by

Figures (14)

  • Figure 1: Example Conversation and the Metadata Process Flow of the Four Types of Customers in Contact Centers
  • Figure 2: Proportion of Known and Uncertain Silent Abandonment in Different Contact Centers
  • Figure 3: ROC Curve on Test Dataset
  • Figure 4: Patience as a Function of the Number of Words Written during Waiting in Queue. XYZ Dataset. (Confidence Intervals are Computed Using Bootstrapping.)
  • Figure EC.1: Conversation Start Day and Hour by Customer Type (Percentages)
  • ...and 9 more figures

Theorems & Definitions (2)

  • Theorem 1
  • Theorem 2