arXiv subjects

Antonio Castellanos

Publications and source records attributed to Antonio Castellanos.

6 recordsLinked to original sources

arXiv2025

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

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.

cs.SI↗

arXiv2024

Silent Abandonment in Contact Centers: Estimating Customer Patience from Uncertain Data

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.

cs.SI↗

arXiv2023

The Co-Production of Service: Modeling Services in Contact Centers Using Hawkes Processes

In customer support contact centers, every service interaction involves a messaging dialogue between a customer and an agent; together, they exchange information, solve problems, and collectively co-produce the service. Because the service progression is shaped by the history of conversation so far, we propose a bivariate, marked Hawkes process cluster model of the customer-agent interaction. To evaluate our stochastic model of service, we apply it to an industry contact center dataset containing nearly 5 million messages. Through both a novel residual analysis comparison and several Monte Carlo goodness-of-fit tests, we show that the Hawkes cluster model indeed captures dynamics at the heart of the service and also surpasses classic models that do not incorporate the service history. Furthermore, in an entirely data-driven simulation, we demonstrate how this history-dependent model can be leveraged operationally to inform a prediction-based routing policy. We show that widely-used and well-studied customer routing policies can be outperformed with simple modifications according to the Hawkes model. Through analysis of a stylized model proposed in the contact center literature, we prove that service heterogeneity can cause this underperformance and, moreover, that such heterogeneity will occur if service closures are not carefully managed.

cs.SI↗

arXiv2008

Computer simulation of model cohesive powders: Plastic consolidation, structural changes and elasticity under isotropic loads

The quasistatic behavior of a simple 2D model of a cohesive powder under isotropic loads is investigated by Discrete Element simulations. The loose packing states, as studied in a previous paper, undergo important structural changes under growing confining pressure P, while solid fraction Φirreversibly increases by large amounts. The system state goes through three stages, with different forms of the plastic consolidation curve Φ(P*), under growing reduced pressure P* = Pa/F0, defined with adhesion force F0 and grain diameter a. In the low-confinement regime (I), plastic compaction is negligible, and the structure is influenced by the assembling process. The following stage (regime II) is independent of initial conditions. The void ratio varies linearly with log P, as described in the engineering literature. In the last stage of compaction (III), a maximum solid fraction is approached, and properties of cohesionless granular packs are retrieved. Under consolidation, while the fractal range of density correlations decreases, force patterns reorganize, and elastic moduli increase by a large factor. Plastic deformation events correspond to very small changes in the network topology, while the denser regions tend to move like rigid bodies. Elastic properties are dominated by the bending of thin junctions in loose systems. For growing RR those tend to reduce to particle chains, the folding of which, rather than tensile ruptures, controls plastic compaction.

cond-mat.mtrl-sci↗

arXiv2008

Effect of the difference in ion mobilities on traveling-wave electro-osmosis

We analyze the AC electro-osmotic motion of a 1:1 aqueous solution, taking into account the difference in mobilities and diffusion coefficients between positive and negative ions. This model serves to understand the behavior of common systems as a solution of NaCl in water. We pay special attention to two cases. First, the case of slightly different mobilities, that can model a KCl solution. Second, the case of a strongly asymmetric solution, with an almost vanishing mobility, applicable to the case of a salt where the negative ion is much more massive than the positive one. For all the cases, we perform the mathematical description and linear analysis of the problem, in order to establish the dependence of the induced velocity with the frequency, wavelength and amplitude of the applied voltage.

physics.flu-dyn↗

arXiv2007

Computer simulation of model cohesive powders: influence of assembling procedure and contact laws on low consolidation states

The structure and mechanical properties of a simple two-dimensional model of a cohesive powder are investigated by molecular dynamics simulations. Micromechanical ingredients involve elasticity, friction, a short range attraction and, possibly, rolling resistance (RR) in contacts. The microstructure of the cohesive packing varies according to the assembling procedure, from rather densely packed if isolated particles are directly compressed to much looser if the formation of large aggregates is allowed prior to compression. A crucial parameter is the ratio P*= Pa/F0 of applied pressure P, acting on grains of diameter $a$, to maximum tensile contact force F0. At low P* the final structure depends on the level of velocity fluctuations at the early stages of cluster aggregation. With RR the coordination number approaches 2 in the limit of low initial velocities or large rolling friction. The force network generally comprises small hyperstatic clusters with forces of the order of F0, joined by nearly unstressed, barely rigid arms. As P* grows, it quickly rearranges into force chain-like patterns. Density correlations witness a fractal structure, with dimension Df, up to some density-dependent blob size. WIth RR Df coincides with the ballistic aggregation result, despite a possibly different connectivity. Possible effects of some parameters on material strength are evoked.

cond-mat.dis-nn↗