Extracting "broken expectations" from call center records

Abstract

Currently, despite the explicit industrial consideration to improve the appeal and usability of technically sound electronics products, users increasingly seem to have dissatisfactory experiences in interacting with them. These unforeseen experiences (attributable to specifications omissions, usability/learnability problems, or specific usage context) lead to a large and increasing share of unknown field complaints. To correct and prevent such complaints or user reports, we promote effective exploitation of call centers: Valuable usage data is retrievable from the field by adopting a usercentered failure classification model, which we developed. We also report on the supporting results of a test from applying our model to a set of call center data.

“Please Connect Me to a Specialist”: Scrutinising ‘Recipient Design’ in Interaction with an Artificial Conversational Agent

“…In institutional, task-oriented settings, pleasurable customer and user experiences are of utmost importance for continued service or product use and customer and user retention (e.g. [28,43,46]). In this paper, we focused on the first step in interaction with an artificial agent in a telephone enquiry service; the formulation of an information enquiry and its repair after the operator's reply.…”

In situ evaluation of recommender systems: Framework and instrumentation

“…Despite the various ways of getting this kind of information from users in the user-centered engineering domain, it is notoriously known to be challenging to systematically extract and quantify user feedback, and even more so to scale this process up within industrial settings where many users, many products, and many complaints or feedback co-exist (Koca and Brombacher, 2008a;Koca et al, 2007). An important goal we aim at with the use of user-initiated feedback forms, or the so-called ''Thumbs Up'' and ''Thumbs Down'' surveys, is to systematically elicit meaningful subjective information from users as to their satisfaction and dissatisfaction levels and causes related to the use of the product that may affect its longterm use and adoption.…”

Grasping product pragmatics

“…The model we developed (a.k.a. designer-centered feedback analysis and classification model) and use for these purposes are described in [18,19]. Additionally, quantitative data may provide information regarding users' momentary affect [6,9], as well as about users' longer term perceptual and evaluative judgments [14].…”

“…by depicting perceptions of the product at different use phases as different instances). In addition, quality analysis can be conducted on the related qualitative survey results, by viewing the product's fault-failure 2 versus success distribution proportions over the categories of our designer-centered feedback classification model [18,19]. Such analysis enables to systematically view the userperceived strong and weak points of the current product concept, and as such enables (i) visualizing how those compare to other products', (ii) analyzing the relative importance of certain failures or successes, and hence prioritizing, and (iii) suggesting relevant actionable items for diverse information stakeholders to work on.…”

“…Failures of good fit, or socalled soft failures, i.e. where a user has problems with a technically sound product, are indeed becoming increasingly common [7,23,4,17,18,20].…”