Maurits Kaptein

Personalizing behavior change technologies

The Personalization in Behavior Change Technologies workshop will focus on how to design, build and study persuasive technologies to adapt to meet the individualized needs of target users. The goal of this workshop is to connect the diverse group of behavior change researchers and practitioners interested in personalization to share their experiences, ideas, and discuss how to move the field forward. We will identify the key challenges in this area and brainstorm solutions to tackle these issues. Discussion and ideas generated from this workshop will be archived online to be available to the larger research community. This workshop ties into a number of special interests for the CHI community, including health, sustainability, intelligent user interfaces, serious games, and persuasive technology.

Extending the Similarity-Attraction Effect: The Effects of When-Similarity in Computer-Mediated Communication

The feeling of connectedness experienced in computer-mediated relationships can be explained by the similarity-attraction effect SAE. Though SAE is well established in psychology, the effects of some types of similarity have not yet been explored. In 2 studies, we demonstrate similarity-attraction based on the timing of activities-when-similarity. We describe a novel experimental paradigm for manifesting when-similarity while controlling for the activities being performed what-similarity. Study 1 N = 24 shows when-similarity attraction in the evaluation of connectedness with others. Study 2 N = 42 identifies an interaction between who-similarity-similarity in personal backgrounds-and when-similarity. Both studies show that real-time computer-mediated interaction can lead to greater feelings of connectedness between people when there is an opportunity to discover when-similarity.

Nice to Know You: Familiarity and Influence in Social Networks

Advertisers on Social Network Sites often use recommendations by others in a users networks to endorse products. While these familiar others are hypothesized to be more effective in influencing users than unfamiliar others, there is a catch: familiarity does not necessarily ensure similarity to the familiar person, a potential problem because the combination of familiarity and dissimilarity has been hypothesized to lead to lowered compliance. In an experiment (N = 44), we test peoples compliance to similar and dissimilar familiar others in an online environment: we show that in both cases, familiarity leads to increased compliance. The work highlights the importance of familiarity on influence and suggests that gaining familiarity even in situations of dissimilarity is effective.

The use of Thompson sampling to increase estimation precision.

In this article, we consider a sequential sampling scheme for efficient estimation of the difference between the means of two independent treatments when the population variances are unequal across groups. The sampling scheme proposed is based on a solution to bandit problems called Thompson sampling. While this approach is most often used to maximize the cumulative payoff over competing treatments, we show that the same method can also be used to balance exploration and exploitation when the aim of the experimenter is to efficiently increase estimation precision. We introduce this novel design optimization method and, by simulation, show its effectiveness.

Tracking the decoy : Maximizing the decoy effect through sequential experimentation

The decoy effect is one of the best known human biases violating rational choice theory. According to a large body of literature, people may be persuaded to switch from one offer to another by the presence of a third option (the decoy) that, rationally, should have no influence on the decision-making process. For example, when asked to choose between a laptop with a good battery but a poor memory and a laptop with a poor battery but a good memory, customers may be induced to shift their preference if the offer is accompanied by a third laptop that has a battery as good as the latter but even worse memory—an effect that has clear applications in marketing practice. Surprisingly, renowned decoy studies have resisted replication, inducing scholars to challenge the scientific validity of the phenomenon and question its practical relevance. Using a treatment allocation scheme that takes inspiration from the lock-in amplification schemes used in experimental physics, we were able to explore the entire range of decoy attribute values and demonstrate that some of the reproducibility issues reported in the literature result from a suboptimal initial conditions. Furthermore, we demonstrate that our approach is able to sequentially identify the features of the decoy that maximize choice reversal. We thus reinstate the scientific validity and practical relevance of the decoy effect and demonstrate the use of lock-in amplification to optimize treatments.

Real-Time Adaptation of Influence Strategies in Online Selling

Real-time adjustments in online selling are becoming increasingly common. In this paper we describe a novel method of real-time adaptation, and introduce influence strategies as a useful level of analysis for personalization of online selling. The proposed method incorporates three perspectives on real-time adaptation: the content of the appeal (influence strategies), the context in which the optimization is performed (online selling), and the computational method (a Beta-Binomial model in combination with Randomized Probability Matching). We argue that these three perspectives are in constant interplay in any attempt to dynamically optimize online selling outcomes using personalization. Dynamic learning, adaptation and personalization of influence strategies represents are concluded to be prerequisites for e-selling - using the psychology of personal selling interactions in online marketing.

Customizing persuasive messages; the value of operative measures

This paper aims to examine whether estimates of psychological traits obtained using meta-judgmental measures (as commonly present in customer relationship management database systems) or operative measures are most useful in predicting customer behavior.,Using an online experiment (N = 283), the study collects meta-judgmental and operative measures of customers. Subsequently, it compares the out-of-sample prediction error of responses to persuasive messages.,The study shows that operative measures – derived directly from measures of customer behavior – are more informative than meta-judgmental measures.,Using interactive media, it is possible to actively elicit operative measures. This study shows that practitioners seeking to customize their marketing communication should focus on obtaining such psychographic observations.,While currently both meta-judgmental measures and operative measures are used for customization in interactive marketing, this study directly compares their utility for the prediction of future responses to persuasive messages.

Automated adaptive selling

This paper aims to develop and test a method of automating, for online retailers, the practice of adaptive selling, which is typically used by salespeople in face-to-face interactions. This method customizes persuasive messages for individual customers as they navigate a retailer’s website.,This paper demonstrates a method for the online implementation of automated adaptive selling using sales influence tactics. Automated adaptive selling is compared to nonadaptive selling in three e-commerce field studies.,The results reveal that adaptive selling is more effective than nonadaptive selling. The click-through rates increased significantly when adaptive selling was used.,This paper highlights the effectiveness of existing theories concerning adaptive human-to-human selling and their utility to online selling. The authors demonstrate the added value of adaptive selling in e-commerce, thereby opening up a novel area of research into adaptive selling online. While the paper focuses on the adjustment of sales influence tactics, other factors could be investigated for adjustment in future research (e.g. prices).,The methods, described in detail, are readily available for implementation by online retailers. The implementations are timely and increasingly valuable as e-commerce expands into interpersonal channels (e.g. instant messengers and social media).,To the authors’ knowledge, this paper is the first to formally implement automated adaptive selling as described in the ISTEA model in an e-commerce setting.

Personalized Product Recommendations: Evidence from the Field

Targeting personalized product recommendations to individual customers has become a mainstream activity in online stores as it has been shown to increase click-through rate and sales. However, as personalization becomes increasingly commonplace, customers may feel personalized content intrusive and therefore not responding or even avoiding them. Many studies have investigated advertising intrusiveness and avoidance but a research gap on the effect of degree of personalization on customer responses based on field evidence exists. In this paper, 27,175 recommendation displays from five different online stores are analyzed. The results show that the further the customer is in the purchasing process, the more effective personalization is if it is based on information about the present rather than past browsing session. Moreover, recommendations in passive form are more effective than recommendations in active form suggesting the need to dispel the perception of intrusiveness.

Uncovering noisy social signals: Using optimization methods from experimental physics to study social phenomena

Due to the ubiquitous presence of treatment heterogeneity, measurement error, and contextual confounders, numerous social phenomena are hard to study. Precise control of treatment variables and possible confounders is often key to the success of studies in the social sciences, yet often proves out of the realm of control of the experimenter. To amend this situation we propose a novel approach coined “lock-in feedback” which is based on a method that is routinely used in high-precision physics experiments to extract small signals out of a noisy environment. Here, we adapt the method to noisy social signals in multiple dimensions and evaluate it by studying an inherently noisy topic: the perception of (subjective) beauty. We show that the lock-in feedback approach allows one to select optimal treatment levels despite the presence of considerable noise. Furthermore, through the introduction of an external contextual shock we demonstrate that we can find relationships between noisy variables that were hitherto unknown. We therefore argue that lock-in methods may provide a valuable addition to the social scientist’s experimental toolbox and we explicitly discuss a number of future applications.