Search

In this research, we formulate budget allocation decisions as an optimal control problem using a generalized Vidale–Wolfe model (GVW) as its advertising dynamics under a finite time horizon. One key element of our modeling work is that the proposed optimal budget allocation model (called GVW-OB) takes into account the roles of two useful indexes of the GVW model representing the advertising elasticity and the word-of-mouth (WoM) effect, respectively, in determining optimal budget. Moreover, we discuss desirable properties and provide a feasible solution to our GVW-OB model. We conduct computational experiments to assess our model’s performance and its identified properties, based on real-world datasets obtained from advertising campaigns by three e-commerce companies on Google AdWords, Facebook Ads and Baidu Ads, respectively. Experimental results show that (1) our GVW-OB strategy outperforms four baselines in terms of both payoff and ROI in either concave or S-shaped settings; (2) linear budget allocation strategies favor concave advertising responses, while nonlinear strategies support S-shaped responses; (3) a larger ad elasticity empowers higher levels of optimal budget and corresponding market share and thus achieves higher payoff and ROI, so does a larger WoM effect; and (4) as the total budget increases, the resulting payoff by the GVW-OB strategy increases monotonically, but the ROI decreases, which is consistent with the law of diminishing marginal utility. From a methodological perspective, our GVW-OB strategy provides a feasible solution for advertisers to make optimal budget allocation over time, which can be easily applied to a variety of advertising media. The identified properties and experimental findings of this research illuminate critical managerial insights for advertisers and media providers.

We carry out our study on interactions via second screens by analyzing more than 3M, 800K and 50K Super Bowl related posts from Twitter, Instagram, and Tumblr respectively to measure the influence of second screen based interaction on social media attitude. We took commercials, musicals and game as the categories of interest into consideration for temporal analysis in Pre, During, and Post Super Bowl phases. Research results show that the change in attitude is significant among categories in phases. Posts containing URLs and posts that contain only texts with no URLs, retweets or response show a positive association with viewers’ second screen attitude in Pre and Post phases. This research is important in identifying the interplay that technology has on the use of social interactions for sharing feelings and information via cross screens across disparate social platforms at different phases of an in real life event.

In sponsored search advertising, keywords serve as an essential bridge linking advertisers, search users and search engines. Advertisers have to deal with a series of keyword decisions throughout the entire lifecycle of search advertising campaigns. This paper proposes a multi-level and closed-form computational framework for keyword optimization (MKOF) to support various keyword decisions. Based on this framework, we develop corresponding optimization strategies for keyword targeting, keyword assignment and keyword grouping at different levels (e.g., market, campaign and adgroup). With two real-world datasets obtained from past search advertising campaigns, we conduct computational experiments to evaluate our keyword optimization framework and instantiated strategies. Experimental results show that our method can approach the optimal solution in a steady way, and it outperforms two baseline keyword strategies commonly used in practice. The proposed MKOF framework also provides a valid experimental environment to implement and assess various keyword strategies in sponsored search advertising., Comment: 21 pages, 3 figures,1 table

User needs inform designers and developers of essential functionalities for requirements engineering. In this work, we summarize key concepts and challenges relating to manual and automatic user needs detection methods. We discuss six challenges with manual and eight challenges with automated methods. Despite the promise of automated methods, the challenges imply that artificial intelligence and machine learning are not yet mature enough to replace manual methods, such as interviews and focus groups, for discovering user needs in requirements engineering.

Controlling the quality of social media feeds poses an issue for many users. Platforms such as Twitter give users some options to influence their feeds. Still, the selection of content predominantly relies on implicit rather than explicit user actions, as manual options for "cleaning the feed" are often cumbersome and difficult to use for most users. Here, we present Take Back Control, a web browser extension that gives users control to hide undesirable content from their social media feeds. The extension combines JavaScript (for hiding the content) and machine learning (for deciding what content to hide). Our current demonstration includes three filter types: Toxic, Political, and Negative content, with a possibility to add more filters, all of this with the overarching aim of helping end users control the information visible in their social media feeds.

We compare a data-driven persona system and an analytics system for efficiency and effectiveness for a user identification task. Findings from the 34-participant experiment show that the data-driven persona system affords faster task completion, is easier for users to engage with, and provides better user identification accuracy. Eye-tracking data indicates that the participants focus most of their attention on the persona content while focusing more on navigation features when using the analytics system. The combined results provide empirical support for the use of data-driven personas for a user identification task, which we surmise to be a result of the persona system following a user-centered design paradigm instead of an information-centered paradigm. That analytics system afforded capabilities and insights that the persona system did not suggest that the triangulation of features may lead to a better overall user understanding.

As complex data becomes the norm, greater understanding of machine learning (ML) applications is needed for content marketers. Unstructured data, scattered across platforms in multiple forms, impedes performance and user experience. Automated classification offers a solution to this. We compare three state-of-the-art ML techniques for multilabel classification - Random Forest, K-Nearest Neighbor, and Neural Network - to automatically tag and classify online news articles. Neural Network performs the best, yielding an F1 Score of 70% and provides satisfactory cross-platform applicability on the same organisation's YouTube content. The developed model can automatically label 99.6% of the unlabelled website and 96.1% of the unlabelled YouTube content. Thus, we contribute to marketing literature via comparative evaluation of ML models for multilabel content classification, and cross-channel validation for a different type of content. Results suggest that organisations may optimise ML to auto-tag content across various platforms, opening avenues for aggregated analyses of content performance.

Investigating users’ engagement with interactive persona systems can yield crucial insights for the design of such systems. Using eye-tracking, researchers can address the scarcity of behavioral user studies, even during times when physical user studies are difficult or impossible to carry out. In this research, we implement a webcam-based eye-tracking module into an interactive persona system, facilitating remote user studies. Findings from the implementation can show what information users pay attention to in persona profiles.

Observing user interactions with interactive persona systems offers important insights for the design and application of such systems. Using an interactive persona system, user behavior and interaction with personas can be tracked with high precision, addressing the scarcity of behavioral persona user studies. In this research, we introduce and evaluate an implementation of persona analytics based on mouse tracking, which offers researchers new possibilities for conducting persona user studies, especially during times when in-person user studies are challenging to carry out.

During exceptional times when researchers do not have physical access to users of technology, the importance of remote user studies increases. We provide recommendations based on lessons learned from conducting online user studies utilizing four online research platforms (Appen, MTurk, Prolific, and Upwork). Our recommendations aim to help those inexperienced with online user studies. They are also beneficial for those interested in increasing their proficiency, employing this increasingly important research methodology for studying people’s interactions with technology and information.

Practitioners in user-centric industries have increasingly recognized the applicability of personas. However, the methods used to create personas in different domains remain inconsistent and unsystematic. We analyzed 51 studies focused on designing personas for professional purposes and found the practice most prevalent in the user experience design, healthcare, market research, and social media strategy domains. Within these domains, user experience design personas are characterized by their focus on user activity goals, health personas on medical patients’ physical symptoms, market research personas on customers’ lifestyles, and social media strategy personas on interactions within and between online communities. We identify and compare the elements in the personas. Based on these, we provide guidelines for professionals interested in developing personas for understanding barriers to positive user experience, recruiting users, and building online communities, including how to represent persona details related to lifestyle and health, contexts of product usage, and scaling of online data. Practitioners in user-centric industries have increasingly recognized the applicability of personas. However, the methods used to create personas in different domains remain inconsistent and unsystematic. We analyzed 51 studies focused on designing personas for professional purposes and found the practice most prevalent in the user experience design, healthcare, market research, and social media strategy domains. Within these domains, user experience design personas are characterized by their focus on user activity goals, health personas on medical patients’ physical symptoms, market research personas on customers’ lifestyles, and social media strategy personas on interactions within and between online communities. We identify and compare the elements in the personas. Based on these, we provide guidelines for professionals interested in developing personas for understanding barriers to positive user experience, recruiting users, and building online communities, including how to represent persona details related to lifestyle and health, contexts of product usage, and scaling of online data.

The competitiveness of the video game market has increased the need for understanding players. We generate player personas from survey data of 15,402 players' 195,158 stated game preferences from 130,495 game titles using the methodology of automatic persona generation. Our purpose is to demonstrate the potential of data-driven personas for segmenting players by their game preferences. The resulting prototype personas provide potential value for game marketing purposes, e.g., targeting gamers with social media advertising, although they can also be used for understanding demographic variation among various game preference patterns.

Artificial generation of facial images is increasingly popular, with machine learning achieving photo-realistic results. Yet, there is a concern that the generated images might not fairly represent all demographic groups. We use a state-of-the-art method to generate 10,000 facial images and find that the generated images are skewed towards young people, especially white women. We provide recommendations to reduce demographic bias in artificial image generation.

Quantitative methods are becoming more common for persona creation, but it is not clear to which extent online data and opaque machine learning algorithms introduce bias at various steps of data-driven persona creation (DDPC) and if these methods violate user rights. In this conceptual analysis, we use Gillespie's framework of algorithmic ethics to analyze DDPC for ethical considerations. We propose five design questions for evaluating the ethics of DDPC. DDPC should demonstrate the diversity of the user base but represent the actual data, be accompanied by explanations of their creation, and mitigate the possibility of unfair decisions.

We address a recommendation task for next likely flight destination to customers of a major international airline company. We compare performance using historical flight data and an actual user evaluation. Using two years of historical flight data consisting of tens of millions of flights, an ensemble and a collaborative filtering approach obtained an accuracy of 47% and 20% using a test set of 100,000 customers, respectively, highlighting the challenge of the domain. We then evaluated our recommendations on 10,000 actual customers, with a 45-45-10 split among ensemble, collaborative filtering, and control group. The overall predictive power employed with real users was 23%, with the ensemble method having a predictive power of 19% and 30% for collaborative filtering. Results indicate that, in complex and shifting domains such as this one, one cannot rely solely on historical data for evaluating the impact of user recommendations. We discuss implications for recommendation systems and future research in this and related domains.

Personas are a well-known technique in human computer interaction. However, there is a lack of rigorous empirical research evaluating personas relative to other methods. In this 34-participant experiment, we compare a persona system and an analytics system, both using identical user data, for efficiency and effectiveness for a user identification task. Results show that personas afford faster task completion than the analytics system, as well as outperforming analytics with significantly higher user identification accuracy. Qualitative analysis of think-aloud transcripts shows that personas have other benefits regarding learnability and consistency. However, the analytics system affords insights and capabilities that personas cannot due to inherent design differences. Findings support the use of personas to learn about users, empirically confirming some of the stated benefits in the literature, while also highlighting the limitations of personas that may necessitate the use of accompanying methods.

Quantitative persona creation (QPC) has tremendous potential, as HCI researchers and practitioners can leverage user data from online analytics and digital media platforms to better understand their users and customers. However, there is a lack of a systematic overview of the QPC methods and progress made, with no standard methodology or known best practices. To address this gap, we review 49 QPC research articles from 2005 to 2019. Results indicate three stages of QPC research: Emergence, Diversification, and Sophistication. Sharing resources, such as datasets, code, and algorithms, is crucial to achieving the next stage (Maturity). For practitioners, we provide guiding questions for assessing QPC readiness in organizations.

Despite the dedication that some volunteer moderators of online communities display when performing their moderation duties, they become targets of hate and harassment by other users. To understand what causes the change in moderator role from heroes to victims, we analyze the responses of 1,818 moderators on Reddit to an online survey about moderation practices and harassment. We built a statistical model and found 6 significant independent variables that affect harassment on moderators, such as the knowledge of community norms, which increases harassment on moderators the most. Our findings imply that vulnerable moderators in toxic communities need countermeasures against harassment.

Understanding the causes or triggers of toxicity adds a new dimension to the prevention of toxic behavior in online discussions. In this research, we define toxicity triggers in online discussions as a non-toxic comment that lead to toxic replies. Then, we build a neural network-based prediction model for toxicity trigger. The prediction model incorporates text-based features and derived features from previous studies that pertain to shifts in sentiment, topic flow, and discussion context. Our findings show that triggers of toxicity contain identifiable features and that incorporating shift features with the discussion context can be detected with a ROC-AUC score of 0.87. We discuss implications for online communities and also possible further analysis of online toxicity and its root causes.

In an experiment, we present 38 marketing and data analysts professionals with two online AI-driven persona interfaces, one using numbers and the other using text. We employ eye tracking, think-aloud, and a post-engagement survey for data collection to measure perception and visual engagement with the personas along 7 constructs. Results show that the use of numbers has a mixed effect on the perceptions and visual engagement of the persona profile, with job role as a determining factor on whether numbers/text affect end users for 2 of the constructs. The use of numbers has a significant positive effect on user perceptions of usefulness by analysts but a significantly negative effect on user perceptions of completeness for both marketers and analysts. The use of numbers decreases the perceived completeness of the personas for both marketer and analysts. This research has both theoretical and practical consequences for AI-driven persona development and their interface design, suggesting that the inclusion of numbers can have a desirable effect for certain roles but with possible negative effects on user perceptions.

Creating personas from large amounts of online data is useful but difficult with manual methods. To address this difficulty, we present Automatic Persona Generation (APG), which is an implementation of a methodology for quantitatively generating data-driven personas from online social media data. APG is functional, and it is deployed with several organizations in multiple industry verticals. APG employs a scalable web front-end user interface and robust back-end database framework processing tens of millions of user interactions with tens of thousands of online digital products across multiple online platforms, including Facebook, Google Analytics, and YouTube. APG identifies audience segments that are both distinct and impactful for an organization to create persona profiles. APG enhances numerical social media data with relevant human attributes, such as names, photos, topics, etc. Here, we discuss the architecture development and central system features. Overall, APG can benefit organizations distributing content via online platforms or with online content that relates to commercial products. APG is unique in its algorithmic approach to processing social media data for customer insights. APG can be found online at https://persona.qcri.org.

Following the proliferation of personified big data and data science algorithms, data-driven user personas (DDPs) are becoming more common in persona design. However, the DDP templates are seemingly diverse and fragmented, prompting a need for a synthesis of the information included in these personas. Analyzing 31 templates for DDPs, we find that DDPs vary greatly by their information richness, as the most informative layout has more than 300% more information categories than the least informative layout. We also find that graphical complexity and information richness do not necessarily correlate. Furthermore, the chosen persona development method may carry over to the information presentation, with quantitative data typically presented as scores, metrics, or tables and qualitative data as text-rich narratives. We did not find one “general template” for DDPs and defining this is difficult due to the variety of the outputs of different methods as well as different information needs of the persona users.

Algorithmic fairness criteria for machine learning models are gathering widespread research interest. They are also relevant in the context of data-driven personas that rely on online user data and opaque algorithmic processes. Overall, while technology provides lucrative opportunities for the persona design practice, several ethical concerns need to be addressed to adhere to ethical standards and to achieve end user trust. In this research, we outline the key ethical concerns in data-driven persona generation and provide design implications to overcome these ethical concerns. Good practices of data-driven persona development include (a) creating personas also from outliers (not only majority groups), (b) using data to demonstrate diversity within a persona, (c) explaining the methods and their limitations as a form of transparency, and (d) triangulating the persona information to increase truthfulness.

Customers increasingly rely on reviews for product information. However, the usefulness of online reviews is impeded by fake reviews that give an untruthful picture of product quality. Therefore, detection of fake reviews is needed. Unfortunately, so far, automatic detection has only had partial success in this challenging task. In this research, we address the creation and detection of fake reviews. First, we experiment with two language models, ULMFiT and GPT-2, to generate fake product reviews based on an Amazon e-commerce dataset. Using the better model, GPT-2, we create a dataset for a classification task of fake review detection. We show that a machine classifier can accomplish this goal near-perfectly, whereas human raters exhibit significantly lower accuracy and agreement than the tested algorithms. The model was also effective on detected human generated fake reviews. The results imply that, while fake review detection is challenging for humans, “machines can fight machines” in the task of detecting fake reviews. Our findings have implications for consumer protection, defense of firms from unfair competition, and responsibility of review platforms.

We analyze more than 3,000,000 Twitter, 800,000 Instagram, and 50,000 Tumblr posts concerning a single major in-real-life event, Super Bowl XLIX, to determine attitude. We consider three event phases ( Pre , During , and Post ). Findings show link-based recommendations and undirected broadcast patterns positively correlate with attitude in the Pre and Post phases, respectively. The usage of these specific features highlights the differing information needs of viewers during these phases, specifically the sharing of information in the Pre phase and the sharing of opinions in the Post phase. The volume of postings indicates a negative attitude for all social media platforms, demonstrating that adverse information is more likely to be shared than positive information; this finding contradicts prior findings. This second screens phenomenon research is important in identifying the sharing information on multiple social media platforms during an in real life event.

In this critique, we conceptually examine the use of personas in an age of availability of large-scale online analytics data. Based on the criticism and benefits outlined in prior work, we formulate the major arguments for and against the use of personas, analyze these arguments, and demonstrate areas for the productive employment of personas by leveraging digital analytics data in their creation. From our review of the prior literature and the given availability of online customer data, our key tenet is that personas are located between aggregated and individual statistics. At their best, personas capture the coverage of the customer base attributed to aggregated data representations while retaining the interpretability of individual-level analytics. Persona creation benefits from both novel computational techniques and data sources. To demonstrate this, we propose and implement automatically generated personas primarily based on quantitative data. We also review key persona validation issues and examine how these issues can be addressed with automated persona generation using real user data from online analytics platforms. Finally, we outline areas of future research in the persona domain within the field of digital marketing and advertising. Overall, to survive in the rapidly developing marketing industry and online customer analytics, personas must evolve by adopting new practices. There are implications for this evolution of personas in a variety of domains, including design, content creation, and digital marketing.

Personified big data and rapidly developing data science techniques enable previously unforeseen methodological developments for longitudinal analysis of online audiences. Applying data-driven persona generation on online customer statistics from a real organizational social media channel, we demonstrate how personas can be deployed to understand online customer patterns over time. We conduct 32 monthly rounds of data collection of customer demographics and content consumption patterns on the YouTube channel of a major publishing organization posting thousands of items of content and then algorithmically generate 15 personas monthly. We analyze the data-driven persona for changes monthly, yearly, and lifetime (period). Results show an average 40% change in the personas, and 78% of the personas experience more change than consistency for topic interests. The implications are that organizations frequently publishing online content should employ automatic data collection and periodic persona creation to ensure their customer understanding is current. For this, algorithmic data-driven systems that leverage methods for persona creation are recommended.

We develop a framework to reduce the number of customer segments to the smallest quantity without losing essential information of the underlying population in the electronic marketplace. As a use case of this approach, we create personas for these segments to enhance customer understanding. We use (a) matrix factorization to identify customer behaviors and construct customer segments, (b) statistical heuristics to collapse into meaningful segments, and (c) automation to enrich by generating a persona profile for each segment. We evaluate our approach in a case study using more than 21 million online flight bookings of a major airline company resulting in a 57.5% decrease from 1194 to 507 segments, thereby reducing segment noise. Three customer mega-segments emerge: Behaviorally same – Demographically different, Behaviorally different - Demographically same, and Behaviorally and Demographically different. As one of the first efforts at the essential task of customer segmentation reduction for large customer populations, findings have implications for organizations desiring to employ segmentation and/or personas for enhanced customer understanding.

Purpose This research is based on the premise that current metrics for search engine advertising (SEA) are misleading and do not sufficiently allow managers to evaluate traffic and conversions simultaneously. This study aimed to conceptually develop and assess conversion potential (CvP) as a unifying construct for both measuring and evaluating the performance of SEA campaigns. Design/methodology/approach A data set of nearly seven million records covering almost three years of a multi-million-dollar keyword marketing campaign from a major US retailer was used to validate the construct of CvP. Findings Results empirically validate how CvP measures both campaign traffic and sales in SEA, using the optimization factor of ad rank, which is one of many possible factors. Research limitations/implications Although the data set is large and covers a lengthy period of time, it is limited to one company in the retail sector. Practical implications The research instantiates CvP as a metric for overall SEA account performance while demonstrating that it is a practical tool for future campaign planning. The metric simultaneously incorporates a sales ratio and a traffic ratio. Originality/value This is the first study to formalize and provide a working definition of CvP in the academic literature. The contribution is a theoretical and practical managerial framework to mutually evaluate, measure and make decisions about SEA efforts.

The umbrella term "computational advertising" encompasses a spectrum of computational systems, technologies, and methods of advertising and promotional behaviors and decision-making activities. The guest editors of this special issue on computational advertising discuss the field in general and the highlighted articles in particular.

Evaluate the effectiveness of question routing systems in the social Q&A process.Find that individuals are more willing to share their knowledge under question routing context whereas less connected.Build an effective model to automatically identify active knowledge sharers from non-shares using non-Q&A features from four dimensions: profile, posting behavior, language style, and social activities. The increasing volume of questions posted on social question and answering sites has triggered the development of question routing services. Most of these routing algorithms are able to recognize effectively individuals with the required knowledge to answer a specific question. However, just because people have the capability to answer a question, does not mean that they have the desire to help. In this research, we evaluate the practical performance of the question routing services in social context by analyzing the knowledge sharing behavior of users in social Q&A process in terms of their participation, interests, and connectedness. We collect questions and answers over a ten-month period from Wenwo, a major Chinese question routing service. Using 340,658 questions and 1,754,280 replies, findings reveal separate roles for knowledge sharers and consumers. Based on this finding, we identify knowledge sharers from non-sharers a priori in order to increase the response probabilities. We evaluate our model based on an analysis of 3006 Wenwo knowledge sharers and non-sharers. Our experimental results demonstrate knowledge sharer prediction based solely on non-Q&A features achieves a 70% success rate in accurately identifying willing respondents.

The use of second screen devices with social media facilitates conversational interaction concerning broadcast media events, creating what we refer to as the social soundtrack. In this research, we evaluate the change of the Super Bowl XLIX social soundtrack across three social media platforms on the topical categories of commercials, music, and game at three game phases ( Pre , During , and Post ). We perform statistical analysis on more than 3M, 800K, and 50K posts from Twitter, Instagram, and Tumblr, respectively. Findings show that the volume of posts in the During phase is fewer compared to Pre and Post phases; however, the hourly mean in the During phase is considerably higher than it is in the other two phases. We identify the predominant phase and category of interaction across all three social media sites. We also determine the significance of change in absolute scale across the Super Bowl categories (commercials, music, game) and in both absolute and relative scales across Super Bowl phases ( Pre , During , Post ) for the three social network platforms (Twitter, Tumblr, Instagram). Results show that significant phase-category relationships exist for all three social networks. The results identify the During phase as the predominant one for all three categories on all social media sites with respect to the absolute volume of conversations in a continuous scale. From the relative volume perspective, the During phase is highest for the music category for most social networks. For the commercials and game categories, however, the Post phase is higher than the During phase for Twitter and Instagram, respectively. Regarding category identification, the game category is the highest for Twitter and Instagram but not for Tumblr, which has dominant peaks for music and/or commercials in all three phases. It is apparent that different social media platforms offer various phase and category affordances. These results are important in identifying the influence that second screen technology has on information sharing across different social media platforms and indicates that the viewer role is transitioning from passive to more active.

Despite the considerable interest in the detection of toxic comments, there has been little research investigating the causes -- i.e., triggers -- of toxicity. In this work, we first propose a formal definition of triggers of toxicity in online communities. We proceed to build an LSTM neural network model using textual features of comments, and then, based on a comprehensive review of previous literature, we incorporate topical and sentiment shift in interactions as features. Our model achieves an average accuracy of 82.5% of detecting toxicity triggers from diverse Reddit communities.

We analyze the effect of a smile in personas pictures on persona perceptions, including credibility, likability, similarity, and willingness to use. We conduct an online experiment with 2,400 participants using a 16-item survey and multiple persona profile treatments of which half have a smiling photo and half do not. We find that persona profiles with a smiling photo result in an increase in perceived similarity with, likability of, and willingness to use the personas. In contrast, a smile does not increase the credibility of the personas. Our research has implications for the design of persona profiles and adds to previous findings of persona research that the picture choice influences individuals' persona perceptions in profound ways.

We investigate the existence of demographic bias in automatically generated personas by producing personas from YouTube Analytics data. Despite the intended objectivity of the methodology, we find elements of bias in the data-driven personas. The bias is highest when doing an exact match comparison, and the bias decreases when comparing at age or gender level. The bias also decreases when increasing the number of generated personas. For example, the smaller number of personas resulted in underrepresentation of female personas. This suggests that a higher number of personas gives a more balanced representation of the user population and a smaller number increases biases. Researchers and practitioners developing data-driven personas should consider the possibility of algorithmic bias, even unintentional, in their personas by comparing the personas against the underlying raw data.

Creating personas from actual online user information is an advantage of the data-driven persona approach. However, modern online systems often provide big data from millions of users that display vastly different behaviors, resulting in possibly thousands of personas representing the entire user population. We present a technique for reducing the number of personas to a smaller number that efficiently represents the complete user population, while being more manageable for end users of personas. We first isolate the key user behaviors and demographical attributes, creating thin personas, and we then apply an algorithmic cost function to collapse the set to the minimum needed to represent the whole population. We evaluate our approach on 26 million user records of a major international airline, isolating 1593 personas. Applying our approach, we collapse this number to 493, a 69% decrease in the number of personas. Our research findings have implications for organizations that have a large user population and desire to employ personas.

We use machine learning to predict the search engine rank of webpages. We use a list of keywords for 30 content blogs of an e-commerce company in the gift industry to retrieve 733 content pages occupying the first-page Google rankings and predict their rank using 30 ranking factors. We test two models, Light Gradient Boosting Machine (LightGBM) and Extreme Gradient Boosted Decision Trees (XGBoost), finding that XGBoost performs better for predicting actual search rankings, with an average accuracy of 0.86. The feature analysis shows the most impactful features are (a) internal and external links, (b) security of the web domain, and (c) length of H3 headings, and the least impactful features are (a) keyword mentioned in domain address, (b) keyword mentioned in the H1 headings, and (c) overall number of keyword mentions in the text. The results highlight the persistent importance of links in search-engine optimization. We provide actionable insights for online marketers and content creators.

Predicting user confusion can help improve information presentation on websites, mobile apps, and virtual reality interfaces. One promising information source for such prediction is eye-tracking data about gaze movements on the screen. Coupled with think-aloud records, we explore if user's confusion is correlated with primarily fixation-level features. We find that random forest achieves an accuracy of more than 70% when prediction user confusion using only fixation features. In addition, adding user-level features (age and gender) improves the accuracy to more than 90%. We also find that balancing the classes before training improves performance. We test two balancing algorithms, Synthetic Minority Over Sampling Technique (SMOTE) and Adaptive Synthetic Sampling (ADASYN) finding that SMOTE provides a higher performance increase. Overall, this research contains implications for researchers interested in inferring users' cognitive states from eye-tracking data.

Increased access to data and computational techniques enable innovations in the space of automated customer analytics, for example, automatic persona generation. Automatic persona generation is the process of creating data-driven representations from user or customer statistics. Even though automatic persona generation is technically possible and provides advantages compared to manual persona creation regarding the speed and freshness of the personas, it is not clear (a) what information to include in the persona profiles and (b) how to display that information. To query into these aspects relating information design of personas, we conducted a user study with 38 participants. In the findings, we report several challenges relating to the design of automatically generated persona profiles, including usability issues, perceptual issues, and issues relating to information content. Our research has implications for the information design of data-driven personas.

As the quantity of social and online analytics data has drastically increased, a wide variety of methods are deployed to make sense of this data, typically via computational and algorithmic approaches. However, in many cases, these approaches trade one form of complexity for another by ignoring the principles of human cognitive processing. In this perspective manuscript, we propose an approach of employing Personas as an alternative form of making large volumes of online user analytics information useful to end users of the user and customer analytics, with results applicable in software development, business sectors, communication industry, and other domains where understanding online user behavior is deemed important. Toward this end, we have developed a system that automatically generates data-driven Personas from social media and online analytics data, capable of handling hundreds of millions of user interactions from tens of thousands of pieces of content on YouTube, Facebook and Google Analytics, while retaining the privacy of individual users of those channels. Our approach (1) identifies and prioritizes user segments by their online behavior, (2) associates the segments with demographic data, and (3) creates rich Persona profiles by dynamically adding characteristics, such as names, photos, and descriptive quotes. This chapter characterizes the currently open research problems in automatic Persona generation, such as de-aggregation of data, cross-platform data mapping, filtering of toxic comments, and choosing the right information content according to end-user needs. Addressing these problems requires the use of state-of-the-art techniques of computer and information science within one system and benefits greatly from inter-disciplinary collaboration. Overall, the research agenda set in this work aims at achieving the vision for automatic user profiling using diverse online and social media platforms and advanced data processing methods for the end goal of making complex analytics data more useful for human decision makers, especially those working with online content.

The massive volume of online analytics data about customers has led to novel opportunities for user segmentation. However, getting real value from data remains challenging for many organizations. One of the recent innovations in online analytics is data-driven persona generation that can be used to create high-quality human representations from online analytics data. This manuscript (a) summarizes the potential of data-driven persona generation for online analytics, (b) characterizes nine open research questions for data-driven persona generation, and (c) outlines a research agenda for making persona analytics more useful for decision makers.

In this research study, the authors investigate the association between external searching, which is searching on a web search engine, and internal searching, which is searching on a website. They classify 295,571 external – internal searches where each search is composed of a search engine query that is submitted to a web search engine and then one or more subsequent queries submitted to a commercial website by the same user. The authors examine 891,453 queries from all searches, of which 295,571 were external search queries and 595,882 were internal search queries. They algorithmically classify all queries into states, and then clustered the searching episodes into major searching configurations and identify the most commonly occurring search patterns for both external, internal, and external-to-internal searching episodes. The research implications of this study are that external sessions and internal sessions must be considered as part of a continuous search episode and that online businesses can leverage external search information to more effectively target potential consumers.

Many people turn to their social networks to find information through the practice of question and answering. We believe it is necessary to use different answering strategies based on the type of questions to accommodate the different information needs. In this research, we propose the ASK taxonomy that categorizes questions posted on social networking sites into three types according to the nature of the questioner's inquiry of accuracy, social, or knowledge. To automatically decide which answering strategy to use, we develop a predictive model based on ASK question types using question features from the perspectives of lexical, topical, contextual, and syntactic as well as answer features. By applying the classifier on an annotated data set, we present a comprehensive analysis to compare questions in terms of their word usage, topical interests, temporal and spatial restrictions, syntactic structure, and response characteristics. Our research results show that the three types of questions exhibited different characteristics in the way they are asked. Our automatic classification algorithm achieves an 83% correct labeling result, showing the value of the ASK taxonomy for the design of social question and answering systems.