Research

The research of the Management Analytics Center is based on the cooperation of the individual chairs and their respective teams. More information about the individual authors can be found within the respective project. An overview of contact information of the individual chairs in their respective areas can be found on the team website.

    Technologien

  • Data Mining

    LeAP – Learning Analytics Profile

    Author(s): 

    Dirk Ifenthaler

    Cooperation Partners:

    Chair of Learning, Design and Technology

     

    Description:

    The field of learning analytics is generating growing interest in data and computer science as well educational science and hence, becoming an important aspect of modern e-learning environments. However, implementing learning analytics into existing legacy systems in learning environments with high data privacy concerns is quite a challenge. This research shows how a learning analytics application has been implemented into such an existing university environment by adding a plugin to the local digital learning environment which injects user centric information to specific objects within students’ and teachers’ learning environment. The LeAP platform can be used in various learning and teaching scenarios in order to provide adaptive support for learners and teachers.

     


    Advertising Over Time: Moving in the speed-of-life?

    Author(s): 

    Leonie Gehrmann, Florian Stahl

    Cooperation Partners: 

    Hebrew University (Israel), Economist

     

     

    Description:

    We analyze how print advertising content has adapted to changes in speed-of-life, best defined as the “rapidity and density of experiences, meanings, perceptions, and activities” (Werner et al. 1985), over time.

    On the one hand, the development of our image mining algorithm promises relevance for a variety of alternative use cases, since it is easily extendable to images beyond the print advertisements in our data set. Similarly, the analysis could be continued for other magazines, as well as social media postings, further refining our algorithm. On the other hand, our research also has direct implications for marketing practice. Increasing advertising clutter is a largely recognized phenomenon and our findings are likely to provide insights into the existing competition for consumer attention.


    Variation in review writing behavior across time-of-the-day, week day, season, weather conditions, a city’s life-quality and other factors

    Author(s): 

    Andreas Bayerl, Florian Stahl, Jacob Goldenberg

    Cooperation Partners:

    IDC Israel, Kununu.de

     

    Description:

    We analyze a data set of in total more than 10 Mio online reviews from two different platforms collecting User-Generated Content to find differences in review writing behavior across time-of-the-day, week day, season, weather conditions, a city’s life quality and other factors. Online reviews are meant to be an unbiased source of information. We will show how in fact, they can be influenced by externalities that have nothing or very little to do with the reviewed instance. By doing so we will follow a suggestion by Miller (2011) from his Science article saying that the power of the Internet and unstructured data can help take research to the next level. Methods employed during this project are among others, web scraping to acquire data, Natural Language Processing (NLP) to analyze data and Machine Learning to develop a model predicting review scores based on externalities. 

     


    The Impact of the Offered Product Mix on the Acquisition, Usage and Retention of Information Goods

    Author(s): 

    Daniela Schmitt, Raghuram Iyengar, Florian Stahl

    Cooperation Partners: 

    Wharton School (Philadelphia)

     

    Description:

    A primary challenge for information goods providers (e.g., news websites) is maintaining a balance between their free users and paid subscribers by determining which of the former are likely to convert to the latter and when. We address questions of what kinds of content do users consume and how much do the inter-session consumption dynamics impact if/ when they subscribe. Model-free evidence shows that an increase in paid content is associated with an increase in the probability of subscribing (acquisition) and retention, which is moderated by the volume of users' consumption. However there is also a segment of users that abandons the website on seeing an increase in the amount of paid content. These results suggest that the amount of paid content has heterogeneous effects across users on conversion and consumption. We specify a hidden Markov model that captures the underlying dynamics in behavior as users engage with the website.

    Theoretically, our results help understand different users' responses in usage behavior to product mix changes both in the short and in the long run. From a managerial perspective, our results provide actionable implications for firms that employ the freemium business model (e.g. newspaper websites, video or music streaming services) on how a change in their product mix can impact the user experience and behavior.


    Competence development in VET enculturation processes

    Author(s): 

    Viola Deutscher

    Cooperation Partners: 

    Esther Winther, University of Duisburg-Essen; Vocational schools/companies with VET programmes

     

    Description:

    The DFG funded research project analyzes the competence development in vocational education training (VET) as well as relevant success factors for competence development for commercial apprentices (n = 877). With this objective, the project refers to the modeling of VET enculturation processes at both group and individual level, including the description of company learning conditions.

    A validated competence assessment is used simultaneously with an instrument for recording the enculturation conditions of vocational learning at the beginning, at the middle and at the end of initial vocational training (longitudinal design). Via this design, practical findings about the genesis of professional competences can be generated that help companies to improve their educational programmes.


    A friendly turn: advertising bias in the news media

    Author(s): 

    Florens Focke, Alexandra Niessen-RuenziStefan Ruenzi

    Cooperation Partners: 

    Chair of International Finance & Chair of Corporate Governance

    Description:

    This paper investigates whether newspapers report more favorably about advertising corporate clients than about other firms. Our identification strategy based on high-dimensional fixed effects and high frequency advertising data shows that advertising leads to more positive press coverage. This advertising bias in reporting is found among local and national newspapers. Further results show that advertising bias manifests particularly in less negative reporting after bad news events such as negative earnings surprises or extremely negative stock returns. Our findings cast doubt on the independence of the press from corporate pressure and hint at important information frictions.


    Prediction of Consumer Behavior

    Author(s): 

    Yasid Soufi, Florian Kraus

    Cooperation Partners: 

    Chair of Sales & Services Marketing

     

     

    Description:

    Gaining insights into predictors of a customer’s purchase behavior and personal traits enables a more flexible and adaptive approach to the sales process. Utilizing data on behavioral patterns, consumer characteristics and self-reported traits, we aim to understand and anticipate customer behavior in an app-based context. This is made possible by the application of machine learning algorithms to the enriched data set and leads to a better understanding of mechanisms underlying consumer decisions and the leveraging of predictive abilities in the optimization of the sales process.

    Applications arise in many areas, in which behavioral patterns can be observed and more enriched customer data could possibly be gathered. This data, whilst valuable, remains often unsaved and unexplored. Insights may be valuable, both from a customer and provider perspective, as optimized processes lead to higher customer satisfaction and operational efficiency.


    Marketing in Finance

    Author(s): 

    Yasid Soufi, Florian Kraus

    Cooperation Partners: 

    Chair of Sales & Services Marketing

     

     

    Description:

    Skillful communication with participants of financial markets enables the development of new financing sources and ameliorated sentiment towards a company’s activities and value potential. Well-executed promotion in the context of Investor Relations can reduce opportunity costs when going public and afterwards drive stock returns as well as ease return volatility.

    High-quality Investor Relations (IR) can be a critical success factor for corporations with high exposure to reception in the financial community. Firms aiming to undergo an Initial Public Offering (IPO) as well as firms, which are already listed, are legally bound to provide filings to authorities and the public, that can be utilized as a means of communication and promotion of a firm’s value potential.  Crawling large amounts of textual data and deploying methodologies of textual analysis, we aim to explore, how marketing should be communicating in the frame of Investor Relations.


    The scale of sharing transaction via online platforms and issues for taxation in the digital economy

    Author(s): 

    Christopher Ludwig, Christoph Spengel, Rainer Bräutigam

    Cooperation Partners: 

    Chair of Business Administration and Taxation II

    Description:

    This project is based on publicly accessible Airbnb data regarding the type of lodging offered on the platform, the price per night, the facilities and equipment, the location of the accommodation as well as booking trends in 20 German cities. On the basis of this data we carry out projections to estimate the annual turnover of hosts as well as the potential tax revenue arising from this turnover. The annual turnover of all accommodations in the 20 considered cities is approximately 683 million euros. Based on the annual turnover projections, the analysis shows that the estimated tax revenue from income and value-added tax liability, is enormous for lodgings offered on the exemplarily analyzed sharing platform.
    Given the size of tax revenues at stake, we develop recommendations on how to adjust existing tax systems in order to adapt them to the digital economy and to ensure compliance among taxpayers.


    The Complexity of Multinational Business Groups

    Author(s): 

    Marcel OlbertChristoph Spengel

    Cooperation Partners: 

    Chair of Business Administration and Taxation II

     

    Description:

    By tracking the development of organizational structure over time, the insights from this project help to understand how multinational firms structure their value chains and businesses models globally. The findings can be used to evaluate policy measures and to provide recommendations to legislators.

    This project uses data on financial characteristics, legal status, and shareholder structures of more than 250 million firms around the globe to investigate the determinants of organizational structures and firms’ complexity. To this end, a sophisticated algorithm is used to iteratively link information on individual firms to create a group of affiliated firms that ultimately belong to the same global owner. This study attempts to explain why multinational firms employ complex organizational structures and ambiguous legal strategies such as founding subsidiaries in tax havens.


    Digitalization of the Tax Function

    Author(s): 

    Christoph Spengel, Alexander Maedche, Sven Scheu, Sarah Winter

    Cooperation Partners: 

    Institute of Information Systems and Marketing (IISM) at KIT, EnBW AG

    Description:

    Tax departments are facing new regulatory requirements to provide an increasing amount of information to tax authorities. Thus, being able to process structured and unstructured tax data in a timely manner and improve decision making processes is essential. 

    We collaborate with a major German energy supplier to develop new concepts for the digital transformation of the tax function. The project focuses on optimizing processes in the tax department through digital means and enhancing existing tax information systems and intragroup information processing. The project also aims at developing data-driven intelligent assistant functions to support decision-making processes using Process Mining, Business Intelligence, and more generally spoken Data Science.


    Can European Banks’ Country-by-Country Reports Reveal Profit Shifting? An Analysis of the Information Content of EU Banks’ Disclosures

    Author(s): 

    Verena Dutt, Katharina Nicolay, Heiko Vay, Johannes VogetChristoph Spengel

    Cooperation Partners: 

    Chair of Business Administration & Taxation II, Taxation & Finance

    Description:

    The insights of this project are especially relevant in the context of the ongoing political discussions whether to introduce a public country-by-country (CbC) reporting for all large multinational firms in the EU.

    The project uses hand-collected data – facilitated with an automated process – from mandatory CbC-reports of European banks. The data offer a precise look into the amount of profits and turnover and the number of employees in the jurisdictions where multinational banks operate and allow to estimate the profitability of different locations. The goal of this project is to identify if public CbC-reporting provides incremental information on the profit shifting behavior of multinational banks compared to commercial databases like Orbis.


    Driving forward while looking back: Employee responses to error amidst when monitored by emergent technologies

    Author(s): 

    Rajiv Sabherwal, Hartmut Höhle, Tobias Nisius, Florian Pethig, Fareed Zandkarimi

    Cooperation Partners: 

    Aalberts Material Technology GmbH

    Description:

    In this project we are examining the effect of reclamations on several behaviors of production staff. Using process mining allows us to measure production-level behavior of workers. These measures are collected over a three-year period and are planned to be analyzed using a difference-in-difference (DID) design. The expected results should broaden our understanding of employee responses to error. 

    We are looking into core-production data of a very large company. Applying statistical methods on production-level data, combined with the application of process mining, is one major contribution of our work. Also, the expected outcomes related to the research problem, i.e., employee responses to reclamations, provide practitioners new insights on making effective decisions before and after an error is reported.

    During the data collection phase, we collected 1.4 million production cases via access to the company’s enterprise resource planning (ERP) system. In order to apply process mining, 500+ lines of SQL code were developed to transform the data into processable event logs. Also, a Java program was written to collect appropriate data for the DID analysis. The aggregated dataset is currently analyzed using Stata.


    “Helpful or Harmful? Negative Behavior Toward Newcomers and Welfare in Online Communities

    Author(s): 

    Florian Pethig, Kai-Lung Hui, Andreas Lanz, Hartmut Höhle

    Cooperation Partners: 

    Hong Kong University of Science and Technology, HEC Paris

    Description:

    New members are important for the survival of online communities. However, hostility toward newcomers is pronounced in many online communities, often exercised through downvotes, rejections, and negative comments from established members. Online communities have realized that such negativity can take a toll on newcomers. In this paper, we study a new intervention aimed at reducing hostility toward newcomers: a “newcomer nudge,” which informs members when they are interacting with a newcomer post and nudges them to be more lenient toward its author. We use a natural experiment research design and analyze 5,027 newcomer posts published in a 90-day time window before and after the introduction of the nudge. We observe a strong increase in upvotes and number of responses as well as a decrease in negative sentiment in the responses. Taking advantage of the panel data, we find that newcomers who are socialized with the nudge are significantly more likely to post again in the following 12 months than newcomers socialized without the nudge.

  • Internet of Things

  • Artificial Intelligence

    Prediction of Consumer Behavior

    Author(s): 

    Yasid Soufi, Florian Kraus

    Cooperation Partners: 

    Chair of Sales & Services Marketing

     

    Description:

    Gaining insights into predictors of a customer’s purchase behavior and personal traits enables a more flexible and adaptive approach to the sales process. Utilizing data on behavioral patterns, consumer characteristics and self-reported traits, we aim to understand and anticipate customer behavior in an app-based context. This is made possible by the application of machine learning algorithms to the enriched data set and leads to a better understanding of mechanisms underlying consumer decisions and the leveraging of predictive abilities in the optimization of the sales process.

    Applications arise in many areas, in which behavioral patterns can be observed and more enriched customer data could possibly be gathered. This data, whilst valuable, remains often unsaved and unexplored. Insights may be valuable, both from a customer and provider perspective, as optimized processes lead to higher customer satisfaction and operational efficiency.


    Advertising Over Time: Moving in the speed-of-life?

    Author(s): 

    Leonie Gehrmann, Florian Stahl

    Cooperation Partners: 

    Hebrew University (Israel), Economist

     

     

    Description:

    We analyze how print advertising content has adapted to changes in speed-of-life, best defined as the “rapidity and density of experiences, meanings, perceptions, and activities” (Werner et al. 1985), over time.

    On the one hand, the development of our image mining algorithm promises relevance for a variety of alternative use cases, since it is easily extendable to images beyond the print advertisements in our data set. Similarly, the analysis could be continued for other magazines, as well as social media postings, further refining our algorithm. On the other hand, our research also has direct implications for marketing practice. Increasing advertising clutter is a largely recognized phenomenon and our findings are likely to provide insights into the existing competition for consumer attention.

  • Process Mining

    Digitalization of the Tax Function

    Author(s): 

    Christoph Spengel, Alexander Maedche, Sven Scheu, Sarah Winter

    Cooperation Partners: 

    Institute of Information Systems and Marketing (IISM) at KIT, EnBW AG

    Description:

    Tax departments are facing new regulatory requirements to provide an increasing amount of information to tax authorities. Thus, being able to process structured and unstructured tax data in a timely manner and improve decision making processes is essential. 

    We collaborate with a major German energy supplier to develop new concepts for the digital transformation of the tax function. The project focuses on optimizing processes in the tax department through digital means and enhancing existing tax information systems and intragroup information processing. The project also aims at developing data-driven intelligent assistant functions to support decision-making processes using Process Mining, Business Intelligence, and more generally spoken Data Science.


    Driving forward while looking back: Employee responses to error amidst when monitored by emergent technologies

    Author(s): 

    Rajiv Sabherwal, Hartmut Höhle, Tobias Nisius, Florian Pethig, Fareed Zandkarimi

    Cooperation Partners: 

    Aalberts Material Technology GmbH

    Description:

    In this project we are examining the effect of reclamations on several behaviors of production staff. Using process mining allows us to measure production-level behavior of workers. These measures are collected over a three-year period and are planned to be analyzed using a difference-in-difference (DID) design. The expected results should broaden our understanding of employee responses to error. 

    We are looking into core-production data of a very large company. Applying statistical methods on production-level data, combined with the application of process mining, is one major contribution of our work. Also, the expected outcomes related to the research problem, i.e., employee responses to reclamations, provide practitioners new insights on making effective decisions before and after an error is reported.

    During the data collection phase, we collected 1.4 million production cases via access to the company’s enterprise resource planning (ERP) system. In order to apply process mining, 500+ lines of SQL code were developed to transform the data into processable event logs. Also, a Java program was written to collect appropriate data for the DID analysis. The aggregated dataset is currently analyzed using Stata.


    Process Mining meets Procurement: Application of Process Mining Methods to Measure Inefficiencies in Procurement Processes

    Author(s): 

    Prof. Dr. Christoph BodeProf. Dr. Hartmut HöhleJun.-Prof. Dr. Jana-Rebecca RehseJonas Ronellenfitsch, Fareed Zandkarimi

    Cooperation Partners: 

    Schott AG, Hilti AG, HeidelbergCement AG und Heidelberger Druckmaschinen AG

    Description:

    Procurement processes, such as purchase-to-pay (P2P), are of central importance to companies. Problems and inefficiencies in these processes can not only cause delays in production, but also lead to significant additional costs. However, limited research has been conducted on these processes, particularly at the operational level, meaning that it is not known what inefficiencies occur in the actual execution of the processes, where they come from, and how they can be prevented if necessary.
    To address this question, an interdisciplinary research team of process mining and procurement experts is investigating how methods of process mining, i.e., the analysis of process execution data, can be applied to identify, measure, explain, and ideally eliminate inefficiencies in procurement processes. To this end, data from P2P processes of three large industrial companies are analyzed and examined for common patterns. Based on this research, new key performance indicators (KPIs) are developed that precisely quantify the inefficiency of a P2P process. These KPIs can now be used, for example, to identify temporal patterns in process inefficiency or to highlight the impact of an external influence, such as the Corona pandemic, on procurement processes.

  • Blockchain

  • Digital Platforms

    Competence development in VET enculturation processes

    Author(s): 

    Viola Deutscher

    Cooperation Partners: 

    Esther Winther, University of Duisburg-Essen; Vocational schools/companies with VET programmes

    Description:

    The DFG funded research project analyzes the competence development in vocational education training (VET) as well as relevant success factors for competence development for commercial apprentices (n = 877). With this objective, the project refers to the modeling of VET enculturation processes at both group and individual level, including the description of company learning conditions.

    A validated competence assessment is used simultaneously with an instrument for recording the enculturation conditions of vocational learning at the beginning, at the middle and at the end of initial vocational training (longitudinal design). Via this design, practical findings about the genesis of professional competences can be generated that help companies to improve their educational programmes.


    LeAP – Learning Analytics Profile

    Author(s): 

    Dirk Ifenthaler

    Cooperation Partners:

    Chair of Learning, Design and Technology

     

    Description:

    The field of learning analytics is generating growing interest in data and computer science as well educational science and hence, becoming an important aspect of modern e-learning environments. However, implementing learning analytics into existing legacy systems in learning environments with high data privacy concerns is quite a challenge. This research shows how a learning analytics application has been implemented into such an existing university environment by adding a plugin to the local digital learning environment which injects user centric information to specific objects within students’ and teachers’ learning environment. The LeAP platform can be used in various learning and teaching scenarios in order to provide adaptive support for learners and teachers.


    Ad-blockers and Content Differentiation

    Author(s): 

    Gokhan Gecer, Florian Kraus, Pinar Yildirim

    Cooperation Partners: 

    Wharton School (Philadelphia)

     

    Description:

    An increasing number of people use ad-blockers recently which becomes a threat to the main revenue stream of publishers. In order to deal with this challenge, some publishers offer an ad-free (premium) version of their websites to customers who pay a subscription fee. Others ask the users to turn off their ad-blocker. We have developed a game-theoretical model of a media ecosystem that contains publishers and users. Publishers maximize their profits from advertising income and/or subscription fees. Users consume the content provided by the publishers deciding on the amount of content they consume. They are utility maximizers. In our model, the publisher could choose one of two strategies against the ad-blockers: (1) show the content only to those who turn the ad-blockers off, and (2) show the content to users who pay a subscription fee beside the ones who turn the ad-blockers off. We have analyzed the model in monopoly and duopoly markets.

    The results show that when producing content is costly, the publisher decreases its content quality, accordingly the number of users who consume the content decreases. Hence, the publisher offers a subscription option to the users. In a duopoly market, the substitutability of the content is also important. Our results show that when a publisher produces less substitutable content for a low cost (YouTubers and bloggers), it does not offer a subscription option. On the other hand, when the content is more substitutable (newspapers and tabloids) and when the content is less substitutable but it is costly to produce it (websites with a specific focus), the publisher offers a subscription option.


    Maximizing Customer Lifetime Value through Strategic Channel Management: How to Incentivize Customers to Use a Mobile App versus a Website

    Author(s): 

    Gokhan Gecer, Florian Kraus

    Cooperation Partners: 

    Chair of Sales & Services Marketing

     

     

    Description:

    Customer engagement with mobile devices has changed customers' habits in online purchasing, in particular the usage of mobile apps for shopping increases the customer lifetime value (CLV). Hence, we suggest the idea that online retailers should steer their customers to the mobile channel by offering a permanent discount over the mobile app. Although this strategy decreases the short-term income, it may increase the CLV. Based on this suggestion, we develop a probability-based CLV model to show to what type of customers an online retailer should offer a discount over the mobile app as well as the optimal value of that discount.

    Following an analytical modeling approach, we are able to show that the online retailer should offer such a discount to a customer who is either very likely or unlikely to increase her purchasing probability. Also, the firm could offer such a discount to encourage the customer to switch to the mobile app. Online retailers should not offer the discount to those who already have a high purchasing probability because their CLV is already huge. Last but not least, an online retailer should not follow this strategy to gain new customers (customer acquisition) but rather apply it to increase customer retention.


    The scale of sharing transactions via online platforms and issues for taxation in the digital economy

    Author(s): 

    Christopher Ludwig, Christoph Spengel, Rainer Bräutigam

    Cooperation Partners: 

    Chair of Business Administration and Taxation II

    Description:

    This project is based on publicly accessible Airbnb data regarding the type of lodging offered on the platform, the price per night, the facilities and equipment, the location of the accommodation as well as booking trends in 20 German cities. On the basis of this data we carry out projections to estimate the annual turnover of hosts as well as the potential tax revenue arising from this turnover. The annual turnover of all accommodations in the 20 considered cities is approximately 683 million euros. Based on the annual turnover projections, the analysis shows that the estimated tax revenue from income and value-added tax liability, is enormous for lodgings offered on the exemplarily analyzed sharing platform.
    Given the size of tax revenues at stake, we develop recommendations on how to adjust existing tax systems in order to adapt them to the digital economy and to ensure compliance among taxpayers.

  • Machine Learning

    LeAP – Learning Analytics Profile

    Author(s): 

    Dirk Ifenthaler

    Cooperation Partners:

    Chair of Learning, Design and Technology

     

    Description:

    The field of learning analytics is generating growing interest in data and computer science as well educational science and hence, becoming an important aspect of modern e-learning environments. However, implementing learning analytics into existing legacy systems in learning environments with high data privacy concerns is quite a challenge. This research shows how a learning analytics application has been implemented into such an existing university environment by adding a plugin to the local digital learning environment which injects user centric information to specific objects within students’ and teachers’ learning environment. The LeAP platform can be used in various learning and teaching scenarios in order to provide adaptive support for learners and teachers.


    Advertising Over Time: Moving in the speed-of-life?

    Author(s): 

    Leonie Gehrmann, Florian Stahl

    Cooperation Partners: 

    Hebrew University (Israel), Economist

     

     

    Description:

    We analyze how print advertising content has adapted to changes in speed-of-life, best defined as the “rapidity and density of experiences, meanings, perceptions, and activities” (Werner et al. 1985), over time.

    On the one hand, the development of our image mining algorithm promises relevance for a variety of alternative use cases, since it is easily extendable to images beyond the print advertisements in our data set. Similarly, the analysis could be continued for other magazines, as well as social media postings, further refining our algorithm. On the other hand, our research also has direct implications for marketing practice. Increasing advertising clutter is a largely recognized phenomenon and our findings are likely to provide insights into the existing competition for consumer attention.


    Variation in review writing behavior across time-of-the-day, week day, season, weather conditions, a city’s life-quality and other factors

    Author(s): 

    Andreas Bayerl, Florian Stahl, Jacob Goldenberg

    Cooperation Partners:

    IDC Israel, Kununu.de

     

    Description:

    We analyze a data set of in total more than 10 Mio online reviews from two different platforms collecting User-Generated Content to find differences in review writing behavior across time-of-the-day, week day, season, weather conditions, a city’s life quality and other factors. Online reviews are meant to be an unbiased source of information. We will show how in fact, they can be influenced by externalities that have nothing or very little to do with the reviewed instance. By doing so we will follow a suggestion by Miller (2011) from his Science article saying that the power of the Internet and unstructured data can help take research to the next level. Methods employed during this project are among others, web scraping to acquire data, Natural Language Processing (NLP) to analyze data and Machine Learning to develop a model predicting review scores based on externalities. 


    Prediction of Consumer Behavior

    Author(s): 

    Yasid Soufi, Florian Kraus

    Cooperation Partners: 

    Chair of Sales & Services Marketing

     

     

    Description:

    Gaining insights into predictors of a customer’s purchase behavior and personal traits enables a more flexible and adaptive approach to the sales process. Utilizing data on behavioral patterns, consumer characteristics and self-reported traits, we aim to understand and anticipate customer behavior in an app-based context. This is made possible by the application of machine learning algorithms to the enriched data set and leads to a better understanding of mechanisms underlying consumer decisions and the leveraging of predictive abilities in the optimization of the sales process.

    Applications arise in many areas, in which behavioral patterns can be observed and more enriched customer data could possibly be gathered. This data, whilst valuable, remains often unsaved and unexplored. Insights may be valuable, both from a customer and provider perspective, as optimized processes lead to higher customer satisfaction and operational efficiency.


    Marketing in Finance

    Author(s): 

    Yasid Soufi, Florian Kraus

    Cooperation Partners: 

    Chair of Sales & Services Marketing

     

     

    Description:

    Skillful communication with participants of financial markets enables the development of new financing sources and ameliorated sentiment towards a company’s activities and value potential. Well-executed promotion in the context of Investor Relations can reduce opportunity costs when going public and afterwards drive stock returns as well as ease return volatility.

    High-quality Investor Relations (IR) can be a critical success factor for corporations with high exposure to reception in the financial community. Firms aiming to undergo an Initial Public Offering (IPO) as well as firms, which are already listed, are legally bound to provide filings to authorities and the public, that can be utilized as a means of communication and promotion of a firm’s value potential.  Crawling large amounts of textual data and deploying methodologies of textual analysis, we aim to explore, how marketing should be communicating in the frame of Investor Relations.


    Is the stock market biased against diverse top management teams?

    Author(s): 

    Prof. Oliver Spalt, Alberto Manconi, Antonino Rizzo

    Cooperation Partners:

    Chair of Financial Markets and Financial Institutions

    Description:

    Using a novel text-based measure of top management team diversity, covering over 70,000 top executives in over 6,500 U.S. firms from 1999 to 2014, we show that analyst forecasts are systematically more pessimistic for firms with more diverse top management teams ( diverse firms ), especially for inexperienced analysts. Institutional investors, especially if located in conservative areas, are less likely to hold diverse firms, even though diverse firms do not exhibit inferior returns. Consistent with downward-biased expectations, abnormal returns on information-release days are systematically positive for diverse firms. Combined, our results suggest stock markets are biased against diversity in top management teams.



    Qualitative Information Disclosure and Tax Aggressiveness: Is Mandating Additional Information Disclosure Useful?

    Author(s): 

    Elisa Casi, Katarzyna Bilicka, Carol Seregni, Barbara StageChristoph Spengel

    Cooperation Partners:

    Chair of Business Administration and Taxation II

    Description:

    In reponse to the debates about tax avoidance of multinational companies, legislators have introduced several tax transparency measures. With the insights of our study, we aim at providing policy advice on the desirability of mandating qualitative tax disclosure.

    We study the effects of mandatory qualitative tax information disclosure on tax avoidance. We consider a tax transparency reform in UK that required firms to disclose tax strategy reports. We manually collect around 2,000 tax strategy reports and apply textual analysis techniques involving supervised machine learning tools.

  • Mobile Technologies

    Competence development in VET enculturation processes

    Author(s): 

    Viola Deutscher

    Cooperation Partners: 

    Esther Winther, University of Duisburg-Essen; Vocational schools/companies with VET programmes

    Description:

    The DFG funded research project analyzes the competence development in vocational education training (VET) as well as relevant success factors for competence development for commercial apprentices (n = 877). With this objective, the project refers to the modeling of VET enculturation processes at both group and individual level, including the description of company learning conditions.

    A validated competence assessment is used simultaneously with an instrument for recording the enculturation conditions of vocational learning at the beginning, at the middle and at the end of initial vocational training (longitudinal design). Via this design, practical findings about the genesis of professional competences can be generated that help companies to improve their educational programmes.


    Maximizing Customer Lifetime Value through Strategic Channel Management: How to Incentivize Customers to Use a Mobile App versus a Website

    Author(s): 

    Gokhan Gecer, Florian Kraus

    Cooperation Partners: 

    Chair of Sales & Services Marketing

     

     

    Description:

    Customer engagement with mobile devices has changed customers' habits in online purchasing, in particular the usage of mobile apps for shopping increases the customer lifetime value (CLV). Hence, we suggest the idea that online retailers should steer their customers to the mobile channel by offering a permanent discount over the mobile app. Although this strategy decreases the short-term income, it may increase the CLV. Based on this suggestion, we develop a probability-based CLV model to show to what type of customers an online retailer should offer a discount over the mobile app as well as the optimal value of that discount.

    Following an analytical modeling approach, we are able to show that the online retailer should offer such a discount to a customer who is either very likely or unlikely to increase her purchasing probability. Also, the firm could offer such a discount to encourage the customer to switch to the mobile app. Online retailers should not offer the discount to those who already have a high purchasing probability because their CLV is already huge. Last but not least, an online retailer should not follow this strategy to gain new customers (customer acquisition) but rather apply it to increase customer retention.

  • Social Media

    Variation in review writing behavior across time-of-the-day, week day, season, weather conditions, a city’s life-quality and other factors

    Author(s): 

    Andreas Bayerl, Florian Stahl, Jacob Goldenberg

    Cooperation Partners:

    IDC Israel, Kununu.de

     

     

    Description:

    We analyze a data set of in total more than 10 Mio online reviews from two different platforms collecting User-Generated Content to find differences in review writing behavior across time-of-the-day, week day, season, weather conditions, a city’s life quality and other factors. Online reviews are meant to be an unbiased source of information. We will show how in fact, they can be influenced by externalities that have nothing or very little to do with the reviewed instance. By doing so we will follow a suggestion by Miller (2011) from his Science article saying that the power of the Internet and unstructured data can help take research to the next level. Methods employed during this project are among others, web scraping to acquire data, Natural Language Processing (NLP) to analyze data and Machine Learning to develop a model predicting review scores based on externalities. 


    The Impact of the Offered Product Mix on the Acquisition, Usage and Retention of Information Goods

    Author(s): 

    Daniela Schmitt, Raghuram Iyengar, Florian Stahl

    Cooperation Partners: 

    Wharton School (Philadelphia)

     

     

    Description:

    A primary challenge for information goods providers (e.g., news websites) is maintaining a balance between their free users and paid subscribers by determining which of the former are likely to convert to the latter and when. We address questions of what kinds of content do users consume and how much do the inter-session consumption dynamics impact if/ when they subscribe. Model-free evidence shows that an increase in paid content is associated with an increase in the probability of subscribing (acquisition) and retention, which is moderated by the volume of users' consumption. However there is also a segment of users that abandons the website on seeing an increase in the amount of paid content. These results suggest that the amount of paid content has heterogeneous effects across users on conversion and consumption. We specify a hidden Markov model that captures the underlying dynamics in behavior as users engage with the website.

    Theoretically, our results help understand different users' responses in usage behavior to product mix changes both in the short and in the long run. From a managerial perspective, our results provide actionable implications for firms that employ the freemium business model (e.g. newspaper websites, video or music streaming services) on how a change in their product mix can impact the user experience and behavior.