Filtern
Erscheinungsjahr
Dokumenttyp
Publikationsweg
- Open Access (6)
Schlagworte
- Analytics (2)
- Outlier Detection (2)
- Pricing (2)
- Ageing consumers (1)
- Amazon MTurk (1)
- Business and International Management (1)
- Clustering (1)
- Consumer well-being (1)
- Data Visualization (1)
- Economics and Econometrics (1)
Institut
This study presents a dynamic, model-based view of consumers’ ageing developments, focused on gender differences, to uncover the pathways and socioeconomic transitions that female and male consumers take through old age. The analysis of longitudinal survey data spanning 15 years uses a latent Markov dynamic cluster model with transitions over time. The resulting life courses allow an exploration of lifestyle-related changes in multiple consumer well-being variables beyond age 50. Substantial well-being differences appear in the ageing paths of men and women. In both cases, a dominant chronological sequence through old age is complemented by less common transitions, rarely associated with advanced age. Although the model does not use chronological age as an independent variable, it outperforms purely agebased, or age- cohort-, and period-based models in predicting old-age consumer wellbeing. These results highlight the importance of considering within-cohort diversity when modelling the accompaniments of old age: while some older consumers enjoy active lifestyles, others of similar age succumb to depression and loneliness, rendering age an insufficient predictor of well-being states. In the future, the presented model could be matched with other, even cross-sectional, consumer survey data to help predict various dynamics in the ageing consumer population.
Bike-sharing is a popular component of sustainable urban mobility. It requires anticipatory planning, e.g. of station locations and inventory, to balance expected demand and capacity. However, external factors such as extreme weather or glitches in public transport, can cause demand to deviate from baseline levels. Identifying such outliers keeps historic data reliable and improves forecasts. In this paper we show how outliers can be identified by clustering stations and applying a functional depth analysis. We apply our analysis techniques to the Washington D.C. Capital Bikeshare data set as the running example throughout the paper, but our methodology is general by design. Furthermore, we offer an array of meaningful visualisations to communicate findings and highlight patterns in demand. Last but not least, we formulate managerial recommendations on how to use both the demand forecast and the identified outliers in the bike-sharing planning process.
Engagement is important for the success of social media offers. Accordingly, the goal of responsible social media managers is to generate the maximum engagement with their content. Extant research has already addressed the factors that are related to the creation of compelling content. However, very little attention has been paid to the caption, and thus to a part of the social media post that can be easily changed.
Our aim is to measure the impact of caption characteristics on engagement. We use, text length, number of emojis, and hashtags as caption variables, as well as comments, likes, and shares as engagement variables. In two different datasets, one from a merchandise online store (Facebook, Study 1) and another from a travel influencer (Instagram, Study 2) we find a similar robust result: the analyses reveal s a non-linear relationship between caption text length and engagement that follows an inverted U-shape. It follows that, contrary to previous assumptions, longer captions of around 600 characters should be used to in¬crease engagement. Estimations show that an increase in engagement of about 20% is possible.
Entrepreneurs misdiagnosing their ventures’ innovativeness: The roles of causation and effectuation
(2023)
This text discusses the concept of misdiagnosis in entrepreneurial ventures, where managers' unrealistic optimism can lead to misaligned strategies and late pivots. The study explores the connection between misdiagnosis and decision logics (effectuation and causation) through motivated reasoning and beliefs. The researchers conducted a study with 266 ventures from innovative industries and measured misdiagnosis by comparing managers' judgments with external experts' assessments of venture innovativeness. The results show that causation logic positively influences misdiagnosis, but this effect is reduced when coupled with flexibility. The study highlights the importance of objective performance measures and suggests that motivated reasoning and beliefs explain the link between causation logic and misdiagnosis. Additionally, older and larger ventures with a prevalence of causation logic may experience more misdiagnosis and potential delays in innovation efforts. Flexibility is crucial for older ventures to remain innovative.
How To Optimize Your Social Media Caption To Generate More Engagement :) #Captionize #SocialMedia
(2023)
Maximizing engagement is an important issue for social media managers. Science has already provided guidance on how to make content more engaging. However, one easy-to-edit part of the social media post has received little attention: The caption. We show that the caption has an underestimated impact on engagement and we specifically address text length. The relationship between text length and engagement is not linear as described in previous literature but follows an inverted U-shape. Using field data, we estimate the optimal length of captions and provide surprising results. We show that captions are mostly too short instead of too long and that an optimal length can increase the engagement rate by up to 17%.
Social media managers are constantly seeking ways to maximize engagement with their content. While prior research has shed light on strategies for creating engaging content, the impact of captions on engagement has been under-explored. Our study focuses specifically on the relationship between caption length and engagement in social media posts. Contrary to previous literature, we find that the relationship between caption length and engagement follows an inverted U-shape. Using field data from an omnichannel retailer (Facebook) and a travel influencer (Instagram), we determine the optimal caption length for maximizing engagement and present surprising results. Our results are robust for different model and identification specifications, indicating that captions are often too short, rather than too long, and that optimizing caption length in our case can increase engagement by up to 25%. These findings have important implications for social media managers and influencers seeking to enhance the impact of their posts.
Online influencer marketing has matured, and many managers now seek better performance with nano-influencers in seeding campaigns. Compared to macro-influencers, nano-influencers create more private (direct online or offline) brand messages that supposedly deliver more “persuasive punch”. But such private messages lack the usual online tracking per message, forcing managers to ask nano-influencers for self-reports, or to estimate sales at a campaign level if they want to assess their own decisions. Therefore, little is known about how to set up effective nano-influencer campaigns: does it pay to increase the public message share in nano-influencer activity, in order to benefit from private and public influencer activity? Or does it defeat its purpose, diminishing nano-influencers’ persuasiveness and total campaign results? This study analyses a unique campaign-level data set of various seeded marketing campaigns comprising the activity of 700,000 nano-influencers. We model the impact of nano-influencers’ activity on incremental
sales and return-on-investment (ROI) at the campaign
level. To answer our questions, we include effects of public
message share and optimal public message share and their
interaction with activity; instrumenting both for identification,
and controlling for campaign, marketing, and brand context.
We find conditional optimal public message shares for both
sales and ROI. The average campaign benefits most when
increasing influencer activity (through design or incentives)
at public message shares between 4% and 5%. This shows
the importance of striking the right balance between nanoinfluencers’ online and offline activity, and suggests a similar
balance for combining nano- and macro-influencers in
future campaign formats
Online influencer marketing has matured, and many managers now seek better performance with nano-influencers in seeding campaigns. Compared to macro-influencers, nano-influencers create more private (direct online or offline) brand messages that supposedly deliver more “persuasive punch”. But such private messages lack the usual online tracking per message, forcing managers to ask nano-influencers for self-reports, or to estimate sales at a campaign level if they want to assess their own decisions. Therefore, little is known about how to set up effective nano-influencer campaigns: does it pay to increase the public message share in nano-influencer activity, in order to benefit from private and public influencer activity? Or does it defeat its purpose, diminishing nano-influencers’ persuasiveness and total campaign results? This study analyses a unique campaign-level data set of various seeded marketing campaigns comprising the activity of 700,000 nano-influencers. We model the impact of nano-influencers’ activity on incremental sales and return-on-investment (ROI) at the campaign level. To answer our questions, we include effects of public message share and optimal public message share and their interaction with activity; instrumenting both for identification, and controlling for campaign, marketing, and brand context. We find conditional optimal public message shares for both sales and ROI. The average campaign benefits most when increasing influencer activity (through design or incentives) at public message shares between 4% and 5%. This shows the importance of striking the right balance between nano-influencers’ online and offline activity, and suggests a similar balance for combining nano- and macro-influencers in future campaign formats.
Today’s retailers have a strategic imperative to integrate their channels. Some have implemented electronic shelf labels (ESL) to replace paper tags to technologically enable the omnichannel transformation by aligning the presentation of price and product information between online and offline channels. However, consumer reactions to ESL are yet unexplored. They could be positive or negative: on one hand, the fear of frequent price changes, a known phenomenon in e-commerce, could spread to offline channels and reduce consumer purchase intent and overall revenue; on the other hand, ESL could prevent showrooming by signaling price consistency and offering consistent information (e.g., including reviews) between the on- and offline channels. We explore a retailer data set that allows isolating the “mere ESL effect”, as the retailer’s pricing strategy remained unchanged over the introduction of ESL (i.e., no dynamic pricing), but the presentation of the price and product information was integrated through ESL. A difference-in-difference analysis establishes that revenue in product categories in which ESL was introduced grows at the expense of those product categories in which it was not introduced. Visitor numbers are not affected by introducing ESL. This finding supports the adoption of e-commerce capabilities in a brick-and-mortar store as it could help prevent shopper behavior aimed at exploiting channel differences (i.e., showrooming for price or more information).
Marketers know the majority of impactful product and brand recommendations happen in offline conversations. To become part of these conversations, marketers increasingly utilize thousands of nano-influencers---ordinary consumers who “buzz” about a brand---with specialized marketing formats: seeded marketing campaigns (SMCs). The idea is to equip a large number of nano-influencers with the product or brand to be marketed, then encourage them to share their brand experiences and excitement with their peers. Overall, the created seed activity from SMCs has been shown to increase sales (Dost et al. 2019, Chae et al. 2017, Godes and Mayzlin 2009) at comparably low cost, which makes SMCs an attractive marketing mix addition for managers.
The key challenge with SMCs, and a major difference to marketing campaigns with
professional online influencers, is that nano-influencer activity is not contractually specified, and managers cannot economically engage with thousands of nano-influencers individually. Instead, managers have to rely on their initial campaign setup decisions to activate nano-influencers.
SMC setup comprises SMC size and selection (how many nano-influencers, and who from where to select?), SMC content (which products and incentives to send the nano-influencers?), and SMC timing (how long should the campaign run?). Managers actively seek advice from literature to improve their setup decisions; however, extant research offers them only incomplete setup recommendations. This leaves managers at a loss and complicates result- and cost-conscious campaigns set-ups.
With a unique dataset of 148 SMCs comprising the activity of nearly 700,000 nanoinfluencers and covering about a twentieth of all SMCs in the EU during data collection, we comprehensively model how setup decisions affect all relevant campaign outcomes---seed activity, sales, cost, and return-on-investment---while considering marketing and brand context.
Our cross-campaign data enables a comparison of various SMC setups to find out which of managers' available setup decisions aid in planning effective and/or efficient SMCs.
Results show that campaign success lies mostly in managers' hands, as setup decisions show the highest impact on campaign outcomes. Brand context plays only a minor role---contrasting organic word-of-mouth settings. Marketing context, such as parallel advertising or product price, matter differently, depending on the campaign goal. We identify trade-offs between effective brand building (maximum seed activity), effective promotion (maximum
sales), and efficient (maximum return-on-investment) campaigns, reconciling conflicting suggestions from extant research. Our model predicts that managers could increase seed activity by 97%, sales by 162%, or return-on-investment by 100% compared to current practice.
Promoting Price Discounts across Channels: The Role of Discount Level and Product Sales Frequency
(2023)
The most common promotional activity are price discounts, and retailers have to choose whether to further promote their price discounts with ads in offline (e.g., print) or online (e.g., banner ads) channels. However, retail managers lack guidance for which promotional channel (offline or online) would support price discounts best and what types of products (high or low sales frequency) would benefit most. Using a field experiment, we disentangle the interacting effects of price discounts (at various discount levels) and their supporting promotional channel. We find that digi-tal promotions of price discounts are more effective than non-digital print campaigns to increase sales beyond the base price discount effect. In addition, the product’s sales frequency matters: rela-tively, digital promotions best support price-discounted low-sales-frequency products, and steeply discounted high-sales-frequency products receive additional support from offline ads.
Bike-sharing is a popular component of sustainable urban mobility. It requires anticipatory planning, e.g. of terminal locations and inventory, to balance expected demand and capacity. However, external factors such as extreme weather or glitches in public transport, can cause demand to deviate from baseline levels. Identifying such outliers keeps historic data reliable and improves forecasts. In this paper we show how outliers can be identified by clustering terminals and applying a functional depth analysis. We apply our analysis techniques to the Washington D.C. Capital Bikeshare data set as the running example throughout the paper, but our methodology is general by design. Furthermore, we offer an array of meaningful visualisations to communicate findings and highlight patterns in demand. Last but not least, we formulate managerial recommendations on how to use both the demand forecast and the identified outliers in the bike-sharing planning process.
Challenging the Location Paradigm: Parsimoniously Predicting Store Performance with Urban Scaling
(2021)
Location is considered the most important driver of retail store performance; hence, retailers invest in extensive location research, utilizing expensive rich data. This research proposes an alternative, singular, freely obtainable predictor for location potentials. Drawing on the system scaling literature outside of marketing, we suggest that measures of the urban scale in a store’s trading area are a substitute for a multitude of traditional location measures, such as demographic or socio-economic variables. We demonstrate that our scale measure, the route factor calculated from road map data, performs on par with a common set of traditional predictors in a large dataset of supermarket sales. Moreover, our theory correctly predicts, and the analysis shows, a collinearity problem that has gone unrecognized in traditional store performance models. We validate our approach on a second variety-store dataset that covers a wider range of location conditions for generalizability.
Many empirical studies filter participants (e.g., for incorrect attention checks or quick re-sponses), especially when using participant pools such as Amazon MTurk. Yet, there is no consensus on whether and how to filter. This might originate from different perspectives on filtering participants: it may be evaluated positively (e.g., as it might be necessary to prevent inattentive participants from biasing results) or negatively (e.g., as it may enable p-hacking). This research aims to bridge these opposites: first, we empirically compare the effects of different filters and filter levels on validity, reliability, power and effects sizes of the results.
Second, we introduce the Filter Curve and our R-package “FiltR” as a means to recognize filtering which might be used to p-hack results. We suggest that filtering is not per se bad – although some filters decrease reliability and validity – but that researchers should be trans-parent in how sensitive results are for different filter combinations.
Revenue management strongly relies on accurate forecasts. Thus, when extraordinary events cause outlier demand, revenue management systems need to recognise this and adapt both forecast and controls. Many passenger transport service providers, such as railways and airlines, control the sale of tickets through revenue management. State-of-the-art systems in these industries rely on analyst expertise to identify outlier demand both online (within the booking horizon) and offline (in hindsight). So far, little research focuses on automating and evaluating the detection of outlier demand in this context. To remedy this, we propose a novel approach, which detects outliers using functional data analysis in combination with time series extrapolation. We evaluate the approach in a simulation framework, which generates outliers by varying the demand model. The results show that functional outlier detection yields better detection rates than alternative approaches for both online and offline analyses. Depending on the category of outliers, extrapolation further increases online detection performance. We also apply the procedure to a set of empirical data to demonstrate its practical implications. By evaluating the full feedback-driven system of forecast and optimisation, we generate insight on the asymmetric effects of positive and negative demand outliers. We show that identifying instances of outlier demand and adjusting the forecast in a timely fashion substantially increases revenue compared to what is earned when ignoring outliers.
Fluent Contextual Image Backgrounds Enhance Mental Imagery and Evaluations of Experience Products
(2018)
Recent findings show that individuals aged 55 and older use more and more online media and digital technologies; however, this may be biased as only those using these media and technologies may be reacting to the surveys documenting the proliferation of digital media. This might be particularly pronounced for older people affected by one or several divides, such as the urban-rural digital divide, economic divides, or the grey divide between young and old demographics. Here, we design a project to survey them—with the help of a citizen science approach in several small German cities. Results show fewer online and digital information sources and communication channels are used than expected, except for messengers. Interestingly, there is substantial heterogeneity within the older group, as adults aged 52 to 74 and those aged 74 to 85 differ in media use and preference.
Information Systems research continues to rely on survey participants from crowdsourcing platforms (e.g., Amazon MTurk). Satisficing behavior of these survey participants may reduce attention and threaten validity. To address this, the current research paradigm mandates excluding participants through filtering heuristics (e.g.,
time, instructional manipulation checks). Yet, both the selection of the filter and the filtering threshold are not standardized. This flexibility may lead to suboptimal filtering and potentially “p-hacking”, as researchers can pick the most “successful” filter. This research is the first to tests a comprehensive set of established and new filters against key metrics (validity, reliability, effect size, power). Additionally, we introduce a multivariate machine learning approach to identify inattentive participants. We find that while filtering heuristics require high filter levels (33% or 66% of participants), machine
learning filters are often superior, especially at lower filter levels. Their “black box” character may also help prevent strategic filtering.
Do true cost campaigns (TCCs)—which display prices at the point of purchase that include social and environmental negative externalities—nudge consumers toward more expensive sustainable products? From a theoretical point of view, the answer is promising: Communicating true costs means introducing external reference prices that provide a benchmark for consumers to assess price acceptability. Showing true costs triggers a general reference to the price of sustainability, and the higher price of sustainable products becomes at least partially explained by their lower “hidden costs” (i.e., costs to compensate for all environmental and social impacts). In two empirical studies, we demonstrate that for TCCs to be effective, the hidden costs for the sustainable products must be lower than those for the conventional alternatives. Interestingly, under this condition, TCCs have an effect in markets characterized by a larger (study 1) and a smaller (study 2) green gap. In both studies, we find that increased perceived price fairness explains the effect of TCCs, as measured by the relative preference for the sustainable compared to the conventional product. In addition, we see that the price difference between the two products plays a significant role in forming this preference judgment, independent of other factors included in the model and especially independent of TCC.
Many research problems are characterized by complex relationships between time series variables, such as simultaneity (e.g., feedback loops between communication channels) and state-dependence (e.g., marketing interactions with observed and unobserved sales channel variables). We introduce empirical dynamic models (EDM) to management science and marketing research. EDM is a nonlinear methodology that helps researchers to investigate simultaneous (i.e., bidirectional and same-period) and state-dependent (i.e., nonlinear and interacting) relationships with aggregate time series data. The authors demonstrate EDM capabilities and boundaries within the challenging omnichannel case. To study omnichannel systems, researchers often must rely on aggregate data: Despite more individual tracking, the data is often not available for offline channels or comprehensively integrated across channels. A simulation study, that derives aggregate time series from an individual data generation mechanism, explores conditions and boundaries under which EDM is suitable for identifying relationships between variables (i.e. in the Granger sense), predicting variable evolution (i.e. in the Machine Learning sense), and attributing marginal effects from one variable on another (i.e. in the Neyman-Rubin sense). We benchmark EDM against vector autoregression, regression, and machine learning models and provide application criteria for EDM. Next, we confirm the capabilities of EDM in an empirical investigation of interrelated brick-and-mortar, online, and mobile channels from a large European fashion retailer, finding evidence for mostly synergetic but strongly state-dependent relationships among the channels.
Empirical Dynamic Modelling for Exploring Complex Time Series in Management and Marketing Research
(2022)
Many research problems are characterized by complex relationships between time series variables, such as simultaneity (e.g., feedback loops between communication channels) and state-dependence (e.g., marketing interactions with observed and unobserved sales channel variables). The authors introduce empirical dynamic models (EDM) to management and marketing research. EDM is a nonlinear methodology that helps researchers to investigate simultaneous (i.e., bidirectional and same-period) and state-dependent (i.e., nonlinear and interacting) relationships with aggregate time series data. The authors demonstrate EDM capabilities and boundaries within the challenging omnichannel case. To study omnichannel systems, researchers often must rely on aggregate data: Despite more individual tracking, the data is often not available for offline channels or comprehensively integrated across channels. A simulation study, that derives aggregate time series from an individual data generation mechanism, explores conditions and boundaries under which EDM is suitable for identifying, predicting and attributing relationships between variables. We benchmark EDM against vector autoregression, regression, and machine learning models and provide application criteria for EDM. Next, the authors confirm the capabilities of EDM in an empirical investigation of interrelated brick-and-mortar, online, and mobile channels from a large European fashion retailer, finding evidence for mostly synergetic but strongly state-dependent relationships among the channels.
Location is an impactful but irrevocable driver of retail store performance. Unless retailers rely on their gut feelings for finding high potential locations, they have to invest in extensive location research, calibrating performance models on expensive rich data (e.g., income or education of households in each prospective trading area). To prevent “the death of the high street”, also public administrators care for location potentials. This research proposes a parsimonious new model for location potentials, drawing from emerging urban scaling literature outside of marketing. We show that a measure of the local urban scale explains stores’ sales, local competitive intensity, and defining aspects of store lifecycles (managers’ location choice, sales ramp-up to a steady state after opening, store closure). We demonstrate these capabilities of the scaling approach using six datasets, including data from two retail chains (grocery and variety stores), public data, map data, and an experiment with retail managers. Our parsimonious model compares well to more complex multivariate benchmarks and remains more robust across modeling choices. We put forth a scale measure that can be cheaply obtained from map data, offering accessible applications for retail and public policy managers (e.g., “heat maps” across all potential locations in a city) and to marketing research in general (e.g., as input or control variable for geo or mobile marketing).
In Zeiten von Inflation, stagnierenden Umsätzen und intensivem globalen Wettbewerb müssen Unternehmen sich von ihren Konkurrenten abheben, um Wettbewerbsvorteile zu erzielen . Ein interessanter Ansatz hierbei ist der neu entstandene Pick Your Price (PYP) Mechanismus, der ähnlich wie das bekannte Pay What You Want (PWYW) als partizipativer Preisbildungsmechanismus gilt und den Käufern die Möglichkeit g ibt den Preis massiv zu beeinflussen Käufer bevorzugen eine aktive Beteiligung am Preisbildungsprozess anstelle des passiven Akzeptierens fester Preise. Eine Teilnahme der Käufer an diesem Prozess erhöht zudem ihre Wahrnehmung von Fairness und Zufriedenheit. Allerdings stellt sich die Frage, welcher Preismechanismus Festpreis, PWYP oder PYP von Unternehmen genutzt werden sollte, um die Kaufabsicht und erwarteten Zahlungen zu optimieren. Um diese Frage zu beantworten, wurde ein Online Experiment durchgeführt und mittels ANOVA und Mediationsanalyse ausgewertet. Die se Studie vergl eicht d ie genannten Strategien anhand von Produkten mit unterschiedlichen Preisniveaus (Deodorant, Smartphone, Auto). Ziel war es, die Auswirkungen der Preismechanismen auf die Kaufabsicht, die erwarteten Zahlungen und die Wahrnehmung der Käufer in Bezug auf Kontrolle , A ufwand und Fairness, während des Kaufprozesses zu bewerten. Die Ergebnisse zeigen, dass PWYW die höchsten Werte für Preiskontrolle, kognitiven Aufwand und Kaufabsicht liefert. Die höchste Zahlungserwartung wurde hingegen für PYP ermittelt. Darüber hinaus zeigt diese Stud ie, dass die wahrgenommene Kontrolle die Kaufabsicht und die erwartete Zahlung am stärksten beeinflusst , gefolgt von
Aufwand und Fairness. Weiterhin werden Erkenntnisse für Forscher und Praktiker diskutiert.
Social media has become an integral part of Internet usage, with billions of users worldwide. Therefore, it is increasingly important for marketers to understand how to optimize one of the most important variables for success: Engagement. This study examines the relationship between user attention and engagement on social media, considering the effect of caption text length and emojis. Using an eye-tracking experiment, we measure user attention and engagement (likes, clicks, comments) for 161 social media posts. Our results confirm that attention affects engagement. Furthermore, we show that text length and emojis moderate this relationship. This study makes an important contribution to understanding one of the most important mechanisms for driving engagement.
Transport service providers, such as airlines and railways, often use revenue manage-ment to control offers and demand in mobility networks. Such systems rely on accu-rate demand forecasts as input for the underlying optimisation models. When changes in the market place cause demand to deviate from the expected values, revenue man-agement controls no longer fit for the resulting outliers. Analysts can intervene if they deem the demand forecast to be inaccurate. However, existing research on judgemen-tal forecasting highlights fallibility and bias when human decision makers are not sys-tematically supported in such tasks. This motivates the need for automated alerts to highlight outliers and thereby support analysts.
Network effects complicate the problem of detecting outlier demand in practice. Pas-sengers often book travel itineraries that stretch across multiple legs of a network. Thereby, they requests products that require multiple resources – seats on several legs. In the rail and long-distance coach industries, the large number of possible itiner-aries makes it likely for outlier demand to affect multiple legs. At the same time, outli-er demand from a single itinerary may be difficult to recognise as it mixes with regular demand on individual legs. To support outlier detection in transport networks, we pre-sent a method to aggregates outlier detection across highly correlated legs. We propose to use the results of this analysis to construct a ranked alert list that can support ana-lyst decision making. We show that by aggregating we are able to improve detection performance compared to considering each leg in isolation.
Network effects complicate transport demand forecasting in general and outlier detection in particular. For example, a sudden increase in demand for a specific destination will not only affect the legs arriving at that destination, but also connected legs nearby in the network.
Network effects are particularly strong when service providers, such as railway or coach companies, offer many multi-leg itineraries. In such situations, automated alerts can help analysts to adjust demand forecasts and enable reliable planning.
In this presentation, we outline a novel two-step method for automatically detecting outlying demand from transportation network bookings. The first step clusters network legs according to the observed booking patterns. The second step identifies outliers within each cluster to create a ranked alert list of affected legs. We illustrate the method using empirical data obtained from Deutsche Bahn. In addition, we present a detailed simulation study that quantifies the improvement from the clustering step and implications of ranking to measure the criticality of the outliers. Our results show that the proposed approach outperforms independently analysing each leg, especially in highly connected networks where most passengers book multi-leg itineraries.
Bidirectional Links between Aggregate Advertising and Goods Consumption at the National Level
(2024)
A longstanding question in marketing and economics is whether all of advertising affects aggregate consumption. If it does not, the advertising–consumption system would be a zero-sum game; if it does, advertising can grow markets and represents an economic force of interest for marketing, economics, and public policy. Ashley, Granger, and Schmalensee (1980) test advertising’s potential causal impact on aggregate goods consumption. Challenging prior empirical studies that suggested advertising drives aggregate consumption (Taylor & Weiserbs, 1972), Ashley et al. show that advertising does not influence consumption; rather, consumption causes advertising in a landmark methodological study of the Granger causality test. Yet recent conceptualizations suggests this empirical result and especially the linear autoregression-based Granger method to be inconclusive. Wilkie and Moore (1999) propose a complex aggregate systems view to discussing macro impacts of marketing, and in such systems the Granger tests typically fails (Sugihara et al. 2012). The present research extends prior studies with a new more appropriate nonlinear method (convergence cross-mapping) to test the causal links between television or digital advertising and three types of aggregate consumption—durables, non-durables, and service consumption. The novel test, unlike the Granger causality test, confirms the existence of a bidirectional causal link between advertising and aggregate goods consumption. Specifically, and similar to the brand level, advertising effects on consumption are stronger for durable than for non-durable goods.
Consumer decision-making is influenced by user-generated information in reviews of products and services. We predict that sales levels are associated with two forms, respectively, of information inconsistency (confliction and competition) and consistency (confirmation and complementary) often found in product and service reviews. For our investigation, we classified information contained in tablet computer reviews as conflicting, competing, confirming, or complementing using BERT (Bidirectional Encoder Representations from Transformers) and word2vec (Word to Vector) techniques. Estimating an aggregated-level sales rank model over a 24-week period on reviews and sales data obtained from Amazon, we show that variations in tablet sales can be partially explained with our information classification. These findings provide new insights for user-generated content research.
This paper presents an automated approach for providing ranked lists of outliers in observed demand to support analysts in network revenue management. Such network revenue management, e.g. for railway itineraries, needs accurate demand forecasts. However, demand outliers across or in parts of a network complicate accurate demand forecasting, and the network structure makes such demand outliers hard to detect. We propose a two-step approach combining clustering with functional outlier detection to identify outlying demand from network bookings observed on the leg level. The first step clusters legs to appropriately partition and pools booking patterns. The second step identifies outliers within each cluster and uses a novel aggregation method across legs to create a ranked alert list of affected instances. Our method outperforms analyses that consider leg data without regard for network implications and offers a computationally efficient alternative to storing and analysing all data on the itinerary level, especially in highly-connected networks where most customers book multi-leg products. A simulation study demonstrates the robustness of the approach and quantifies the potential revenue benefits from adjusting demand forecasts for offer optimisation. Finally, we illustrate the applicability based on empirical data obtained from Deutsche Bahn.
Many behavioral heuristics are fast and frugal ways for buyers to cope with uncertainty. Therefore, the extent to which pricing can rely on heuristic consumer reactions depends on the specifics and the extent of uncertainties in the market. In markets with a high degree of commoditization, uncertainties about the product are by definition low. However, as buyers are also uncertain about their own preferences, behavioral approaches can explain consumer reactions to price, even in a context of commoditization.
In this chapter, we propose and test a dual process framework for pricing around buyers’ reservation prices. Such a dual process includes both rational and heuristic modes of choice. According to our framework, reactions to prices are more or less influenced by heuristic choice, depending on their relation to a buyers’ reservation price. This framework can explain the interrelations between rational and heuristic modes of choice, willingness-to-pay ranges, and latitudes of price acceptance. We also base a set of pricing and combined price-and-communication strategies on this framework. One interesting implication is that these pricing strategies are specific to target groups among buyers: as preferences and reservation prices are heterogeneous even in commodity markets, there are different potential buyers with a more rational or more heuristic mode of choice, for every price imaginable.
How can the pricing scope be further leveraged in times of increased price transparency and growing price awareness on the
consumer side? To answer this question, this paper uses the concept of willingness-to-pay ranges (as opposed to points). Three quantitative studies show that various marketing activities that allow consumers to assess a product on a more abstract (less concrete) level shift the upper limits of the intervals upwards and thus increase the scope for price setting. These (price) upper limits are particularly high when the central product advantages are emphasized.