PreprintPDF Available

Digital Technologies for Pricing Problems-A case study on increasing the level of digitization at a leading German retail company

Authors:
Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

In many parts of the retail sector, the market structure is oligopolistic which leads to high interdependencies of the decision parameters (price, product, promotion, placement). Still, the analysis of interactions and the resulting automation of decisions has not yet progressed much. The ever-increasing intra-and inter-competition, increased costs, and a lower differentiation margin have led to the need to rethink the low priority of these parameters at one of Germany's largest retailers. The presented case study focuses on a research and consulting project on the digitization of price management. In an initial step, the transparency within the software systems was increased by making the competitors' prices available. This was followed by increasing automatization using the store's individual competitive factors. In a final step, a proof of concept was sketched to cope with the immense amount of data and the complexity of algorithms by using in-memory database technology. Within the project, it became clear that digitalization is a continuous process, as the company is not fundamentally at risk in the situation under consideration. Moreover, digitization is not independent of the degree of digitization that has already been achieved. The consulting project showed major results: about 25% of the total manual workload was reduced and an overall margin improvement of 0.2 percentage points could be achieved.
Content may be subject to copyright.
Digital Technologies for Pricing Problems - A case study on increasing the level of
digitization at a leading German retail company
Felix WEBER
Institute for Computer Science and Business Information Systems, University of Duisburg-Essen
Essen, NRW, Germany
Reinhard SCHÜTTE
Institute for Computer Science and Business Information Systems, University of Duisburg-Essen
Essen, NRW, Germany
ABSTRACT
In many parts of the retail sector, the market structure is
oligopolistic which leads to high interdependencies of the
decision parameters (price, product, promotion, placement). Still,
the analysis of interactions and the resulting automation of
decisions has not yet progressed much. The ever-increasing intra-
and inter-competition, increased costs, and a lower
differentiation margin have led to the need to re-think the low
priority of these parameters at one of Germany’s largest retailers.
The presented case study focuses on a research and consulting
project on the digitization of price management. In an initial step,
the transparency within the software systems was increased by
making the competitors’ prices available. This was followed by
increasing automatization using the store’s individual
competitive factors. In a final step, a proof of concept was
sketched to cope with the immense amount of data and the
complexity of algorithms by using in-memory database
technology. Within the project, it became clear that digitalization
is a continuous process, as the company is not fundamentally at
risk in the situation under consideration. Moreover, digitization
is not independent of the degree of digitization that has already
been achieved. The consulting project showed major results:
about 25% of the total manual workload was reduced and an
overall margin improvement of 0.2 percentage points could be
achieved.
Keywords: digitalization, retail, pricing, price management,
digital technologies
1. INTRODUCTION
Increasing competition within retailing
The German retail sector is characterized by an oligopolistic
market with strong intra-competition between existing retailers
and rising inter-competition between traditional and new “pure”
digital players [1]. The focus of this case study lies in the retail
sector of food retailing, in particular.
Through the withdrawal of Kaiser’s Tengelmann in 2016 and the
looming market entry of Amazon with Amazon Fresh, the
competition has intensified. Furthermore, a waning scope for
differentiation between operating types [2], increased costs, an
overall increased price awareness [3], and the influence of the
company's price image on the customer’s choice for a retail chain
can be observed.
Information technology and its impact on the way we
compete
Another major development within the retail domain is expected
to be driven by technological innovations. Benson-Armer et al.
[4] predicted that 3D printing, the Internet of Things, advanced
robotics, big data, advanced marketing, the mobile world, and
artificial intelligence will be primary digital technologies that
will have a high impact on the consumer industry by 2030.
Utilization of the above-mentioned technological innovations
leads to a situation where data that was formerly not available
becomes available. With this development, not only the
manageable data volume, but also the execution capabilities of
the application systems must keep up [5]. These different
developmental trends have led to the need to rethink the
traditionally low priority factors within the domain of retailing of
process automatization, analytical optimization approaches, and
data analyses.
This article locates in the context of the scientific domain of trade
marketing, which is defined as the process of analysis, target
formulation, strategy selection and composition and control of
the marketing mix in a trading enterprise [6, 7]. In the retail
sector, marketing has the task of developing concepts that are
market-leading and thus provide secure competitive advantages
that are sustainable [8]. Marketing has always been one of the
most important areas of activity and a central influencing factor
in the retail sector [7, 9]. However, due to the rise of competitive
pressure, its importance has increased even further, forcing retail
companies to use all available marketing tools. Particularly
noteworthy in this respect is the price, which represents the
dominant marketing instrument in the competition between retail
companies today [10].
Digitization is also taking place in this area, and there is generally
much more data available in the required form than there was
previously. As can be seen in the following text, this includes the
availability of competitive prices within the system, the
availability of prices from direct and indirect competitors
collected from the Internet, the expansion of master data, and
other possible scenarios that are becoming available with the
development of the Internet of Things.
The central decisions that must be made within the scope of trade
marketing encompass the four areas of the marketing mix. The
basis is the central concept of the 4Ps by McCarthy [11],which
structures marketing into four separable components: "product",
"price", "place" and "promotion". The “product” component
includes many decisions, both operational and strategical,
regarding the assortment. The initial choice of the assortment, the
decision to list a new product, and the decision to remove an
article from the assortment are all part of the considered object
area. The “price” component includes all decisions that influence
the transaction price: the initial pricing, normal sales pricing,
promotional pricing, and markdown pricing at the end of the
product’s life cycle. Under the “place” component, all decisions
and actions of the company made in connection with the
distribution of a product or a servicefrom the initial supplier to
the consumerare considered. Relevant decisions in the context
of “promotion” include any decisions on the design and
propagation of any information relating to the product, like
corporate communication, advertising, sales promotion,
sponsoring, or public relations. With the above outlined trends
and the room to maneuver within the marketing mix in mind, the
low utilization of analytics and data-driven decision-making in
the current retail practice do not seem feasible and competitive
for the future. The purchasing department at a trading company
is responsible for defining the whole assortment, including the
decisions about adding and removing products and any related
decisions about the pricing and promotional activities. The
pricing decisions include the pricing at any stage of the product’s
lifecycle. The decisions about the promotional activities
comprise the selection of suitable products, the pricing, and the
medium advertised in. As a particularity of the German food
retail market, promotional activities are mainly implemented as
physical leaflet ads distributed to all households in a specific
catchment area.
The case study
In this case study, one of Germany’s largest retailers is sketched.
The retail group operates a range of different distribution chains,
hypermarkets, supermarkets, and discount formats. This leads to
a large assortment of widths and depths: 8,000 products at the
discount chain, 35,000 products in the regular supermarket chain,
and up to 80,000 products at the hypermarket chain. The regional
subsidiary considered in this case study operates more than 1,000
stores and has an annual turnover of more than 3 billion euros.
On a technical level, more than 1 million articles exist in the
databases (including temporary products). Any kind of manual
analyses of and decisions on articles, assortments, promotions, or
prices are not expedient. Moreover, each of the stores serves up
to 10,000 customers daily, who have an average shopping basket
size of 16 items.
2. PROJECT OBJECTIVE AND SITUATION
FACED
The projects objective
Due to the abovementioned increasing competition level, the
purchasing department responsible for the price management of
the case study company requested an extensive evaluation of the
current practices and recommendations for action. Also, the
regarded trading company is currently about to migrate all legacy
applications towards a single vendor system architecture. In light
of this ongoing development, the project also aims to improve the
outcome of this transformation project from a business
perspective. The current decision-making about the assortment,
prices, and promotions is led by “gut feeling” and rudimentary
Excel and VBA-Scripts, instead of advanced analytics. The
company recognize that this setting is not competitive in the long
run and thus, has initiated the hereinafter described project. The
use of a consulting project with a research institution was
required due to the analytical complexity and the independency
of any vendor relationships and biases.
Situation faced
The abovementioned factors, such as the huge assortment and the
fast succession of sales, multiplied by the thousands of stores
operated leads to a data volume that cannot be handled by most
analytical applications. The absence of the necessity to analyze
this data in the past and, especially, the lack of the technical
practicability [12, 13] with most of the current retail information
system architectures are the reasons why the priority within the
domain of retailing for automated and analytically-based
decisions has traditionally been low [5]. Ultimately, the
complexity imposed by the interdependences between a huge
amount of possible influencing factors, for example, the
composition of the assortment [14], the placing within the store,
the competition and their actions, the promotional activities or
the composite effects, is a major challenge as well.
For a further understanding, we first consider the business
processes that are currently applied in common retail companies
for decision-making regarding the price management. In today's
decision-making processes, the market behavior of the
competitors within an oligopoly is expressed. According to the
classical structureconductperformance (SCP) paradigm of
competition theory, the market structure and market behavior are
responsible for market performance (see also [15] for a market
definition). In an oligopoly, the existence of a market leader
operating as a price leader can often be observed [16]. Due to the
width and depth of the assortments, there may not be only one
price leader for all articles within the whole assortment.
In practice, food retailing is structured according to different
operating types. The simplest subdivision here is between full-
range stores and discount stores.
This classification is based on the aggregation of the real
manifestations with similar strategy patterns regarding the
factor’s goods, space, and personnel under one operation type.
The different forms of operation type are thus abstract from
reality but allow an analysis and observation of the market [17,
18].
Discount stores focus on price with a clear emphasis on a fast
moving but flat product range of food and near-food products
Fast Moving Consumer Goods (FMCG). This is only expanded
by a small range of non-food offered as special offers. In contrast,
the full-line stores offer a much deeper assortment of FMCG. The
assortment here also contains service elements (especially, the
service counters for sausage, cheese, and fish) as well as a large
proportion of non-food products. The full-line operating types
can further be subdivided into supermarkets, consumer markets,
and hypermarkets. This differentiation is mainly based on the
physical size and wide of the assortment. All in all, full-range
stores have a considerably greater assortment depth, but
regarding the core FMCG, their widths are not bigger. The
currently applied practices of pricing decisions are mainly
oriented alongside this structure of operation types and
assortments.
Different competitors in the market can be price leaders for
various parts of the assortment. This differentiation among parts
of the assortment results into a basic concept that theoretically
reflects the competition. For example, the German discount
retailersespecially ALDIare the price leaders for the so-
called “discount assortment” (500 to 1,000 articles).
The division into these three shopping basket sizes as a reference
is a pragmatic solution for retail companies. The classification of
price comparisons is based on the standards of an independent
service provider. This service provider collects the pricing data
and delivers the information as a service.
The justification for these three classes is to be regarded as
arbitrary for the time being. The classification is more
pragmatically oriented to reflect the three main operating types
(supermarket, consumer market, and hypermarket) which can be
mapped as a reference shopping basket structure. From the point
of view of the actual competition, however, this abstraction does
not have to be correct.
Universally, three different levels of price comparison exist:
small, medium, and large. The three groups consist of a set of
articles that arein the smallest versionalso on sale at the
German discounter and drugstore (dm-drogerie markt and
Rossmann) chains and therefore are very price sensitive. The
second group of around 1,500 articles, including the articles from
the first group, extend the consideration to articles that the
customers can also buy in competing supermarkets or
hypermarkets. The third group, representing the largest group
with 2,000 to 2,500 articles of the assortment, include all relevant
(individually defined by the retail chain itself by observing the
prices of the different competitors) articles that are comparable
for the customers. The last group, representing the remaining
assortment, are of the least attention, and the prices are mainly
the recommended retail prices of the manufacturers.
3. ACTION TAKEN
The action taken in this case study can be structured into three
phases that build on each other in terms of time and content:
transparency creation, automatization, and optimization.
As a first phase, the availability and thus, the transparency of the
prices is increased. By adding the data of the competitors’ prices
into the central ERP-System (SAP Retail), the decisions on prices
are still a manual process, but the purchaser of the retail company
can directly access the competitor's prices as well as the
anticipated price reactions of the competitors depending on the
price decision being considered. In this initial step, the system
uses manually define price strategies and operative price policies
based on these factors. For example, price reactions are manually
defined depending on the competitor. Based on a competitor’s
price, a default calculation rule automatically determines a
proposition based on price point rounding rules that the purchaser
can manually accept. In this context, a price point is defined as a
special fixed sales price to achieve a psychologic effect on the
customer, which is directly below a round figure (e.g. EUR 2.99).
This added data also lays the foundation for the further described
steps in this case study.
In a second phase, the manual process of the price decision-
making is automated. The starting position and old logic here is
that prices are set without dependence on competition for an
entire type of company. As a result, local competition from a
specific branch does not have any influence on pricing. Rather,
the store type is decisive for the calculation, irrespective of its
local characteristics.
In contrast to the initial approach, individual stores are now
assigned to specific competition groups. Depending on the
location and the competitors in the geographic neighborhood,
price ranges are defined at the product group level (see Figure 1).
With this heuristic approach, pricing is not detached from the
individual features that were previously completely abstracted.
Figure 1: Assigning individual stores to competition groups
The third phase is the ongoing implementation of an
"optimization solution". In contrast to the heuristically
approaches of the steps before, an optimization problem, in a
mathematical sense, consists of the two basic elements of an
objective function, which represent an equation or system of
equations containing the variable to be optimized and the
limitations underlying the problem [2]. The objective function
results from the possible influencing factors and effects on the
price and reactions of customers and competitors. The limitations
result from company-specific characteristics, such as the internal
cost structure, the overriding legal framework conditions, or the
restrictions resulting from the general strategy or objectives, for
example, fixed-price positioning. Currently, no analytical
method for this price optimization exists in the scientific
literature; there are only fragmentary constructs of limited
extend. Thus, there is a need for scientific and empirically
oriented research to understand the problem including all
relevant factors.
The regarded trading company is currently undergoing a
migration process to replace all legacy applications for a single
vendor (SAP) system architecture strategy. In light of this
ongoing migration, the Proof of Concept (PoC) is designed not
to build on the existing and soon to be obsolete architecture, but
rather, keeps a possible target system architecture, the SAP
reference architecture for omnichannel retailing [19, 20], in mind
(see Figure 2). SAP HANA is the underling in-memory, column-
oriented, and relational database management system [21] for all
applications. Its primary function is to act as the database server
for storing and retrieving data from the built-on top applications.
In addition to this, it is also designed to perform advanced
analytics (predictive analytics, spatial data processing, text
analytics, text search, streaming analytics, or graph data
processing) and includes an application server. This makes the
SAP HANA, as a platform, capable of serving as a fast data
source as well as an application server, providing a server-based
JavaScript application server with the SAP HANA XS Engine,
including a front-end solution based on SAPUI5, a proprietary
framework [22].
Figure 2: Proof of Concept Architecture for the Optimization of
Price Decisions
The initially collected and available data (sales, promotions,
master data, competitors’ prices, weather data, social media data)
is extracted from all the different legacy source applications,
external data providers, and beyond that, collected by hand.
With this PoC, the retail company can now conduct detailed
pricing and marketing experiments and use the resulting data
output to validate a holistic model for both price and promotion
optimization.
4. RESULTS ACHIEVED
As shown in this case study, digitization leads to an ever-
increasing interdependence of tasks [5]. At the level of the
business solution, both the process quality and the solution
quality of the underlying business problem significantly
improved. An evaluation of the achieved results is carried out in
the following text in analogy to the three defined steps:
The initial phase of creating transparency by making the
competitors’ prices available achieved, compared to the initial
situation prior to the introduction of the system for pricing policy,
a considerable reduction in the manual workload. Overall, the
workloads of the respective process steps (Figure 3) were
reduced by 25 percentresulting in annual saving of around
30,000€.
Figure 3: Monetary result of creating price transparency
Without the use of technology and the buzzwords associated with
digitization, such as IoT, Artificial Intelligence, or big data, huge
savings have already been made possible in this early phase. The
availability of basic data within the existing systems combined
with smaller technical adjustments will be responsible for a major
impact on the company’s button-line.
In the second phase, another considerable economic efficiency
effect was realized. Here, the increasing digitization of pricing
policies, by allowing the system to adjust prices for each product
group rather than for the entire product range based on a type of
store, was added. In the past, the system only allowed an
indicator at the over-reaching store type level of the kind of
pricing policy to be followed. For example, since the company
has several heterogeneous distribution channels with differing
price policy mechanisms in the market, it was not possible for
one store to sell the drugstore assortment at the pricing level of
the competition and the remaining assortment at a high price
level. This led to the situation where, on the one hand, these
stores offered prices that were too low compared to the
competition (for example, articles for the “Lidl dry goods”
category, although there was no competition in the geographical
range of the store). On the other hand, prices in other categories
were excessively high (since the “dm-drogerie markt” ranges
were offered at uncompetitive prices, where competition was
within reach). Although the degree of automation introduced
here is far behind that intuitively associated with digitization,
massive effects were evident in the improvement of the margin.
The monetary effect forecasted for this change in the calculation
logic was based on a margin improvement of 0.2 percentage
points. This represents an effect of approximately 6 million EUR
per year considering the turnover of 3 billion EUR in total per
annum.
Since the third phase is an ongoing research and implementation
project, the results cannot yet be determined in monetary key
figures. However, comparing the implemented PoC with the old
processes, a major improvement on the technical level can be
attested. The fundamental analysis of the basic decisions about
promotions and pricing can now be underpinned with analytical
information. The central application uniting the data and
information that were initially spread over several systems is also
technically capable of handling the large amount of data. A first
proposal for an overreaching data model is in place, which is also
underpinned by the target system architecture of the company.
With the above outlined process and the data made available, the
retail group can conduct big data analysis on nearly all fields of
interest in regard to promotional activities. The quality of the
business-related solution brings a major improvement by the
provided analytics and their technical implementation. With the
PoC laying the foundation for more applications of data
analytics, including all aspects of retailing from the product,
store, consumer, and competition, to a temporal development of
consumer preferences, a set of examples of the most concern is
presented in the following text.
The preliminary work done with this proof of concept, also lays
the foundation for any future employment of other digital
technologies, for example, the exploitation of new technologies
like the Internet of Things, which might be of interest for brick-
and-mortar retailers, or the implementation of new digital
businesses and digital business models. The deployment of real-
time analytics and predictive analytics combined with Multi- and
Omni-channel retailing is also one of these scenarios. The
sketched case-study also presents a step to counteract the current
low penetration of Artificial Intelligence and Machine Learning
within the domain of retailing [23].
Within price management, the use of personal and real-time
pricing, enabled by the alongside deployment of digital price tags
or customer loyalty cards, is possible, thus eliminating some of
the analytical superiority of pure online retailers, like Amazon.
5. LESSONS LEARNED
The digitization of retail companies is not independent of the
degree of digitization that has already been achieved. It is not
possible to automate decisions directly from a state of largely
manually determined decision-making processes. Within the
framework of the project, it became clear that this is a continuous
process, as the case study company is not fundamentally at risk
from digitization in the situation under consideration.
The learning effects were different in the three project phases. In
the first phase, it became clear that an understanding of the
problem is indispensable before it can lead to automation in the
following phases. A major unexpected result at this point was the
ambiguity regarding the day-to-day business of the purchasing
department of the trading company. During the several
interviews and meetings that lead to the joint development of the
target requirements, the employees in charge of making the
decisions about the above outlined fields of decision could not
adequately declare the procedures that currently lead to
decisions. The lack of being able to linguistically explain the
daily operations led to the elementary question of whether the
employees have a fundamental lack of knowledge of the
underlying decision problems that is only covered or
compensated by the validity of the used truisms and instinct, or
the awareness of a possible replacement of their manual decision-
making with a fully automated processes, followed by a personal
loss of power. Notwithstanding the above, this difficulty in
picturing the current processes and practicesirrelevant if this is
imposed by a real lack of knowledge or personal resistance
constitutes a huge challenge in capturing the current decision
problems and transferring them into an equivalent model within
the application.
In sum, the main learning effect for our research was that
companies remain socio-technical systems in which the benefits
of application systems can only be realized if the interpretation
of the data by the actors and problem-appropriate modeling in the
system are compatible with each other. The application of
advanced analytics and, even, the utilization of the fastest in-
memory database technologies are of less value without the
understanding and deep knowledge of the underlying business
model and daily operations in the domain. In particular, without
domain-specific knowledge, any interpretation of data,
regardless of the incredible amount of data points provided, is
much less expedient. When the awareness is based only on
transaction data, a decrease in the sale of an article on a specific
day would be credited to composite effects between two
products. By studying this same situation with the awareness of
the domain and process related reality, a wide range of external
factors could also be made accountable for the change of the sales
amount—a farmer’s market in the neighborhood, an in-store
tasting, or a simple out of stock situation can lead to the same
data pattern. Without the domain-specific knowledge and by
purely judging from a data scientist domain, the consequences of
this data pattern will turn out to be ultimately different.
The organizational setting used in this study was shown to have
much more influence than initially expected. Any big data
projects will most certainly involve more than the mandating
department. In the above case, at least the local retailers, who
represent a heterogeneous group of merchants, were involved as
well as the central wholesale operations, the purchase
department, the marketing department, and the IT department.
This complex organizational setting is also new to the common
research on price management. All of these stakeholders have
different data, data sources, and applications in places that cannot
be considered to be aligned in any sense. In the situation
explained above, the organizational structure of the trading
group, with separate wholesale and retail operations under one
roof, posed a unique challenge to data analytics. In this context,
the quality of data is of paramount importance for the quality of
decision-making, so the economic benefits of digitization in the
individual phases are more important in terms of quality,
consistency, and coherence. The time dependency of the master
data resulted in another major challenge in terms of distinction
towards a pure research project for analytical processes, as the
same identification number could represent two different articles
over time. The low data quality provided in this project as well
as analogs in many other retailing and trading companies,
resulted from a different set of mostly historically imposed
reasons. The first obstacle to achieving clean master data resulted
from the fact that the system architecture has grown over the
course of time, without any attempts to standardize it. However,
with a set of different systems deployed, no consistent master
data is possible, at least not without a major effort.
In the second phase, which was understood as partial automation,
it became clear how high the economic advantage can be.
Although the degree of automation here was fairly low and the
amount of data used did not correspond to the amount of data
intuitively associated with digitization, massive impacts on the
financial performance of the company in question were observed.
The aforementioned learnings, in turn, opened up the possibilities
for a much more radical digitization consideration, aimed at the
far-reaching substitution of human decision makers. This was the
subject of the third phase. On a technical level, the first
conclusion regarding the to-be implemented PoC is that the
general perception of the targeted software application
architecture based on the central in-memory database was that
the setup is feasible for the intended goals of the proof of concept
and can also power real and real-time business processes. The
infrastructure is fast enough and capable of handling the amount
of data resulting from the Point-Of-Sale transactions. Real-time
analyses are available within the SAP HANA database platform
and allow the utilization of predictive and advanced analytics.
For more mature analytical algorithms, the authors found that the
built-in functions and algorithm are much faster and should be
preferred over custom logic, whenever possible.
5. REFERENCES
[1] R. Schütte and T. Vetter, Analyse des
Digitalisierungspotentials von Handelsunternehmens. In R.
Gläß and B. Leukert, Handel 4.0., pp. 75-113, Berlin:
Springer Gabler, 2017.
[2] H. Meffert, C. Burmann and M. Kirchgeorg, Marketing:
Grundlagen marktorientierter Unternehmensführung
Konzepte - Instrumente - Praxisbeispiele, Wiesbaden:
Springer Fachmedien Wiesbaden, 2015.
[3] S. Daurer, D. Molitor and M. Spann, Digitalisierung und
Konvergenz von Online-und Offline-Welt. Zeitschrift für
Betriebswirtschaft, Vol. 82, No. 4, 2012, pp. 3-23.
[4] R. Benson-Armer, S. Noble and A. Thiel, The consumer
sector in 2030: Trends and questions to consider, 2015.
[5] R. Schütte, Information Systems for Retail Companies. In
E. Dubois and K.Pohl, Advanced Information Systems
Engineering: 29th International Conference, CAiSE
2017, Essen, Germany, June 12-16, 2017, Proceedings.,
Berlin: Springer, 2017.
[6] N. H.Borden, The concept of the marketing mix, Journal of
advertising research, 1964. Vol. 4, No. 2, pp. 2-7.
[7] S. Haller, „Handelsmarketing“. In H.-C. Weis, Modernes
Marketing für Studium und Praxis, Ludwigshafen
(Rhein): Fridrich Kiehl Verlag, 2008.
[8] H. Schröder, Handelsmarketing Methoden und
Instrumente im Einzelhandel. München: Redline
Wirtschaft, 2002
[9] L. Müller-Hagedorn, Handelsmarketing, Stuttgart:
Kohlhammer, 2002.
[10] H. Simon and M. Faßnacht, Preismanagement: Strategie -
Analyse Entscheidung Umsetzung, Wiesbaden:
Gabler. 2009.
[11] E. McCarthy, Basic Marketing: A Managerial Approach,
Indiana: Irwin, 1960.
[12] A. McAfee and E. Brynjolfsson, Big data: the management
revolution, Harvard Business Review, Vol. 90, No. 10,
2012, pp. 60-68.
[13] D. Litzinger, Computergestütztes Promotioncontrolling:
Konzeption eines Informationssystems für das
Controlling von Konsumgüterpromotions, Berlin:
Springer, 2013.
[14] ] J. Becker and R. Schütte, Handelsinformationssysteme.
Domänenorientierte Einführung in die
Wirtschaftsinformatik, Frankfurt am Main: Redline-
Wirtschaft, 2004.
[15] M. Motta, Competition policy: Theory and Practice.
Cambridge: Cambridge University Press, 2004.
[16] P. Belleflamme and M. Peitz, Industrial Organization:
Markets and Strategies, Cambridge: Cambridge
University Press, 2015.
[17] R.P. Lademann, Vollsortimentsstrategien im
LebensmitteleinzelhandelEntwicklung und Perspektiven
in einem Discountumfeld. In J. Zentes, B. Swoboda, D.
Morschett, H. Schramm-Klein, Handbuch Handel, pp.
243-257, Berlin: Springer, 2012.
[18] D. Möhlenbruch, Sortimentspolitik im Einzelhandel:
Planung und Steuerung, Berlin: Springer-Verlag, 2013.
[19] SAP. The SAP Model Company for Omnichannel Retail.
Available from:
https://www.sap.com/documents/2017/05/8ec48bce-bc7c-
0010-82c7-eda71af511fa.html [01.08.2017].
[20] I. Woesner, Retail Omnichannel Commerce Model
Company. Available from:
https://blogs.sap.com/2016/12/22/retail-omnichannel-
commerce-model-company/. [01.07.2017].
[21] H. Plattner and B. Leukert, The In-Memory Revolution:
How SAP HANA Enables Business of the Future, Berlin:
Springer, 2015.
[22] P. Prassol, In-Memory-Platform SAP HANA als Big Data-
Anwendungsplattform. In D. Fasel and A. Meier, Big Data:
Grundlagen, Systeme und Nutzungspotenziale, pp. 195-
209 Wiesbaden: Springer, 2016. pp. 195-209.
[23] F. Weber and R. Schütte (2019), A Domain-Oriented
Analysis of the Impact of Machine LearningThe Case
of Retailing. Big Data and Cognitive Computing 3(1): 11.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Information technologies in general and artifical intelligence (AI) in particular try to shift operational task away from a human actor. Machine learning (ML) is a discipline within AI that deals with learning improvement based on data. Subsequently, retailing and wholesaling, which are known for their high proportion of human work and at the same time low profit margins, can be regarded as a natural fit for the application of AI and ML tools. This article examines the current prevalence of the use of machine learning in the industry. The paper uses two disparate approaches to identify the scientific and practical state-of-the-art within the domain: a literature review on the major scientific databases and an empirical study of the 10 largest international retail companies and their adoption of ML technologies in the domain are combined with each other. This text does not present a prototype using machine learning techniques. Instead of a consideration and comparison of the particular algorythms and approaches, the underling problems and operational tasks that are elementary for the specific domain are identified. Based on a comprehensive literature review the main problem types that ML can serve, and the associated ML techniques, are evaluated. An empirical study of the 10 largest retail companies and their ML adoption shows that the practical market adoption is highly variable. The pioneers have extensively integrated applications into everyday business, while others only show a small set of early prototypes. However, some others show neither active use nor efforts to apply such a technology. Following this, a structured approach is taken to analyze the value-adding core processes of retail companies. The current scientific and practical application scenarios and possibilities are illustrated in detail. In summary, there are numerous possible applications in all areas. In particular, in areas where future forecasts and predictions are needed (like marketing or replenishment), the use of ML today is both scientifically and practically highly developed.
Chapter
Full-text available
Der Lebensmitteleinzelhandel hat sich in den letzten Jahrzehnten teilweise grundlegend verändert. Neben der Integration von Groß- und Einzelhandelsstufen führte der Wandel der Betriebsformen des Handels zu diesen Veränderungen. Unter Berücksichtigung der zwei Betriebstypen Vollsortimenter und Discounter analysiert der nachfolgende Artikel Gründe, aber auch Auswirkungen der Entwicklungen im Lebensmitteleinzelhandel. Anhand verschiedener Kriterien, wie z. B. Sortimentspolitik, Serviceleistungen oder Verkaufsfläche, werden die historische Genese und aktuelle Entwicklungen dargestellt und erläutert. Weiterhin werden die Zukunftsperspektiven der Betriebsformen im Lebensmitteleinzelhandel aufgezeigt.
Chapter
Die aktuelle Diskussion um die „Moden“1 der „Digitalisierung“ und der „Disruption“2 durch neue Technologien ist aus einer wissenschaftlichen Perspektive als kritisch zu beurteilen. Es wird mitunter der Eindruck erweckt, dass es einen kategorischen Unterschied von digitalen und „analogen“ Unternehmen gibt. Dabei werden seit Jahrzehnten Unternehmen oder Betriebe als sozio-technische Systeme in der Betriebswirtschaftslehre verstanden.3 Dieses Verständnis von Unternehmen hat auch weiterhin Bestand und es bedarf keiner Anpassung dieser Definition: es gibt kein ausschließlich digitales Unternehmen. In jedem Unternehmen sind mindestens die Anteilseigner Individuen oder Institutionen.
Digitalisierung und Konvergenz von Online-und Offline-Welt
  • S Daurer
  • D Molitor
  • M Spann
S. Daurer, D. Molitor and M. Spann, Digitalisierung und Konvergenz von Online-und Offline-Welt. Zeitschrift für Betriebswirtschaft, Vol. 82, No. 4, 2012, pp. 3-23.
The consumer sector in 2030: Trends and questions to consider
  • R Benson-Armer
  • S Noble
  • A Thiel
R. Benson-Armer, S. Noble and A. Thiel, The consumer sector in 2030: Trends and questions to consider, 2015.
Information Systems for Retail Companies
  • R Schütte
R. Schütte, Information Systems for Retail Companies. In E. Dubois and K.Pohl, Advanced Information Systems Engineering: 29th International Conference, CAiSE 2017, Essen, Germany, June 12-16, 2017, Proceedings., Berlin: Springer, 2017.
Domänenorientierte Einführung in die Wirtschaftsinformatik
  • J Becker
  • R Schütte
J. Becker and R. Schütte, Handelsinformationssysteme. Domänenorientierte Einführung in die Wirtschaftsinformatik, Frankfurt am Main: Redline-Wirtschaft, 2004.