case study

T E A C H I N G C A S E

Understanding the value and organizational implications of big data analytics: the case of AUDI AG

Christian Dremel1 • Jochen Wulf1 • Annegret Maier2 • Walter Brenner1

Published online: 13 April 2018

� Association for Information Technology Trust 2018

Abstract ‘‘Understanding the value and organizational

implications of big data analytics: the case of AUDI AG’’

presents the case of AUDI AG and its attempts to

implement big data analytics in its organization. The case

highlights the situation of an original equipment manu-

facturer (OEM) in the automotive industry and the

potentials and challenges the emerging technology big

data analytics may entail for such organizations. The case

tries to help students to grasp the technical characteris-

tics, the value, and organizational implications of big

data analytics as well as the distinct types of analytics

services. The case is presented through the eyes of

Hortensie, an aspiring manager at AUDI, who gained

strong interest in the phenomenon of big data analytics

and received the task to position it within AUDI. To

ramp up the topic big data analytics, AUDI is engaging

with industry and design experts as well as an external

consultancy ITConsult.

Keywords Big data analytics � Organizational adoption � Organizational change � Organizational transformation �

Organizational benefits � Predictive analytics � Descriptive analytics � Analytics services � Teaching case

Introduction

On a rainy day in autumn 2014 Hortensie woke up with

only one thought on her mind. 1

Today, was her big day.

She had to present the use cases and value potential of big

data analytics in front of the chief of sales and marketing

(CMO), which she had elaborated with her team over the

last work-intensive months. Nicolas Moreau—the new

CMO at AUDI—not only was known for his positive

attitude toward innovativeness but also for his ability to

find any weakness of potential ideas. He was one of the

persons she did not want to disappoint.

She remembered how everything had started: Soon after

she had become manager at AUDI’s sales and marketing

department, Nicolas joined the company. Coming from one

of the haute écoles of Paris, he had gained experiences as

CMO at Renault and at PSA Peugeot Citroën. He had the

mission to put AUDI ahead in regard to profit, earnings,

and, first and foremost, innovativeness. When one of the

first tasks of Nicolas was the elaboration of the value and

the organizational implications of the emerging technology

big data analytics for AUDI’s sales and marketing

department, Hortensie had willingly accepted the position

as lead of the task force.

& Christian Dremel [email protected]

Jochen Wulf

[email protected]

Walter Brenner

[email protected]

1 Institute of Information Management, University of St.

Gallen, Mueller-Friedberg-Strasse 8, 9000 St. Gallen,

Switzerland

2 Audi AG, 85045 Ingolstadt, Germany

1 This illustrative case is developed on the basis of a longitudinal

case study with AUDI AG (see Dremel et al. 2017). However, due to

reasons of confidentiality, descriptions of the organization’s inner

processes, organizational hierarchies, names, and roles are anon-

ymized. Any views and statements expressed within this teaching

case do not necessarily reflect the views or policies of any individual

or the organization represented by this case.

J Info Technol Teach Cases (2018) 8:126–138

https://doi.org/10.1057/s41266-018-0036-8

She first had heard of the potential of big data analytics

in the article ‘‘Data Scientists: The Sexiest Job of the 21st

Century’’ in the October Issue of Harvard Business Review

in 2012. Since then, she learned about anecdotal evidences

of companies profiting from big data analytics such as

Netflix through their challenge for improving the predic-

tion accuracy of future movie ratings depending of a cus-

tomer’s movie preferences priced with 1,000,000 USD, the

Oakland Athletics as described in the book Moneyball by

Michael Lewis and in the same named movie, or LinkedIn

with the People You May Know feature, which had sug-

gested to her an old friend from her studies at the London

School of Economics just last month.

However, AUDI was a company known for its precision

and quality in building cars for the premium segment as

well as their innovative engineering progress, but not for

their data scientists on-site. So, she never thought of the

possibilities of big data analytics at AUDI. But now, with

the assignment of Nicolas, her thinking changed funda-

mentally. An assignment of the CMO not only meant a

huge responsibility but also a huge commitment of one of

the board members and thus power of persuasion. Hence,

she thought it will be easy to get some data, to try some

analytical scenarios in the context of AUDI, and to put

together a nice slide deck proving the value of big data

analytics. After all, she knew that at least one business unit

at the sales and marketing department was successfully

using analytics, in particular data mining, to improve the

product feature combinations in the car configurator.

AUDI’s slogan is ‘‘Truth in Engineering,’’ which is

well established in the corporate culture and brand image.

Consistent with this slogan, the company aims at further

extending its market leadership by leveraging digital

technology to provide superior products and services to its

customers. Being among the top in its market segment,

AUDI aims not only to differentiate products by innova-

tion from competitors but also to stay competitive

investing in new technologies (see Dremel et al. 2017).

To do so, AUDI AG heavily invests in emerging tech-

nologies to improve its core product, the car, for the profit

of the company. In 2015, AUDI shipped more than 2

million luxury cars to the customers worldwide. Origi-

nally established in 1909 by August Horch, the company

was acquired by Volkswagen in 1966. Headquartered in

Ingolstadt, Germany, it has been operating under the

AUDI name since 1985.

However, the traditional business of AUDI was attacked

by traditional car manufactures such as Daimler, BMW,

Volvo, by innovative market entrants such as Tesla, Fara-

day Future, and by tech-giants like Google or Apple, as

well as Uber and other companies providing innovative

mobility services.

Unraveling big data analytics

The first day of the project, Hortensie had a meeting with

the team members: Matthias, a newly hired employee with

a background in statistics and math and an affinity to data

manipulation, Tobias, an employee with 10 years AUDI

background mainly active in a multitude of projects to

improve AUDI’s retail, and Nadine who worked for

8 years in the marketing strategy department at AUDI.

Before the meeting started, Hortensie remembered one

last explanation of Nicolas, when she was assigned with the

new task (see Exhibit 1): ‘‘I briefly discussed with the other

board members whether they would be willing to invest in

a unified data organization, which could leverage big data

analytics along all departments. You know, though every-

one is thinking of this topic as an interesting one, they want

to minimize the risk to invest budget without any additional

profits or cost reductions. We, as the most innovative

department, will at first elaborate use cases for sales and

marketing alone.’’ She had read the article ‘‘How Smart,

Connected Products are Transforming Companies’’ of

Porter and Heppelmann (2015) last Monday and reflected

since then about a unified data organization. She really

liked the idea to leverage data throughout the whole

company. However, if Nicolas had already made up his

mind and discussed this point with the other board mem-

bers, she could invest her efforts elsewhere.

She abounded her thoughts and started the meeting by

asking one simple question: ‘‘As you all know, Nicolas

gave us the task to elaborate scenarios for big data analytics

at AUDI. But what does big data analytics mean concep-

tually? Is it just the visualization of data? Can we distin-

guish distinct types, supposed there are any?’’ Suddenly,

the whole room was filled with silence. No one in this room

had asked themselves this question before this meeting.

After a brief period of time, however, Matthias started: ‘‘In

my opinion, big data analytics is not just the visualization

of data—if you provide services, which are just visualizing

data, they are reporting services. For analytics, you need at

least an analytical model, which derives causalities within

data, may it be a model examining the present, the past, or

the future.’’ Tobias looked rattled and said: ‘‘And what

about all the tasks I had to coordinate to get a visualization

of our customers configuring their cars in our car config-

urator as well as the technology stack we have to pay every

month to our external provider?’’ Matthias briefly thought

about Tobias’ comment and drew a short image illustrating

the distinct types of analytics on the backboard (see Fig. 1).

Referring to Delen and Demirkan (2013) he explained:

‘‘If you simplify big data analytics services, you can dis-

tinguish descriptive and predictive analytics. Both require a

technological infrastructure, integrated data sources, and of

Understanding the value and organizational implications of big data analytics: the case of… 127

course the visualization of data. This holds also true for

reporting services. However, a descriptive analytics service

possibly explains the past or the presence (i.e., what hap-

pened or what is currently happening and why is it most

probably happening) with the help of an explanatory ana-

lytical model. A predictive analytics service on the other

hand looks into the future using an additional predictive

analytical model. Based on historical data and with the

input of current data it explains or extrapolates what will

happen and why it will happen. However, both require a

certain business acumen since, without any business

questions, no answers can be given through any model. Of

course, reporting services require business acumen as well,

but for the sake of simplicity I neglect it here.’’

Moreover, Tobias added, that you have to consider the

technological characteristics of big data itself as well

because they will pose a challenge regarding the techno-

logical infrastructure (see Table 1).

In particular, initial projects had shown that a car pos-

sibly sends 500 signals per second. Thus, car data will

result in immense high volumes of data requiring appro-

priate big data analytics technologies to enable the appro-

priate analysis. In this context, not every technical system

within every car used the same formats and names for the

same data points resulting in a variety of formats. More-

over, AUDI as recognized brand and through their expen-

sive advertisement videos, for instance, the commercials in

super bowls, generated quite a buzz of data in social media,

which not always had the desired amount of veracity.

Following this meeting, Hortensie looked up the article

‘‘Data, information and analytics as services’’ of Delen and

Demirkan (2013), which Tobias had given her. Soon she

realized, that Tobias had not mentioned a last distinct type

of big data analytics ‘‘prescriptive analytics.’’ Whereas

‘‘predictive analytics’’ used data, text, and media mining as

well as forecasting, ‘‘prescriptive analytics’’ uses either

optimization or simulation or decision modeling to suggest

which action a decision maker should perform based on the

analyzed data. She wondered why Tobias had missed out

this type but realized soon that developing prescriptive

analytics services would be something they could consider

in the long run, but right now the technological infras-

tructure constituting of the technology stack and integrated

data sources did not allow prescriptive analytics. More-

over, Hortensie realized that the future organizational unit

will have to deliver analytics-as-a-service. Thus, the

insights of analytics services will have to be accessible

through a standardized interface such as an analytics

platform.

Bringing big data analytics to AUDI

Soon after, Tobias and Matthias had identified an external

service provider who was one of the leading IT consul-

tancies. After their successful pitch at a meeting with

Hortensie and the other team members, this consultancy

had the mandate not only to help identify potential use

cases for AUDI’s sales and marketing department but

also to identify potential future work models as well as

organizational implications to implement those use cases in

the future in collaboration with the task force. After 2

Fig. 1 The distinct types of analytics services at AUDI AG

128 C. Dremel et al.

months and several interviews, Stefan, the senior consul-

tant of ITConsult, called Hortensie late on a Thursday

afternoon. After some small talk, he started: ‘‘Currently,

AUDI is contacting the customer through a multitude of

contact points such as the website, the car configurator,

their dealers, and their car. The interactions of AUDI with

its customer produces data that allows to identify customer

desires, to elaborate their preferences, and behavior.

However, you currently miss on leveraging the potentials

of data-driven marketing because of either no integration of

relevant data sources or a lack in the competence to do so.

Please have a look at Fig. 2, you can see how, on a very

generic level, a potential unit could address possible cus-

tomers, such as AUDI’s business units, dealers, and

importers in collaboration with an analytics partner, who is

capable of conducting the entire process of data connec-

tion, exploration, analysis, and visualization.’’

He, further noted that, first, an analytics team or a

loosely coupled analytical competence center must be

implemented to provide analytics-as-a-service. How the

implementation takes place is highly depending on how

strategically and centralized the sales and marketing

department wants to pursue big data analytics. Moreover, it

will most probably demand quite some investment for the

sales and marketing department. In this context, he handed

her Exhibits 2, 3, and 4. The following Monday Hortensie

had a meeting with Nicolas. Hortensie showed him the

exhibits: ‘‘Exhibits 2 and 3 illustrate approaches for an

organizational embedding of big data analytics, or, more

precisely, of the competencies and capacity we need to

succeed in big data analytics through a new subsidiary.

This subsidiary, as the strategy department informed me,

could not only support big data analytics use cases but also

new digital business models and services. Exhibit 4 on the

other hand represents the embedding in our existing sub-

sidiary InnovativeCar. 2

That way we would minimize our

investment, however, this means also that InnovativeCar as

supporter of our engineering departments is taken over

tasks, which were traditionally in the hand of the sales and

marketing. Exhibit 3 assigns the data analytics unit to the

strategy department and thus ensures cross-departmental

responsibility. Exhibit 2, however, clearly prioritizes sales

and marketing related big data analytics topics.’’

2 InnovativeCar was specifically created to support new innovative

technologies and concepts for cars (e.g., autonomous driving and

electric mobility).

Table 1 Technological characteristics of big data

Characteristic Description Exemplary source at AUDI

Volume High volumes of unstructured, volatile and heterogeneous data

enable a company to broadly generate insights, for instance

on customer sentiments. The sheer amount of data exceeds

the ability of traditional business intelligence systems to

process this data

With over 100 sensors, a car produces up to 25 GB of data per

hour

Variety Variety stands for the various formats of data resulting from

the many data sources that are often unstructured and

inconsistent in their nature. The usage of further data sources

increases the variety of data and thus the complexity to

analyze these data points. Big data characterizes the shift

away from predefined categorizations and data schemas

A connected car needs to communicate with multiple external

data sources (such as traffic lights or other cars) as well as

with the driver (for instance via speech recognition)

Velocity Velocity states that big data is produced in high speed

requiring real-time analysis to achieve decisive insights

A connected car needs to respond to external conditions in

real-time (e.g. critical traffic situations)

Veracity Data sources like social media produce data that carries no

single truth and thus requires big data technologies to assess

data accuracy

AUDI needs to interpret social media buzz, to filter out

deceptive data, and to interpret ambiguous statements

Fig. 2 Conceptual work model of an analytical competence center

Understanding the value and organizational implications of big data analytics: the case of… 129

After a while Nicolas explained, that his strategy team

had already planned to create a new, more agile company

that would be able to develop, design, and operate digital

services and their business models. In his opinion, that

might be one solution where synergies could help not only

the development of digital services through big data ana-

lytics but also vice versa. He said: ‘‘I think one point is

crucial. Every future digital service will not only require

data, may it be data from the customer, the car, marketing

agency, weather agency, and so on, but also produce data

on its own. The more we can use big data analytics to our

advantage, the better will our services be and thus our

value proposition. That is why I want to implement ana-

lytics not only in a new company but also in my sales and

marketing department, and of course we should not miss

out our colleagues at the IT department. However, I still

would like to start step-by-step. This is such a new topic. I

am not only talking about a re-organization but also about

required skill sets. We are at the very beginning of

implementing a data-oriented mindset at AUDI. This holds

true for digitization but also for big data analytics.’’

Hortensie, replied: ‘‘Alright, in this case it would make

most sense to go with Exhibit 2 and assign it to the con-

nected retail unit. Here, we are already trying to improve

our retail through data analysis, for instance a targeted

marketing approach for new models, the optimization of

car feature combinations, and feature usage analyses.’’

Nicolas said: ‘‘Sounds great to me. Maybe, you could

already start with a pilot project to support our introduction

of the upcoming e-tron model of the A3 in Germany.’’

Hortensie left partially happy and partially confused. Was

she the one who would be the manager of this analytics

unit? Nicolas had not said it explicitly but implicitly. The

next day, she received the official mail making her the

manager of the data intelligence subunit in the connected

retail unit. The start of a long journey.

The next day, Nadine came to Hortensie’s office after

Hortensie had explained the decision to her team in a

previous 2-h meeting: ‘‘Looks like we will have to build up

some new competencies and skills in analytics. However, I

am not sure whether I like it to engage with external

agencies. This would mean that right from the start we

become dependent on external agencies.’’ Tobias replied:

‘‘Though I think you are absolutely right, we have no

choice but to work with consultancies till we have our

subsidiary Analytics GmbH which will ensure the techno-

logical and analytical competencies through their data

scientists, big data architects, and visualization experts.’’

Hortensie had already thought about bringing in external

expertise. However, she knew that this will be a chal-

lenging task as it meant to implicitly state that right now

the IT department and her team cannot provide the tech-

nological and analytical expertise required to succeed.

However, the creation of the subsidiary was not planned

before 2015. After one discussion with Tobias and having

reflected on Fig. 1, Hortensie was sure that enough busi-

ness acumen was already available at the sales and mar-

keting department. In regard to analytical skills as well as

some technological tasks, however, they fell short in

capacity, skills, and, most of all, experience. Hence, she

and her team had to collaborate with external agencies with

the goal to substitute the agencies with the innovative

subsidiary in the medium term. So, at first Hortensie’s team

and the agencies developed pilot uses cases with Tableau

as well as SPSS Modeler as first tools. Soon, they found

pilot customers with whom they carried out first projects in

the sales and marketing department.

Finally, when the subsidiary was created, Hortensie tried

to sketch how the data analytics unit at sales and market-

ing, the IT department, and most importantly the subsidiary

Analytics GmbH would collaborate and work together:

‘‘Internally, it’s a cross-functional team consisting of our

unit, the Analytics GmbH, and the IT department to ensure

the required capabilities. The data scientists and big data

architects need to work with database engineers, the busi-

ness owner, and people from the sales and marketing

department who are advocating for the end user. It will take

a lot of collaboration.’’

Achieving commitment for big data analytics

The respective parties had to adopt an attitude that values

data along with a data-specific technical infrastructure.

AUDI needed to institutionalize a practice of sharing data

in a standardized way across teams, whether it will be sales

or manufacturing, and integrate it in a central database.

These standards, along with technical systems for ware-

housing and organizing data, had to be established as soon

as possible and facilitated through a defined change man-

agement plan.

To elaborate use cases, industry experts, designers, and

AUDI representatives collaboratively conducted an inno-

vation workshop. Hortensie and her team was excited about

the atmosphere of innovation during these days. However,

most of the potential use cases that were identified would

only be possible in the future when some homework such

as setting up a technology stack and integrating all required

data sources would have been done. Nevertheless, the data

analytics unit started to think about the key steps to derive

insights (see Fig. 3).

In between the discussion rounds, a guest speaker from

General Electric mentioned: ‘‘GE makes large machines

like jet engines and locomotives. GE realized that these big

machines are just commodities and the value to their cus-

tomers lies in telling them more about the machine: What’s

130 C. Dremel et al.

wrong with the machine so they can maintain it more

efficiently. How the machine is performing so its perfor-

mance can be increased. When the machine will need

maintenance so they can schedule the downtime. What

they do is put sensors on the machine and what I do is look

at all this data coming in, look at the people who are going

to use that data, find out what metrics are useful for them,

and present it in a way that makes sense for them.’’

Hortensie reflected on this interesting comment—of course

AUDI was as well producing some kind of machine that

the customers are using to get from point A to point B.

Thus, the similarities to this comment were obvious to her.

Another participant mentioned: ‘‘Oftentimes data means

uncertainty. That’s the biggest source of hesitancy in larger

companies. People ask: Is this is a science experiment?

That is a blanket term, with a little bit more understanding

and empathy, companies could actually differentiate

between a moon shot and what actually makes very clear

sense.’’

Hortensie summarized the workshop: ‘‘Both the creative

teams and industry experts emphasized the importance of

first defining and understanding the end user. Data should

be thought of as adding value for the customer: as some-

thing that might be given back to customers as a mean-

ingful service. In order to engage with users, especially as

concerns over data privacy and ownership grow, human-

centered design methods and a focus on user experience

should drive the development of new products and ser-

vices. Your shared advice included activating emotion,

drawing on both convention and novelty, and empowering

users as active participants.’’

This workshop resulted in a multitude of ideas for new

use cases, for instance the optimization of marketing

approaches based on sociodemographic information and

customer sales data or the optimal planning of sales num-

bers based on historic information. Moreover, the teams

had created a quick proof-of-concept for a reporting service

that describes the usage of AUDI connect (see Table 2).

As a starting point, the business analytics unit developed

a pilot use case for the service MicroTargeting (see

Table 3). To do so, the respective data were gathered in the

form of a data snapshot, because the technological infras-

tructure had not yet been set up. The data included internal

data from AUDI AG and importers (e.g., purchase history

and car specifications of the analyzed car) and were enri-

ched with external data (e.g., sociodemographics, socio-

geographics, behavioral variables, innovation affinity, and

price sensitivity). Afterward, the data were visualized with

the help of a visualization software to elaborate the data

richness and quality. Next, an analytical model was

developed and, in a second step, an analysis of the data

screenshot took place using clustering analysis to group

similar customers. At last, the results were visualized a

second time in a customer-specific dashboard.

Although the first pilot projects led to an initial com-

mitment for big data analytics, the unit did not manage to

create solutions that could be leveraged in a standardized

way across all countries. One major reason was the data-

centered development of services. Based on available data,

use cases were identified with one pilot customer. The

interest of all potential customers was not required. How-

ever, the first pilot cases needed data from the sales and

marketing business units. Although, of course, big data

Fig. 3 Deriving value from AUDI’s data

Understanding the value and organizational implications of big data analytics: the case of… 131

analytics would be used for the benefit of the whole

company, Hortensie and her team struggled to get access to

relevant data sources due to a lack of understanding of the

benefit of big data analytics, power issues as well as

departmental boundaries. For instance, Nadine had to col-

lect the data of A3 customers to develop in collaboration

with ITConsult the service MicroTargeting. In a meeting,

she received the answer: ‘‘Sure, we have the data of pre-

vious AUDI A3 customers, but not only we do not have

any statement of our management to share the data but also

I do not get the point how your service should help our

customer targeting at all.’’

Equipped with statistics of the pilot use cases, she

replied: ‘‘We improved our targeting by 10% in Spain and

our project in France shows comparable results. Contacting

customers based on big data analytics instead of just

sending any customer advertisement is what Premium car

manufacturing is about. Or do you want to lose any cus-

tomer just because we are not able to collaborate and share

data?’’ This was not the first time she had to talk straight to

get access to data, and she knew it would not be the last.

That moment she remembered how Stefan had explained to

her: ‘‘AUDI is currently too much characterized by parties

that try to improve the business of their single unit instead

of improving the whole company. Big data analytics,

however, as technology innovation requires a mindset of

data-sharing.’’

Hortensie received quite good feedback for this proof of

concept: ‘‘Microtargeting is like a good sales man in my

dealership…if we have a look at the fluctuation it‘s of real high value to have data-based evidence standardized and

storable.’’ The positive response to MicroTargeting led to

further awareness for data analytics at the sales and mar-

keting department on executive, managerial, and operative

levels. That way business units at the sales and marketing

were more willingly cooperating with Hortensie’s team and

sharing their data and ideas. This resulted in further follow-

up projects requiring both Hortensie’s team and the Ana-

lytics GmbH to hire more employees.

Table 2 The analytics service ‘‘AUDI connect usage’’ (proof-of-concept)

Illustration

Description

This service reports the usage of AUDI connect on a country level. It helps to analyze and plan the coverage of AUDI connect services.

132 C. Dremel et al.

Table 3 The analytics service MicroTargeting (proof-of-concept)

Illustration

Description

The analytics service MicroTargeting is engineered as a combination of a descriptive analytics service and the visualization of additional data. Its goal is the description of the demographic distribution of customers for a specific car model (e.g., A3 e-tron) to sharpen the marketing approach by better addressing the target group. The service was, specifically, developed to gain knowledge of the customer base and their need for certain products. The Microtargeting dashboard allowed for in-depth analysis of potential at both area and customer level. Initially, this service was created because of a decision on executive board level to improve the targeting of potential A3 e-tron customers in Spain.

Challenge

Depending on each market new market specific data had to be gathered and were not always available in the right quality and richness. Consequently, having built the analytical model the biggest and continuous challenge when developing and delivering the MicroTargeting service was the acquisition of market specific data. Therefore, the effort for delivering the service to a new customer did not decrease but stayed instead on the same level. Moreover, the variety of data sources in the different countries of AUDI importers resulted in country-specific solutions, which is against the paradigm of analytics-as-a-service.

Understanding the value and organizational implications of big data analytics: the case of… 133

Achieving the value of big data analytics through scrum

With the growing importance of Hortensie’s unit and big

data analytics at AUDI, the number of projects rose and the

fulfillment of each project in time was near to impossible.

Matthias, who was planned to act as program manager for the

analytics initiative, proposed to try out the agile software

development method scrum to use the resources available in

the best way and to prioritize the resources according to the

strategic relevance for the success of each project. He made a

one pager explaining scrum in general and the roles needed to

develop analytics services (see Table 4). After his explana-

tions, Hortensie realized working with big data necessitates a

variety of design methodologies and approaches including

design thinking and agile software development. She

deemed agile software development and, in particular, scrum

as appropriate because it ensures that requirements and

solutions evolve through collaboration between self-orga-

nizing, cross-functional teams, and it promotes evolutionary

development, prompt delivery, and continuous improve-

ment. Moreover, it encourages rapid and flexible response to

change.

Having the subsidary with analytical and visualization

expertise in place, Hortensie let her team implement the

scrum process to follow up on additional valuable pilot

cases such as stock time reduction of used cars, a detailed

analysis of AUDI connect, prediction of repurchases, and

identification of factors influencing the selling of cars.

Besides others, the service Demand Analysis was devel-

oped with the help of the newly implemented scrum pro-

cess (see Table 5).

With the success of MicroTargeting and Demand

Analysis, Hortensie and her team were increasingly

Table 4 Developing analytics services using scrum

Definition The basic notion of scrum is to deliver incremental value for the product in an iterative way. It is reflected by the basic unit of development: the sprint. Each sprint has a length of two weeks and starts with a sprint planning, which defines the scope of work in the form of user-stories that has to be done in the iteration. All team members update each other daily about their progress in a 15min daily scrum call and once per sprint the backlog refinement and retrospective are performed. Each sprint ends with a sprint review in which the development team presents the work results to the stakeholders. Due to its cross-functional nature the scrum team incrementally builds the full analytics service by ensuring the integration of the data into the data warehouse, the design and implementation of the necessary analytics to the handover of the analytics service to the regular IT operations. The scrum process helps as well to solve the challenge of evolving requirements for analytics services as those new requirements can be implemented incrementally in a new iteration.

Roles and Relations

• Product owner: Staffed by the AUDI sales and marketing department, the product owner bundles the requests and needs of the stakeholders (Business Owners) and formulates them in the form of user stories. He prioritizes those user stories in the product backlog and ensures that features and milestones for the individual projects are met. In parallel the product owner acts as link between the program steering board, the external business owners and the development team.

• Development team: The cross-functional development team consists of members of the AUDI IT and Analytics GmbH and spans competences from operations, data handling/integration, analytics to UX and visualization skills.

• Scrum master: Externally staffed, the scrum master ensures that the regular scrum process is followed and helps the development team to successfully remove impediments to delivering the product goals. He puts a focus on continuous internal improvements of the team.

Definition The basic notion of scrum is to deliver incremental value for the product in an iterative way. It is reflected by the basic unit of development: the sprint. Each sprint has a length of two weeks and starts with a sprint planning, which defines the scope of work in the form of user-stories that has to be done in the iteration. All team members update each other daily about their progress in a 15min daily scrum call and once per sprint the backlog refinement and retrospective are performed. Each sprint ends with a sprint review in which the development team presents the work results to the stakeholders. Due to its cross-functional nature the scrum team incrementally builds the full analytics service by ensuring the integration of the data into the data warehouse, the design and implementation of the necessary analytics to the handover of the analytics service to the regular IT operations. The scrum process helps as well to solve the challenge of evolving requirements for analytics services as those new requirements can be implemented incrementally in a new iteration.

Roles and Relations

• Product owner: Staffed by the AUDI sales and marketing department, the product owner bundles the requests and needs of the stakeholders (Business Owners) and formulates them in the form of user stories. He prioritizes those user stories in the product backlog and ensures that features and milestones for the individual projects are met. In parallel the product owner acts as link between the program steering board, the external business owners and the development team.

• Development team: The cross-functional development team consists of members of the AUDI IT and Analytics GmbH and spans competences from operations, data handling/integration, analytics to UX and visualization skills.

• Scrum master: Externally staffed, the scrum master ensures that the regular scrum process is followed and helps the development team to successfully remove impediments to delivering the product goals. He puts a focus on continuous internal improvements of the team.

134 C. Dremel et al.

confronted with how to manage all the data sources

available at AUDI. Tobias had the task to illustrate the

most important data sources depending on their hidden

business value, capturing the eventual case that an analysis

with this data would become possible. He reasoned, that

the more the data are related to the core product of AUDI—

the car—the more the value rises as AUDI as company is

the only organization that can actually access this kind of

data if they act within privacy and legal guidelines. Con-

sequently, he sent Table 6 to Hortensie.

Presenting Table 6 Tobias explained: ‘‘Car data is by far

the most valuable data source we have right now. However, it

is not only the most valuable but also the most difficult one to

transfer to our big data landscape as our cars are always

moving and the data itself is highly critical to privacy.

Customer and web data on the other hand is mostly available

Table 5 The analytics service demand analysis (proof-of-concept)

Illustration

Description

From its early roots, the analytics service Demand Analysis was designed as predictive analytics service. Its goal was to anticipate order entries for 3 months in advance. This helped to optimize the spending of media budgets as well as sales commissions based on data-driven analysis instead of experience and gut-feeling. The service used sophisticated analysis methods (e.g., machine learning and neural networks).

Challenge

Developing an analytical model to predict the future order entry posed two major challenges. First, data that could be used from AUDI systems was not always complete. Second, the effects that elevated previous order entries were not always known and were thus not correctly interpreted by the analytical model.

Understanding the value and organizational implications of big data analytics: the case of… 135

right now and from a privacy standpoint we have already

performed all measures necessary to stay within legal and

privacy guidelines. As additional data sources 3rd party data

is easily purchasable and from a privacy standpoint already

anonymized enabling us an easy handling.’’

Looking at this table, Hortensie reflected whether

Tobias did not miss something. Sure, the car was the most

important and valuable data source, but she could not

abandon the thought of a missing data source. However,

these data sources were the most important ones for the

sales and marketing department right now. Looking back,

Hortensie was surprised what had happened during the

recent months. Nicolas had enjoyed the use cases she had

presented with her team, she was now responsible for big

data analytics at the sales and marketing department, and

the first pilot projects to enable the transfer of data from

every car model had been started. Yet she knew there was

still a long way ahead to turn AUDI into a digital car

company.

Assignment questions and teaching note

The students are supposed to take the role of the task force

in order to decide which organizational implementation of

big data analytics is probably the best solution for AUDI.

The project team has to identify the advantages and dis-

advantages of the diverse potential of implementing big

data analytics as well as to identify the most valuable use

case for an OEM in the automotive sector.

For the preparation of the class discussion, the students

are recommended to prepare the following assignments:

1. Please describe, which technological characteristics

are commonly used to describe big data. Which

distinct types of analytics services exist? Please,

provide one example for each type how AUDI used

this type of analytics services within this case.

2. Following Davenport and Harris (2007, p. 118), Porter

and Heppelmann (2015), and Ross et al. (2013), should

AUDI follow up on a strategic path to implement big

data analytics or use a step-by-step approach champi-

oned by one business department (e.g., sales and

marketing) to prove and implement big data analytics?

3. Please elaborate which of the provided Exhibits (1, 2, 3,

and 4) would have been the best solution for AUDI? Why?

4. Which data sources of AUDI have been identified?

Which are the most valuable ones and why? Please,

provide at least one example for each data source of

how AUDI uses or could use this data source.

5. Is big data analytics, a buzzword or an emerging and

sustainable technology? Which trends in our daily life

as customers might affect the importance of big data

analytics?

6. How is the company going to progress through the

analytics ladder? Which obstacles does the company

encounter and how have those challenges been resolved?

If you are a bona fide instructor, please contact the first

author of this teaching case to receive a copy of the sup-

plementary teaching note.

Table 6 Data sources and their value

Data source Data Value

Car data Car model-code

Car identification number

Purchasing date

Registration date

Car model description

Model year

Engine displacement

Engine power

Very high

Customer/web data Customer ID

Salutation

Birth-date

Type of customer (business; private, small commercial)

Full address

Number of car the customer owns

Car configurator data

Traffic data from website

High

Third party data Sociodemographic data

Social media data

Medium

136 C. Dremel et al.

Appendix

See Exhibits 1, 2, 3 and 4.

Exhibit 1 Organization chart for a unified data organization

Exhibit 2 Organization chart analytics GmbH in collaboration with sales and marketing

Exhibit 3 Organization chart analytics GmbH in collaboration with strategy department

Understanding the value and organizational implications of big data analytics: the case of… 137

References

Davenport, T.H., and J.G. Harris. 2007. Competing on Analytics: The

New Science of Winning. Harvard Business Press.

Delen, D., and H. Demirkan. 2013. Data, Information and Analytics

as Services. Decision Support Systems 55 (1): 359–363.

Dremel, C., Herterich, M., Wulf, J., Waizmann, J.-C., and W.

Brenner. 2017. How AUDI AG Established Big Data Analytics

in Its Digital Transformation. MIS Quarterly Executive 16 (2):

81–100.

Porter, M.E., and J.E. Heppelmann. 2015. How Smart, Connected

Products are Transforming Companies. Harvard Business

Review 93 (10): 96–114.

Ross, J.W., Beath, C.M., and A. Quaadgras. 2013. You May Not Need

Big Data After All. Harvard Business Review 91 (12): 90–98.

Christian Dremel is a research associate at the Institute of Information Management (IWI-HSG) at the University of St.Gallen.

He holds an M.Sc. from the University of Bamberg. In collaboration

with AUDI AG his research focuses on the successful adoption and

assimilation of big data analytics. In particular, he investigates the

organizational transformations (e.g., organizational structures, gover-

nance mechanisms, and capabilities) required to profit from big data

analytics. His research has been published in journals such as the MIS

Quarterly Executives (MISQE) and presented at conferences such as

the International Conference on Information Systems (ICIS).

Jochen Wulf is lecturer and fellow of the International Postdoctoral Fellowship program at the University of St.Gallen. Prior to this, he

was assistant professor at the Institute of Information Management at

University of St.Gallen (IWI-HSG), Switzerland. His research focuses

on socio-technical systems and large-scale data processing systems,

consumer-centricity, and IT service management. Jochen authored

more than 50 scientific publications. His research has been published

in journals such as Business & Information Systems Engineering

(BISE) and Electronic Markets (EM), and presented at conferences

such as International Conference on Information Systems (ICIS) and

European Conference on Information Systems (ECIS).

Annegret Maier is responsible within AUDI AG for the field of Data Strategy and Analytics. Along with her team, she pushes for the usage

of data analytics to support business decisions and how data can be

used to enable personalized digital experiences for the users and

customers throughout all online touchpoints worldwide. The strategic

direction of data usage within Audi, with a special focus on the

connected vehicle and car sensor information, falls into her

responsibility.

Walter Brenner is professor joined St.Gallen University in 2001 after having held chairs at the University of Essen (Germany) and

Freiberg University of Mining and Technology (Germany). He

received a graduate degree in business administration and a doctorate

from the University of St.Gallen. His research focuses on information

management, consumer data, innovation and digital industrial

services. He has authored and edited 30 books and more than 300

publications. Brenner also practices as a consultant and is an

entrepreneur. Prior to joining academia, he was Head of Application

Development at Alusuisse-Lonza AG (Switzerland).

Exhibit 4 Organization chart existing subsidiary InnovativeCar in collaboration with strategy department

138 C. Dremel et al.