CAPSTONE PROJECT

DATA STRATEGY
AGENDA

1. INTRODUCTION
2. Understand the context
3. Define your THE case
4. Generate insights from
data
5. Manage and govern data
6. Select the right
technology
7. LEVERAGE DATA SCIENCE AND
ANALYTICS
8. Build a data-driven
culture
9. Next steps
1. Introduction
BUSINESS SOLUTION
BUSINESS SOLUTION FOR POINT OF SALE
INTRODUCTION

Our company provides its

highly sophisticated, full
cloud-based Business
Solution.
Including Point of Sales,
Retail Automation,
Merchandising/Stock
Management, Backoffice,
Fuel, Carwash, Loyalty, etc.
with a huge variety of
interfaces and web services
for retail, fuel, and
hospitality.
 Mobile Workplace

Dispenser Management
Price Polls
Fuel Management / Digital Signage
Fuel Delivery
Car Wash Information
systems
Customer Flow
Analysis
Motion Logic

Promotions / Electronic Shelf

Campaigns Labeling

Self Check-out POS

Click&Collect Mobile POS
POS
Self Check-out
Loyalty
Terminals Smartphone App

Mobile Connected Cars

Merchandise Management
Delivery of Goods Data Entry
Sales Based Ordering

Payment Options Tank Management

Back office Site

Supplier Management
2. Understanding the
context
WHAT DATA DO WE HAVE, AND HOW IS IT
BEING USED
DATA OVERVIEW
UNDERSTANDING THE CONTEXT

A LOT OF DATA Business

solution
▪ Our company has access to a lot and all kinds of Social data Support
data. media data data
▪ The green data (green data bubbles on the right)
are all related to our business solution which is Corporate
used by our customers to do their own business Ms azure
website
with their customers (consumers). data
data DATA
▪ The grey data (grey data bubbles on the right)
are “internal” data, that means this data is
collected in interaction with our customer and is Market
Corporate
pricing
used for administrative purposes (e.g. billing) or data
data
for marketing (e.g. campaigns, product
management). End-
Customer
consumer
▪ In this capstone project the focus is set on the data
data
BUSINESS SOLUTION DATA.
BUSINESS SOLUTION DATA (extract)
UNDERSTANDING THE CONTEXT

LICENSE usage CONTENT

▪ Sessions (concurrent, SaaS) ▪ Logins/logouts ▪ Setting/ ▪ User’s files
▪ Plans (features, interfaces, etc.) ▪ Menu/buttons- configuration ▪ Communication
function ▪ Location
selection ▪ User rights
▪ Used features
▪ Text search
queries

DEVICE, CONNECTIVITY, CONFIGURATION INTEREST AND FAVORITES Error reports & performance
▪ Used hardware ▪ Preferred program language ▪ Performance problems
▪ Operating system ▪ Favorites in menu bars ▪ Type and severity of problems
▪ Browser version ▪ Details of software or hardware
▪ IP address related to an error
▪ Device identifiers (e.g. IMEI) ▪ Content of files using when an error
occurred
▪ Regional and language settings
▪ Data about other software on device
3. DEFINE THE USE CASE
WHO ARE WE TARGETING FOR WHAT
INTERACTION, AND WHAT BUSINESS
RESULTS CAN BE ACHIEVED
USE CASE
DEFINE THE USE CASE

“focus on usage data of the user’s

interaction with the solution”
In this use case I focus on usage data of the user’s
interaction with the business solution (logins, “OPTIMIZING USER
concurrent usage, duration, clicks, etc.).
For this use case there are a lot of "clean" and EXPERIENCE!”
complete data based on log-files/-data; data is
accessible already today; it is real-time data that
allows to see the result of every change and
optimization immediately. We own the data.
This data will help to improve the performance and
usability and user experience by e.g. analyzing and
optimizing the clicks to start a function or to find a
product.
4. GENERATE INSIGHTS FROM
DATA
WHAT INSIGHTS COULD EXISTING OR NEW
DATA GENERATE FOR OUR USERS
 Insights
GENERATE INSIGHTS FROM DATA

USER BEHAVIOR
User behavior analytics usually helps enterprises
detect insider threats, targeted attacks and financial
fraud.
These insights we will also get in our data project.
But we like to focus on data that gives us insights
regarding user experience as well as performance
and stability of our solution.
As a first step this use case will lead to improvement
of operations and improvement of user experience
and efficiency: we will learn more about the user
behavior and user interaction with the solution.
As a second step this new knowledge will lead to
offering additional licenses and plans, and to
12 increasing the margin (optimizing/changing licensing
model).
5. Manage and govern data
HOW WE PLAN TO MANAGE AND GOVERN
THE DATA EFFECTIVELY
GOVERNANCE (1/3)
MANAGE AND GOVERN DATA

Data governance team

As we work through setting up our data governance
program, aligning the efforts of the data governance
team to tangible business value, we will be focusing
on solving specific problems. This will ground the
team and reinforce the perception of them as
enablers and accelerators of ongoing data projects,
rather than having their work viewed as
administrative process overhead.
We will keep sponsors (member of C-Level) updated
and bring stakeholders in gradually.
We will start small and focus on the data and data
attributes that are critical for success. It’s likely we
won’t have all the answers up front, so starting
OUR MOTTO:
small allows us to continuously learn and improve
and more easily curate the engagement.
“Think Big, Start Small,
and Scale Fast!”
GOVERNANCE (2/3)
MANAGE AND GOVERN DATA

Initial steps
As we learned in this course, we will clarify the
following subjects first:
(1) Data Management Maturity
(2) Capacity to Change
(3) Collaborative Readiness
(4) Business Alignment
(5) Data Quality
(6) Regulatory Compliance
We will use the opportunities that we’ve identified
tied to business value, build a roadmap to show how OUR MOTTO:
our initial small initiatives fit into the longer-term
vision at scale. By leveraging a series of small early “focus on business
successes, the experience and confidence will gain
during that time and will enable us to scale fast. value!”
GOVERNANCE (3/3)
MANAGE AND GOVERN DATA

Goal of governance
The goal is not to have the most comprehensive
program but the most used.
We will make content engaging and easy to
understand and tooling easy to use to keep people
regularly involved and participating.
Data governance is not a one-time thing. It is an
ongoing program that could involve people across
various business functions.

OUR MOTTO:
“easy and involving!”
6. Select the right
technology
WHAT TECHNOLOGY OR TECHNOLGOIES
COULD BE USED TO HELP CREATE VALUE FOR
OUR USERS
TECHNOLOGY (1/4)
SELECT THE RIGHT TECHNOLOGY

Our solution is used in various industries by various types of users and

for various purposes. Taking into account the countless variables as
well, the amount of data is massive and the agile business needs of the
future makes the support of technology necessary and helps to be
more efficient and effective.

The size, complexity, and distributed nature of our data, combined with
increasing requirements for continuous intelligence and speed of
action, mean the manual data governance and management practices
of the past can never keep up with the agile business needs of the
future. The underlying technology foundation supporting execution of
the data strategy requires AI powered automation to scale.

Therefore, it is very helpful and necessary to have machine learning in

place, where the algorithms will build a model in order to make
predictions or decisions without being explicitly programmed to do so.
TECHNOLOGY (2/4)
SELECT THE RIGHT TECHNOLOGY

Data analytics
In our industry we mainly see PowerBi and Salesforce. As we are a very
Microsoft oriented company, PowerBi would be the first choice.

As far as I understand by now, PowerBi itself is perfect for descriptive

data analysis.

For the purpose of predictive and for sure for prescriptive analysis we
would add the machine learning service of Microsoft AZURE. I'm not yet
there to fully understand this service, but it looks promising for our
purpose.
TECHNOLOGY (3/4)
SELECT THE RIGHT TECHNOLOGY

EXTRACT, TRANSFORM, AND LOAD AT SCALE

We need a centralized, analytics-grade data warehouse that supports
technical and non-technical teams. At this point it should be obvious
that we need all our data, from every source, connected seamlessly
and available for instant insight. Broadly, this means that we need a few
things:

(1) The ability to get data out of places and into some other central
place.

(2) The ability to clean, connect, store, manage, and maintain all this
data in a single unified place.

(3) The ability to export this data into the places where we’ll use it, for
both technical and non-technical teams.
TECHNOLOGY (4/4)
SELECT THE RIGHT TECHNOLOGY

EXTRACT, TRANSFORM, AND LOAD AT SCALE

With Azure HDInsight, a wide variety of Apache Hadoop environment
components support ETL at scale. The use of HDInsight in the ETL
process is summarized by this pipeline:

Source: [Link]
7. LEVERAGE DATA SCIENCE
AND ANALYTICS
WHAT ROLE COULD DATA SCIENCE AND
ANALYTICS PLAY IN INFORMING BUSINESS
DECISION
GENERATE VALUE
LEVERAGE DATA SCIENCE AND ANALYTICS

SEIZE NEW OPPORTUNITIES

The new opportunity is understanding the underlying data and create (business) value out of the
data. So far, our company has a lot of data, but unstructured and not properly accessible to the
business. The business needs easy, clear data and information to make decisions. If the business
needs to do additional calculations by itself or the data is not fully accessible, the business will not
use the data or make wrong decisions.
Data Science and Analytics will help to provide the right information based on the right data to be
able to make fast and right decisions.
In particular, predictive forecasting or classification will help our business to predict future behavior
of the solution/system and customers/users and therefore support performance and usability.
8. BUILD A DATA-DRIVEN
CULTURE
HOW WE WILL BUILD A CULTURE TO ENABLE
DATA-DRIVEN DECISION-MAKING
3 MAJOR GOLDEN RULES
BUILD A DATA-DRIVEN CULTURE

LEADING BY EXAMPLE HELP EMLOYEES HELP DATA STEWARDS

The management team THEMSELVES Data Stewards can train
needs to lead by The employees need to the teams and explain/
example. This motivates understand the benefit answer concrete
the rest of the team to of data and how to deal questions they may
understand the with data. They need to have. This support is
importance of data and learn how to generate needed to get the
follow the example. value out of data and teams "grounded" in
how to deal with the the data culture and
tools. enable them profit from
data.
9. NEXT STEPS
WHAT ARE THE NEXT STEPS WE CAN TAKE
TO MOVE THIS STRATEGY FORWARD
NEXT STEPS
NEXT STEPS
(1) Map opportunities against business value and organizational
readiness
(2) Get sponsor and stakeholder commitment and participation
(3) Set up a data governance program
(4) Cataloging and cleansing to ensure complete and accurate data,
which impacts the accuracy of analytics activities
(5) Automate and scale with technology
(6) Keep the momentum
While it might be tempting to start with an opportunity that may offer
the most potential value, careful consideration must be given to
whether it is highly complex or takes a great deal of effort. It’s better to
start with a project that is less involved, even if the payoff may be
smaller. Getting a few wins under OUR belt quickly will help build
credibility for the program. Having that momentum is critical for the
longer-term success.
THANK YOU
DATA STRATEGY