Microsoft Fabric

Unlocking Transformative
Data Value with Microsoft Fabric
A three-phased approach

1
 Microsoft Fabric

Forward
Today’s world is awash with data—ever-streaming from thousands of services. Customers must stitch together a
the devices we use, the applications we build, and the complex set of disconnected services from multiple vendors
interactions we have. themselves and incur the costs and burdens of making
these services function together.
Organizations across every industry have harnessed
this data to digitally transform and gain competitive That’s why we’ve introduced a new innovation­­—
advantages. And now, as we enter a new era defined by AI, Microsoft Fabric—an end-to-end, human-centered
this data is becoming even more important. analytics product that brings together data and analytics
tools in one place. Fabric integrates proven technologies
Generative AI and language model services, such as like Azure Data Factory, Azure Synapse, and Microsoft
Azure OpenAI Service, are enabling customers to use Power BI into a single unified product, empowering data
and create everyday AI experiences that are reinventing and business professionals alike to unlock the potential of
how employees spend their time. Powering organization- data and lay the foundation for the era of AI.
specific AI experiences requires a constant supply of clean
data from a well-managed and highly integrated analytics The data platform for the era of AI: Introducing
system. But most organizations’ analytics systems are a Microsoft Fabric
labyrinth of specialized and disconnected services.

And it’s no wonder given the massively fragmented data

and AI technology market with hundreds of vendors and

2
Contents
Forward2 EXPERIMENT & LEARN20 BUILD & SCALE35
Current state of data, analytics, and business intelligence 4
Marketing 21 Building a foundation of alignment 36
The challenge: an organically evolved data estate 5
Operations22 Aligning with your organization’s leaders and data stewards 37
Complex data and analytics challenges  6
HR23 Conversations and questions to drive your business value
Now trending: data mesh, data fabric, data hub 7 roadmap.38
Sales 24
Modern data architectures are not mutually exclusive. Conversations and considerations to drive your
Identify technical experts and business users for the POC 25 infrastructure roadmap 40
They are collectively transformative. 8
Document clear goals for your POC and the criteria that will Mapping your current data and analytics estate 41
Introducing Microsoft Fabric – a unified analytics solution
define its success  26
for the era of AI 9
Ideate on planned transformation and outcomes 42
Using common analytics scenarios and proposed activities
A unified SaaS-based solution that stores all organizational
to guide your Proof of Concept 27 Establishing a network of transformation champions 43
data where analytics workloads operate 11
Lakehouse 28 Training your network of champions 44
Three- phased approach to unlocking data value 12
Data Warehouse 30 You are on your way to transformative data value 45
ENVISION13
Data Science 32
The art of the possible 14
Real time analytics 34
Healthcare 15

Financial Services 16

Public Sector 17

Retail 18

Sustainability/Energy 19 3
Current state of data,
analytics, and business
intelligence
Data has become the oxygen of digital • Personalized Employee Experiences
transformation • Smart Facilities
Data is now truly driving digital transformation: • Actionable Seller Intelligence
from creating delightful customer experiences and
• Personalized Support Experiences
enhancing employee productivity to generating
• Finance Intelligence
operational efficiencies. ​But it’s not just the raw data
itself. It’s the integrated insights and intelligence • Digital Security and Risk Intelligence
that can be activated across an entire connected Yet, research shows that only 32% of organizations
organization. Unlocking the value from your data reported being able to realize tangible
and bringing it into the era of AI is the key to and measurable value from data. And that is
driving innovation within your organization and where the challenges begin to reveal themselves.
for your customers. Source: Accenture

4
The challenge: an
organically evolved
data estate
The most common challenge in enterprise data on-premises data sources from different vendors
estates is that data has organically evolved. in different formats that hold critical information.
Companies have data estates with a ton of data and It is very difficult to find deep and accurate insights
infrastructure redundancy. There are often multiple without a single source of truth. Stitching together
teams operating silos of data that are not truly the unique analytics tools organizations need is
connected. The platforms required to transform data complicated. Costs associated with procuring and
into actionable insights are inherently very technical managing these capabilities can be exorbitant. And
and fragmented. Usually, teams of engineers and there is a significant risk associated with the lack
developers are needed to deploy and manage these of governance.
platforms. Organizations have many siloed cloud or

• Every analytics project has many subsystems

• Every subsystem needs a different class of product
• Products often come from multiple vendors
• Integration across products is complex, fragile and expensive
5
Complex data and
analytics challenges

45+55+Q
Organizations know that to succeed and thrive in • Difficulty in enabling ​data-driven organization
today’s constantly changing landscape, they must due to siloed data and lack of unified tools ​ 55% of companies have a

55%
be able to transform the massive amount of data • Balancing the need for data access and self- mostly manual approach to
they have into deep, impactful insights. We’ve talked service analytics while remaining governed​
discovering data within their
to a wide range of customers who have led their • Limited scalability of legacy solutions as
organizations in data-first modernization. And we data demand rises exponentially
enterprise.
Source: Accenture
understand that businesses today face a variety of
• Breaking down data siloes across the

68+32+Q
data and analytics challenges that can be difficult organization into a unified source of truth
to navigate:
• Delivering on the promise of analytics with
• Scaling data and analytics across the limited resources
organization while reducing costs and optimizing
existing data and management
Investing in a comprehensive data and analytics 32% of companies reported
platform can address the hurdles encountered in

32%
• Gaining BI adoption to streamline data
being able to realize tangible
your current data ecosystem. With higher adoption
sharing and agility to insights across Lines Of of cloud-based business intelligence and analytics and measurable value from
Business and teams
tools, it’s imperative to work with a partner that has a data.
• Encouraging data literacy by making data trusted portfolio of security products—that can enable Source: Accenture
more accessible and easier to handle across faster, data-driven decisions in a secure manner.
technical and non-technical users

6
Now trending:
data mesh, data fabric, data hub
There are three data estate architectures and concepts that
organizations are applying to the modernization of their
data estate. The data mesh, data fabric, and data hub.

Data Mesh Data Fabric Data Hub

A series of domains assigned to individual LOBs that enables A system for automating data management tasks, such as An open and governed lakehouse platform that provides
access to the data they need with maximum autonomy by unifying and cleaning disparate sources as well as authorizing the storage foundation for serving data to multiple domains
upholding the four principles of a data mesh. data access, that helps a business make the most of its existing efficiently, so domain users can easily consume it.
data sources without needing to migrate them.

7
Modern data architectures are
not mutually exclusive. They are
collectively transformative.

25+75+Q
We know that “by 2024, 75% of organizations will have established a While there is copious academic debate surrounding these modern
centralized data and analytics (D&A) center of excellence to support architectures, the reality is the combination of them will drive the
federated D&A initiatives and prevent enterprise failure.” Federated–– best solution.

75%
otherwise known as standardized––data is a critical component of
The Microsoft approach is based on best practices refined over more
creating a centralized data and analytics (D&A) center of excellence,
than two decades of firsthand experience in building global-scale
which is increasingly becoming a key priority for organizations. By
products and services. It starts with a holistic view of the organization,
implementing a standardized approach to managing data, organizations
factoring in the people, processes, culture, and technology. Then, by
can ensure that their data is consistent, accurate, and reliable, making it
applying data governance, security, and compliance across every layer of
easier to analyze and derive insights from.
the stack, it ensures customers have a truly innovative environment that
empowers everyone to do their best work.

Source: Gartner, Our Top Data & Analytics Predicts

8
 Data Integration
This solution offers comprehensive data migration and

Introducing Microsoft integration to enable a lake-first pattern. Azure Data Factory

connectors enable data integration, while Azure Synapse Link
connectors enable “no code” and “always synchronized” data

Fabric – a unified analytics

integration for operational databases.

OneLake

solution for the era of AI

All data is ingested into a data lake on Azure Data Storage Gen–
–a cost- and performance-optimized data lake storage service––
for the most demanding business intelligence, machine learning,
and artificial intelligence workloads.

Analytics
Enabling the art of the possible, exceeding customer expectations,
Data scientists can bring their preferred compute frameworks,
and delivering transformative business value just got easier with the
languages, runtimes, and tools into the data lakehouse and
introduction of Microsoft Fabric.
further enhance the data through feature engineering and
Microsoft Fabric brings together the best parts of data mesh, data statistical techniques.
fabric, and data hub to provide a one-stop-shop for data integration,
data engineering, real-time analytics, data science, and business Business Intelligence
intelligence needs without compromising the privacy and security of Best-in-class integrated solutions to responsibly democratize
your data. By joining top companies using Microsoft Fabric, you will business intelligence with self-serve tools and experiences for
enable teams to experience an end-to-end, highly integrated, single data analysts and data citizens.
solution that is easy to understand, onboard, create, and operate.
Governance
With this new solution, you will establish a unified source of truth
Microsoft Purview then provides a single pane governance
by bringing together all analytics workloads in a lake-first foundation.
solution to help effectively scan and manage your data
Teams will also be able to reduce the time and effort it takes to uncover
estat­e —even as it grows and scales.
impactful insights through democratized access to data. And this can
be done confidently through a secure, governed solution. 9
Microsoft Fabric

A single, lake-first foundation to unify data

Engineering tools to integrate and reformat data

Data science and ML tools to model and transform data

Business intelligence and automation tools to visualize,

monitor, and act on data

Security and governance tools to safeguard data 10

A unified SaaS-based solution that
stores all organizational data where
analytics workloads operate
Microsoft Fabric combines Data Factory, Synapse Analytics, Data
Explorer, and Power BI into a single, unified experience, on the
cloud. The open and governed data lakehouse foundation is a cost-
effective and performance-optimized fabric for business intelligence,
machine learning, and AI workloads at any scale. It is the foundation for
Microsoft Fabric
migrating and modernizing existing analytics solutions, whether this be
data appliances or traditional data warehouses.

By establishing connectivity and integration, organizations can transform

their unstructured and siloed data into a valuable strategic asset through: Synapse Data Synapse Data Synapse Data Synapse Real
Data Factory Power BI Data Activator
Engineering Science Warehousing Time Analytics
• Data modernization backed by the Microsoft Azure Cloud

• Cloud native applications at any scale

• Responsible, powerful AI to make more informed decision-making

• Analytics and insights at a faster rate OneLake

• Responsible machine learning and artificial intelligence

Unified
• Governance backed by Microsoft Purview
Product experience | Compute and storage | Governance and security | Business model
11
Three-phased approach to
unlocking data value
What will you empower your organization to achieve? What innovation
will come from transformation of your data estate and acceleration of the
Learn about the opportunities that are possible
value you derive from your data?

ENVISION
for your organization when you transform your
data estate to unlock the value of your data. This guide breaks down your journey into three phases: Envision,

And how Microsoft Fabric makes it all possible. Experiment & Learn, and Build & Scale. As we walk through each of the
phases, we will help you discover and consider all the factors involved
in aligning teams with a unified source of truth, establishing a thriving
center of enablement for every user with secure, democratized insights
across your organization, and unlocking transformative data value.
Begin to develop data culture and build trust

EXPERIMENT & LEARN through a smaller scale Proof of Concept that delivers
immediate business value.

Get aligned with your organization’s leaders

and data stewards on the business value you’ll

BUILD & SCALE continue to drive over time, map your current
and future data estate, and establish a network of
transformation champions.
12
PHASE 1

Envisioning the
opportunity begins with
customer centricity
To begin envisioning what innovation may come from
unlocking your data’s value, it’s important to start with
understanding the motivations, challenges and experiences
of your employees, customers, partners, and suppliers.

By grounding your data estate transformation initiative in the

outcomes that are made possible for all constituents – you
will help keep customer centricity at the forefront during the
Build and Scale phase.

Experiment &
Envision Learn Build & Scale
PHASE 1: ENVISION

The art of the possible

What are your customers, employees, partners and
suppliers’ unmet needs and how can you uniquely solve
those unmet needs though unified data, comprehensive
insights and faster decision making? Let’s have a closer
look at what Microsoft Fabric makes possible for the
following industries.

Healthcare

Financial Services

Public Sector

Retail

Sustainability/Energy

14
Experiment &
Envision Learn Build & Scale
 Data challenges
• No comprehensive source for all aggregated data

• Securely providing care teams access to the data they need

• Slow time to insights

How Microsoft Fabric can help

• Leverage OneLake to aggregate all data into a single source
of truth

Healthcare • Empower health team collaboration through democratized insights

• Analyze massive amounts of data at an accelerated pace, using

built-in security and governance policies

The healthcare industry comes with many challenges, learning and artificial intelligence, coupled with large
Key outcomes
as it handles massive amounts of highly sensitive sets of genomic and population data, healthcare
• Enable decision-making backed by data to reshape care and
and regulated data. In addition to maintaining high professionals can gain new insights into drug efficacy.
insights
levels of compliance and security, Microsoft Fabric Healthcare providers will also be empowered to
allows for an improved patient experience by giving collaborate more closely by being able to share and • Improve the patient experience by creating an accessible,
holistic view across healthcare professionals
healthcare organizations access to structured and access patient data and insights.
unstructured medical data in OneLake. Having • Deliver secure, real-time clinical and operational insights using
With Microsoft Fabric, healthcare can be reimagined
a lakehouse foundation enables healthcare Event Hub, IoT Hub, Kafta, and more
by analyzing massive amounts of data at a faster
professionals to run analytics to identify patient
pace than ever before. Paired with improved quality
needs in real-time.
of insights, healthcare professionals are equipped to
Healthcare organizations can also leverage real-world curate personalized treatment plans for all patients.
data collected by clinical drug trials. By using machine
15
Experiment &
Envision Learn Build & Scale
 Data challenges
• Abundance of data and technology siloes

• Securely collecting, storing, and sharing data in a highly

regulated industry

• Slow time to informed decision-making and risk responses

How Microsoft Fabric can help

• Collect data in OneLake to provide one unified source of truth

Financial Services
• Access all portfolio, reference, market, and risk data with open
and governed access controls

• Analyze all different types of data in real-time, using built-in

security and governance policies
The financial services industry (FSI) is heavily financial state, allowing them to upsell or cross-sell
regulated and governed, which makes being products and services that best fit those individual Key outcomes
backed by the security of a Microsoft product that customer needs. • Enable decision-making backed by data to thrive in a
much more important. With Microsoft Fabric, FSI competitive environment
In the insurance industry, organizations will be able to
professionals can access customer portfolio and • Gain a complete, 360-degree view of customers through secure
gain insights about customers by connecting internal
market data with open and governed access data sharing
and external data to provide custom solutions to
controls. They can also strengthen risk prevention
the customer. • Drive accurate insights with speed, while strengthening risk
by leveraging scalable compute and analytics detection and prevention
power modeling. In capital markets, organizations can initiate more
personalized client interactions by gaining more
In the banking industry, organizations will be able to
insights into a client’s profitability and predictive
drive increased customer lifetime value and loyalty
analytics, enabling them to sell the right product at
by providing a holistic view of the customer’s current
the right time to close more deals. 16
Experiment &
Envision Learn Build & Scale
 Data challenges
• Abundance of data and technology siloes

• Securely collecting, storing, and sharing data in a highly

regulated industry

• Slow time to informed decision-making and risk responses

How Microsoft Fabric can help

• Collect data in OneLake to provide one unified source of truth

Public Sector
• Access all portfolio, reference, market, and risk data with open
and governed access controls

• Analyze all different types of data in real-time, using built-in

security and governance policies
For most government agencies, data is scattered. sensors and public utilities/transportation machinery
Data cannot be accessed digitally, and it is not easily can be ingested in OneLake with IoT Hub, for real- Key outcomes
digestible. With Microsoft Fabric, customers will be time analysis. • Enable decision-making backed by data to thrive in a competitive
able to collect and store data in one single location–– environment
From a tax fraud prevention standpoint,
and on the cloud. • Gain a complete, 360-degree view of customers through secure
governments will be able to develop advanced
data sharing
Government agencies will now be able to combine machine learning models to analyze large amounts
research and data across the public health sector and of tax data being held in OneLake to identify and • Drive accurate insights with speed, while strengthening risk
use machine learning and artificial intelligence to prevent tax fraud. detection and prevention

identify health risks and trends.

In addition, governments can leverage data to

drive improved urban planning by offering better
services and experiences. Streaming data from IoT 17
Experiment &
Envision Learn Build & Scale
 Data challenges
• Siloed, incomplete data with complex architecture

• Using data to personalize customer experiences

• Meeting data regulations in areas where your organization

does business

How Microsoft Fabric can help

• Unify data from numerous sources such as purchase orders,
inventory, and manufacturing all in a OneLake

Retail • Gain a 360-degree view of all customers by collecting and

analyzing site, digital, and smart store behavior

• Leverage built-in security and governance to help protect

collected customer data
Retailers face their own set of unique challenges ingesting real-time data from in-store sensors and
when it comes to data. They often struggle with cameras will identify even more in-person areas Key outcomes
closing the gap between online and physical stores, for optimization.
• Inform merchandising and supply chain strategy with real-time
integrating data in a meaningful way, and efficiently
Lastly, and most likely top-of-mind for many retailers, analysis and analytics
managing supply chains. With Microsoft Fabric,
is being empowered to deliver an intelligent supply • Develop content, copy, and products tailored to your customers’
retailers can leverage real-time data to inform their
chain. By using data to anticipate customer demand specific tastes and interests
merchandising strategy, leveraging customer data
on a local and global scale, retailers can get the right
to learn what target audiences value and curating • Be rest assured that in-store and Ecommerce transaction
product, at the right price, to the right customer, at data is secure
personalized experiences.
the right time––closing the gap between demand
Retailers can also optimize digital spend by creating and fulfillment.
models that identify who and where to target buyers
based on data trends. Tracking store activity by
18
Experiment &
Envision Learn Build & Scale
 Data challenges
• Using data to transition to clean, renewable energy

• Forecasting and predicting energy demand

• Identifying new business models backed by data

How Microsoft Fabric can help

• Collect real-time data in OneLake to help identify how clean
energy is used and how to improve its efficacy

Sustainability/Energy • Combine real-time data from disparate sources such as wind

turbines and solar panels and use ML and AI to identify demand

• Identify new business models through AI-driven analysis of

energy distribution, consumption, and customer demand
With Microsoft Fabric, renewable energy providers of energy sources based on usage trends.
can combine real-time data from wind turbines/solar
Most importantly, energy providers can reimagine Key outcomes
panels, as well as customers, to identify future power
energy as we know it by identifying new business • Manage smart and efficiency energy on local and global scales
demands through artificial intelligence and marching
models through AI-driven analysis and real-time
learning models. Energy providers will be able to • Identify how much demand for power will be needed in the
consumption/demand. future – quickly and efficiently
collect large amounts of data from in-home systems
and store it in OneLake for analysis. • Drive efficient, sustainable best practices and reduce
environment impact
Microsoft Fabric will enable energy providers to get
more out of their investments through IoT-enabled
remote tracking to drive AI analytics for sustainable
decision-making. Energy providers will also be able to
collect real-time data in OneLake to improve efficacy
19
Experiment &
Envision Learn Build & Scale
Microsoft Fabric

PHASE 2

Experiment & Learn

The best first step is to experiment and learn
through a Proof of Concept (POC). Starting
with a departmental scenario rather than
an organization-wide scenario can help you
ensure that time spent is minimized and the
value derived is maximized.

Consider the following departmental

scenarios as you embark on your POC.

20
Experiment &
Envision Learn Build & Scale
Marketing

Improve campaign analysis Optimize paid media spend Streamline location analytics Improve website, social, and
and planning email analytics
• Impressions and sales data can be • Track retail brand awareness campaign
• Unify campaign data from different sources combined with media spend and attribution effectiveness at the regional and individual • Collect real-time data like impressions, click-
into a single location to make it easier to data in OneLake. store level by collecting store and behavioral through rates, and video completion rates
track campaign performance vs targets data in one centralized lake storage. from digital ads on websites, social channels,
• Develop models that help identify the best
and budget. areas and channels to move budget to help • Analyze this data with advanced methods and emails in one single location.

• Purchase and attribution data can be gain the most impressions or leads per dollar. and present complex behavioral data in • Perform analysis by analyzing product,
brought in to develop models that help visually stunning, digestible formats. region, and segment level performance to
identify what demographics and/or areas to • Improve understanding of customer develop better targeted ads for customers
target for future campaigns. purchase patterns by combining digital you want.

• Translate data into action by receiving campaign data with in-store visit and
real-time alerts when important campaign Ecommerce data into a single view.
metrics change and adjusting if needed. • Target customers more effectively, pulling
• Pull data silos together by compiling from cloud and on-premises data sources
information in a single, unified view across to see footfall patterns and time specials
your marketing technologies—analyze according to when your target demographic
multiple campaigns together, compare data is shopping.
across markets and platforms, and more.

• Track performance vs. targets and budget,

using fast, interactive reports to help show
ROI and manage spend.

21
Experiment &
Envision Learn Build & Scale
Operations

Undergo production optimization Enhance inventory planning Improve sales and Streamline fulfillment
and performance operations planning and distribution
• Collect and analyze data from different
manufacturing centers to identify production • Unify sales data along with production data • Merge sales, marketing, and manufacturing • Make it easier to identify over- and under-
bottlenecks before they make an impact on to identify which products to prioritize based data from different clouds and systems of used resources using production pick rates
the business. on trends in the market and address stock record to draft new product development or to predict whether to flex up or down and
• Develop predictive models to help them levels before inventory shortages hit. production plans more easily. assess distribution center efficiency.
understand when to scale production • Analyze which products to prioritize by easily • Improve sales and production forecasting • Notify your customers sooner when there is
up and down based on market trends. spotting trends, such as regional spikes in quality by analyzing metrics like past total a shipping delay by leveraging alerts from
Increase efficiency by identifying production demand or increased inventory cycle times sales, production output and shipment times dashboards that can inform you about a
bottlenecks faster using a heads-up for certain items. together using natural language querying shipping gap.
dashboard and custom diagrams. or other advanced techniques to move from
• Assess and adjust levels of safety stock by
• Use visualizations to easily understand data hindsight to insight.
combining recent trend and historical
from manufacturers or production lines in data from cloud, on-premises, or hybrid • Get better visibility into demand and supply
one dashboard to aid in planning. data sources. projections across departments from
• Understand anomalies in real-time anywhere with all data stored in the same
• Receive notifications of major changes
manufacturing data using drill-down logical data lake.
in available inventory across different
capabilities, which better informs required geographies by configuring mobile alerts to
corrective action. help avoid inventory shortages.

22
Experiment &
Envision Learn Build & Scale
HR

Reduce employee turnover and Facilitate workforce/demand Monitor benefits, rewards, Predict organizational
gain visibility into recruiting planning and compensation health trends
and retention
• Combine all your workforce data from • Combine disparate HR data in one location • Predict health trends, such as the type and
• Combine data from financial apps and Dynamics 365 Human Resources into a by seamlessly connecting to both on- frequency of sickness, enabling companies to
recruiting systems into custom reports in just single dashboard for easier viewing and premises and cloud-based data sources, such prepare for outbreaks.
minutes, leveraging rich visualizations that manipulation to drive greater insight into as Human Capital Management systems
drive faster time to insight. your business. like Dynamics 365 Human Resources and
recruiting platforms like LinkedIn Recruiter.
• Explore your staffing funnel or recruiter • Get important updates quickly, such as
performance with an intuitive search-engine unplanned changes to budget or headcount, • Report on the rewards and compensation
format that lets you ask questions of your with real-time data refresh and custom metrics you care about by creating
data and get instant, visual answers. mobile alerts. visualizations and custom dashboards
instantly, instead of being locked into pre-
• Stay informed of fluctuations in key metrics, • Investigate demand planning anomalies
built vendor solutions.
like when retention rates fall below a certain quickly and easily by clicking into detailed
threshold or key positions are filled by setting reports directly from your dashboard, or by • Investigate unexpected spikes in benefit
up mobile notifications. leveraging the search-engine-like capability usage or other areas of interest by instantly
to ask questions of your data and get diving into underlying reports directly from
immediate answers. your dashboard and develop models that
can help predict when these spikes may be
expected in the future.

23
Experiment &
Envision Learn Build & Scale
Sales

Identify better opportunities for Increase localization Develop enhanced pricing Improve sales performance analysis
upselling and cross-selling management plans
• Tailor sales efforts to local markets to uplift • Monitor and compare key metrics from
• Bring together various data points into a and maximize overall sales. • Build a thorough price-management plan to different data sources in real time, including
bigger, more detailed picture to reveal use detailed analytics on sales, promotions, win rates, number of wins, revenue to plan,
• Use sales analytics to understand sales
repeat sales patterns, influences on repeat and competitors. margin, discounts, and more.
conversion at a micro level to structure sales
sales journeys, and where customers are in initiatives, promotional campaigns, and even • Develop a single dashboard that makes it
• Aggregate this data into a single pane of
their journey. stocking levels that best fit local markets. easy for anyone to access and track these
glass to help teams quickly slice and analyze
• Identify upsell and cross-sell opportunities data in various forms to deliver better pricing metrics on their own.
and recommended actions to capture and promotion strategies. • View actuals vs. forecast by territory, product
those opportunities. line, and customer segment, and get answers
to questions like “Which territory had the
highest win rate last month?” in the form of
interactive charts.

24
Experiment &
Envision Learn Build & Scale
Sales

Drive more visibility into Improve quota management: Uncover trends: Discount analysis:
pipeline management:
• Blend sales data from different sources and • Increase deal size via market basket • Compare estimated revenue and
• See your pipeline and quota in one view by deliver real-time insights in dashboards analysis—compile data from multiple sources commission against targets for specific
connecting data from to your CRM account, and reports that make the latest quota to discover products frequently purchased categories, geographies, and account types.
forecasting tools, and more in one data lake. attainment results available on demand. together and identify cross-sell and up-sell
• Measure the impact of discounts and
Keep close tabs on metrics like pipeline value opportunities.
• Get instant insight into key metrics like promotions on demand, traffic, average
and win rates to get a holistic picture of
pipeline by sales stage, average deal size, and • Analyze sales more rapidly to improve sale amount, and other sales KPIs.
progress towards quota, and quickly identify
pipeline velocity so you can focus time and forecasting quality, identify spending trends
at-risk territories. • Leverage historic data collected in legacy
energy on the right opportunities. and variations (e.g., by industry or region),
systems to analyze past discount trends
• Manage team performance with and find new opportunities.
• Combine current and historical data from and make better decisions on what to
personalized dashboards that enable
multiple sources, including Dynamics 365 • Combine transactional information from discount today.
drill-down into variance to target by
Sales, to identify trends and more effectively multiple cloud and on-premises sources to
individual representative. • Improve decision making by modeling
qualify opportunities in or out. uncover trends in top accounts by account
dynamic discount scenarios using what-if
• Define and develop models that help you manager, territory, product line and
analysis and Quick measures.
plan for more accurate and attainable quotas other factors.
in future quarters.

25
Experiment &
Envision Learn Build & Scale
Identify technical
experts and business
users for the POC
Identify availability of technical experts and business users
to support you during the POC execution. Ideally, the
technical decision makers in your organization, either an
IT leader and/or a data leader, will start by creating clear
requirements for the proof of concept. Next, your data
practitioners such as architects, engineers, data scientists
and analysts will execute the POC. And lastly business
decision makers or lines of business (LOB) leaders will play a
large role in evaluating the results.

26
Experiment &
Envision Learn Build & Scale
WORKSHEET

Document clear goals for your POC and the

A few suggested goals
criteria that will define its success
• Reduced number of products used from X to Y
• Time saved from X to Y
Keep in mind that a POC should be a short and focused effort to quickly
prove a limited set of concepts and capabilities. These concepts and • Improve development time from X to Y
capabilities should be representative of the overall workload. Start by • Expand analytical potential between X and Y
creating goals for your POC. through the addition of the new data
• Expand user base from X to Y
Use your goals to identify specific tests and to provide the outputs you
• Satisfy user base demonstrated by X score on
identified. It’s important to make sure that you have at least one test to
user surveys
support each goal and expected output.
• Improve dashboard performance from X to Y
• Decrease time to insight from X to Y
• Business value added. This should be specific to
Goal Output Test 1 Test 2 the scenario you choose
• Performance benchmarking
• Low code/no code data pipelining

27
Experiment &
Envision Learn Build & Scale
 Microsoft Fabric supports four common end-
to-end analytics scenarios:​

Using common analytics 1. Lakehouse​

2. Data Warehouse​

scenarios and proposed 3. Data Science​

4. Real Time Analytics​

activities to guide your Please consider these scenarios as a guide

to support your proof of concept. However,

Proof of Concept keep in mind that while these are the most
common scenarios supported by Microsoft
Fabric, there are many others that technical
and business end-users can navigate through.​

28
Experiment &
Envision Learn Build & Scale
Lakehouse Build and implement an end-to-end lakehouse for
The Microsoft Fabric Lakehouse analytics scenario makes it so that data can be ingested your organization:
into OneLake with shortcuts to other clouds repositories, pipelines, and dataflows in order
1. Create a Microsoft Fabric workspace
to allow end-users to leverage other data. ​
2. Quickly create a lakehouse – an optional module to
Once that data has been pulled into Microsoft Fabric, users can leverage notebooks to
implement medallion architecture (Bronze, Silver,
transform that data in OneLake and then store them in lakehouses with medallion structure.​
and Gold)
From there, users can begin to analyze and visualize that data with Power BI using the Direct 3. Ingest, transform and load data into the lakehouse –
Lake mode or SQL endpoints. bronze, silver and gold zones as delta lake tables for
medallion architecture

4. Explore OneLake, OneCopy of your data across lake

mode and warehouse mode

5. Connect to your lakehouse using TDS/SQL endpoint

6. Create Power BI reports using Direct Lake – to

analyze sales data across different dimensions

7. Orchestrate and schedule data ingestion and

transformation flow with Pipeline

8. Cleanup resources by deleting the workspace and

other items

29
Experiment &
Envision Learn Build & Scale
Lakehouse Data Sources – Microsoft Fabric makes it easy and quick to connect
to Azure Data Services, other cloud platforms, and on premises data
sources to ingest data from.

Ingestion – With 200+ native connectors as part of the Microsoft

The lakehouse end-to-end architecture Fabric pipeline and drag and drop data transformation with
dataflow, you can quickly build insights for your organization.
Shortcut is a new feature in Microsoft Fabric that provides a way to
connect to existing data without having to copy or move it.

Transform and Store – Microsoft Fabric standardizes on Delta

Lake format, that means all the engines of Microsoft Fabric can read
and work on the same dataset stored in OneLake – no need for
data duplicity. This storage allows you to build lakehouses using a
medallion architecture or data mesh based on your organizational
need. For transformation, you can choose either low-code or no-
code experience with pipelines/dataflows or notebook/Spark for a
code first experience.

Consume – Data from your lakehouses can be consumed by Power

BI, an industry leading business intelligence tool for reporting
and visualization. Each Lakehouse comes with a built-in TDS/
SQL endpoint for easily connecting to and querying data in the
lakehouse tables from other reporting tools, when needed. When a
Lakehouse is created a secondary item, called a Warehouse will be
automatically generated at the same time with the same name as
the Lakehouse and this Warehouse item provides you with the TDS/
SQL endpoint.

30
Experiment &
Envision Learn Build & Scale
Data Warehouse Build and implement an end-to-end data
The Data Warehouse analytics scenario takes existing sources that are mounted, while warehouse for your organization:
pipelines and dataflows can bring in all other data that is needed.​
1. Enable Microsoft Fabric in your tenant
IT teams can then define and store procedures to transform the data, which is stored as
Parquet/Delta Lake files in OneLake. ​ 2. Create a Fabric workspace

3. Quickly create a data warehouse

From there, business users can analyze and visualize data with Power BI, again using the
Direct Lake mode or SQL endpoints. 4. Ingest data from source to the data warehouse
dimensional model

5. Transform the data to create aggregated datasets

using T-SQL

6. Perform orchestration, data ingestion, and data

transformation with pipelines

7. Query the data warehouse using T-SQL and a visual

query editor

8. Create Power BI report using Direct Lake mode to

analyze the data in place

9. Cleanup resources by deleting the workspace and

other items

31
Experiment &
Envision Learn Build & Scale
Data Warehouse Data Sources – Microsoft Fabric makes it easy and quick to connect
to Azure Data Services, other cloud platforms, and on premises data
sources to ingest data from.

The data warehouse end-to-end architecture: Ingestion – With 200+ native connectors as part of the Microsoft Fabric
pipeline and drag and drop data transformation with dataflow, you can
quickly build insights for your organization. Shortcut is a new feature in
Microsoft Fabric that provides a way to connect to existing data without
having to copy or move it.

Transform and Store – Microsoft Fabric standardizes on Delta Lake

format, that means all the engines of Microsoft Fabric can read and work
on the same dataset stored in OneLake – no need for data duplicity.
This storage allows you to build a data warehouse or data mesh based
on your organizational need. For transformation, you can choose either
low-code or no code experience with pipelines/dataflows or use T-SQL
for a code first experience.

Consume – Data from your lakehouses can be consumed by Power

BI, an industry leading business intelligence tool for reporting and
visualization. Each data warehouse comes with a built-in TDS/SQL
endpoint for easily connecting to and querying data from other
reporting tools, when needed. When a data warehouse is created a
secondary item, called a default dataset, will be automatically generated
at the same time with the same name of the data warehouse to start
visualizing data with just a couple of mouse clicks.

32
Experiment &
Envision Learn Build & Scale
Data Science
The Data Science analytics scenario can be ingested similarly to the Lakehouse and Data In this scenario, consider performing the following
Warehouse paths. activities:

Once the data is ingested, it is cleaned and prepared using notebooks and then stored in 1. Use the Microsoft Fabric notebooks for data
the lakehouse with medallion structure. science scenarios
After the data is cleaned and stored, machine learning models can be trained and tested 2. Ingest data into Microsoft Fabric lakehouse using
directly on the lakehouse.​ Apache Spark

3. Load existing data from the lakehouse delta tables

4. Clean and transform data using Apache Spark

5. Create experiments and runs to train a machine

learning model

6. Register and track trained models using MLflow and

the Microsoft Fabric UI

7. Run scoring at scale and save predictions and

inference results to the lakehouse

8. Visualize predictions in Power BI using Direct Lake

33
Experiment &
Envision Learn Build & Scale
 Data Sources – Microsoft Fabric makes it easy and quick to connect to Azure
Data Services, other cloud platforms, and on-premises data sources to ingest

Data Science
data from. Using Microsoft Fabric Notebooks you can ingest data from the built-
in Lakehouse, Data Warehouse, and Power BI Datasets as well as various Apache
Spark and Python supported custom data sources.

Explore, Clean & Prepare – The Data Science experience on Microsoft Fabric
supports data cleansing, transformation, exploration and featurization by
The data science end-to-end scenario leveraging built-in experiences on Spark as well as Python based tools like Data
Wrangler and SemPy Library.

Models & Experiments – Microsoft Fabric enables you to train, evaluate and
score machine learning models by using built-in Experiment and Model artifacts
with seamless integration with MLflow for experiment tracking and model
registration/deployment. Microsoft Fabric also features capabilities for model
prediction at scale (PREDICT) to gain and share business insights.

Storage – Microsoft Fabric standardizes on Delta Lake, which means all the
engines of Microsoft Fabric can interact with the same dataset stored in
lakehouse. This storage layer allows you to store both structured and unstructured
data that support both file-based storage as well as tabular format. The datasets
and files stored can be easily accessed via all Microsoft Fabric workload artifacts
like notebooks and pipelines.

Expose Analysis and Insights – Data from the lakehouse can be consumed
by Power BI, an industry leading business intelligence tool for reporting and
visualization. Data persisted in the lakehouse can also be visualized in notebooks
using Spark or Python native visualization libraries like matplotlib, seaborn, plotly,
and more. Data can also be visualized using SemPy library that supports built-in,
task-specific visualizations for the semantic data model, dependencies and their
violations, and classification and regression use cases.

34
Experiment &
Envision Learn Build & Scale
Real Time Analytics
Unlike the Data Science, Lakehouse, and Data Once ingested into Microsoft Fabric, streaming data can
Warehouse analytics scenarios, streaming data can be be stored in Kusto DB and mirrored into the lakehouse.
ingested into the Microsoft Fabric in several ways to After the data has been stored, machine learning
achieve real time analytics.​ models can be trained and tested directly on the
lakehouse with experiments. ​
Users can leverage Event Hub, IoT Hub, pipelines, In this scenario, consider performing the following
dataflows, notebooks, or open-source products like Like the other scenarios, business users can analyze activities:
Kafka, Logstash, and more. and visualize the data with Power BI using the Direct
Lake mode or SQL endpoints. Data can also be exposed 1. Create a KQL Database
through KQL or notebooks using Spark.
2. Create Eventstream
The end-to-end Real-Time Analytics scenario
3. Stream data from Eventstream to KQL Database

4. Check your data with sample queries

5. Save queries as a KQL Queryset

6. Create a Power BI report

7. Create a OneLake shortcut

35
Experiment &
Envision Learn Build & Scale
PHASE 3

Build & Scale

Delivering accelerated data value is not an
overnight process. It involves significant
organizational alignment. It necessitates building
a solid and reliable data foundation. And lastly it
requires consistently delivering business value and
enabling decision making to build trust through
incremental success over time.
36
Experiment &
Envision Learn Build & Scale
Building a foundation of alignment
We recommend the following steps to ensure the foundation for your journey
is solid and you can help your entire organization embrace the culture of
insight and innovation.

1. Get aligned with your organization’s

leaders and data stewards 2. Map your current and future
• Build confidence with department leaders
data estate
• Collaborate on your business value roadmap
• Document the details of your current data estate
3. Establish a network of
• Develop your infrastructure roadmap
• Ideate on planned transformation and outcomes
transformation champions
• Assemble a team of cross-functional subject matter experts

37
Experiment &
Envision Learn Build & Scale
Aligning with your
organization’s leaders and
data stewards Stakeholders crucial to data estate
transformation
• Chief Financial Officer and Other Finance Leaders

• Chief Sales Officer and Other Sales Leaders

• Chief Marketing Officer and Other Marketing Leaders

• Human Resources Leaders

Alignment with your organization’s department leaders • Chief Operations Officer and Other Operations Leaders (including Supply
is crucial to a successful, iterative and value-driven data Chain, Ecommerce, Manufacturing, and more)
estate transformation. You’ll likely need to start by • Customer Service Leaders

4X
helping these important stakeholders see the broader • Research and Development Leaders
vision of what is possible and the economic impact
• Corporate Strategy and Planning Leaders
associated with your aspirations.

According to the Harvard Business Review, data

shows that “companies that embrace a data-driven
culture experience a 4x improvement in revenue
performance and better customer satisfaction”. The
impact of a data culture on an organization is enormous.
Source: How to lead a data-driven digital transformation by Harvard
Business Review
38
Experiment &
Envision Learn Build & Scale
 We recommend running ideation sessions with
department leaders to identify the initiatives that can
drive the most business value once your data estate is
unified. Consider an initiative that can drive some initial
momentum and can be iterated on.

1. How can we enable innovation with data and

analytics for your department?

Conversations and 2. What innovation can we enable?

3. How can I make your line of business more efficient?

questions to drive your

4. What immediate initiatives can we collaborate on and
how will we measure the impact?

business value roadmap

5. What business value will we deliver over time? Cost
reduction? Revenue generation?

6. How can I better enable your line of business to use

data in decision-making?

7. What security and governance concerns should we

be aware of?

8. Is there third-party or IoT data that could enrich

your insights?

39
Experiment &
Envision Learn Build & Scale
 FORESIGHT

WORKSHEET INSIGHT
Create a list of your initiatives to drive business value.
Categorize each initiative by the type of analytics it will
support in the future.
HINDSIGHT
Hindsight: is both descriptive and diagnostic, answering
what happened and why it happened

BUSINESS VALUE
Insight: is predictive, to consider what might happen

Foresight: is prescriptive, to suggest a course of action

This will be the start of your backlog and will help you
prioritize based on the impact it has the potential to make.
While you will ultimately settle on one or two ideas to begin
with, once you have started and begun seeing success, this
backlog of ideas will form a roadmap of what to do next.

IMPLEMENTATION DIFFICULTY

40
Experiment &
Envision Learn Build & Scale
WORKSHEET

Conversations and Data & Cloud

Engineering,
considerations to drive your
ETL, Data Lakes &
infrastructure roadmap Warehouses
Alignment between the Chief Data Officer, Chief
Technology Officer, and Chief Information Officer is
imperative in the transformation of your data estate.
Ensuring partnership through clear functional alignment
as well as decision making authority during these
initiatives will pave the way for a smooth journey.
Insights

Key Considerations and Discussion Points for

Alignment

• Budget Ownership
• Decision Making R&R
• Governance Data Sciences, AI/
• Compliance ML
• Quality Assurance
• Operating Models
• Processes
• Monetization
41
Experiment &
Envision Learn Build & Scale
 Data Source Raw Zone Refined Zone Purposed Zone Consumption Users
Zone

Orchestrate

WORKSHEET

Mapping your current

data and analytics Compute

estate
With your infrastructure team, begin
documenting your current data estate. Storage
Be sure to capture which data sources,
technologies, ELT/ETL scenarios, data
pipelines, and appliances exist. Also,
document how data is consumed by users.

Document the details of your data

estate.
Dev Ops Data Model & Governance Monitor, Logging, Alerting IAM
Which data sources, technologies,
ELT/ETL scenarios exist? How is data
consumed by users?

42
Experiment &
Envision Learn Build & Scale
 Technology In Scope/Out of Scope Planned Transformation Expected Outcomes

WORKSHEET

Ideate on planned
transformation and
outcomes
Begin to document the details of your data
estate transformation. Use the inputs from
your business value roadmap to ensure
you’ve prioritized these initiatives in your
phased infrastructure evolution.

43
Experiment &
Envision Learn Build & Scale
WORKSHEET

Establishing a network of transformation champions

Begin assembling your team of champions for your data estate
transformation. Select people who will work well together, but also
We recommend the following roles as
come from different roles and competencies in the organization and part of your innovation team.
with various skills and experiences. You’ll want to find a balance between • Department leaders (VPs / Directors): who can bring
functional expertise. significant domain expertise

• IT leaders: Infrastructure admin and developers

• Data practitioners: Data scientists, data engineers, business

analysts, data architects who will bring specialized
Team Member Role Functional Skillset
knowledge of a process, function, or task

• Data stewards: Data professionals with knowledge of

security, compliance, governance, and access management

• Business users and data analysts: Power BI power users

44
Experiment &
Envision Learn Build & Scale
Training your
network of 1 2 3 4
champions Introduction to end-
to-end analytics using
Get started with
lakehouses in
Use Apache spark
to work with files in
Work with Delta Lake
tables in Microsoft
Microsoft Fabric Microsoft Fabric Microsoft Fabric Fabric
To help your team get more familiar with
Microsoft Fabric, we’ve put together
learning modules

5 6 7 8 9 10
Use Data Factory Ingest data with Get started with data Get started with Real- Get started with data
Administer Microsoft
pipelines in Microsoft dataflows Gen 2 in warehouses in Microsoft Time Analytics in science in Microsoft
Fabric
Fabric Microsoft Fabric Fabric Microsoft Fabric Fabric

45
Experiment &
Envision Learn Build & Scale
 Microsoft Fabric

Explore the Microsoft Fabric free trial today

You are on
your way to Engage with Microsoft Fabric Community

transformative Sign up for updates

data value
Congratulations on following the journey to data estate
transformation. You are well on your way to leading your
Watch the digital event
organization to accelerated data value and innovation.
As you continue your path, you may find these additional
resources helpful.
Watch the webinar series

46