Implementing the
Segmentation of One
When Opportunity Meets Engagement,
Sparks Ignite!
You Have Questions
Is a single view of customer data a realistic goal in our
current enterprise?
How can we measure and respond to the impact of
mobile, social, cloud, big data and analytics for more
effective customer engagement?
How can I build a team that can do this?
How can we get senior leadership on board,
stakeholders engaged and silos opened?
Where Are You Now?
Systems Of Record
Both Information Management System (IMS) and relational
database management systems (RDMBS) require a well-
defined structure, driven by well-known Enterprise
Applications: CRM, ERP, LMS, HR, accounting, etc.
Enterprise applications led to Systems of Records, which are
typically
 Complete
 Well-known
 Governed
 Managed
However, this ubiquitousness removes any competitive
advantage since they no longer provide meaningful
differentiation.
Enterprise Data Warehouse
Making Systems of Record data available for enterprise
analytics is largely focused on ETL: extracting, transforming
and loading data from these Online Transaction Processing
(OLTP) systems into centralized data marts and warehouses.
Analysts can then target those Online Analytic Processing
stores for complex analytic and ad-hoc queries rapidly
typically using Multidimensional Expressions (MDX) and Linq.
This approach is viable when the source data is typical of
OLTP systems: structured (typically but not exclusively in
relational formats) and at reasonable volumes.
Segmentation of Groups
Systems of Record from Enterprise Applications can provide
a Single Source of Truth particularly when Master Data
Management and Data Governance best practices are in
place.
OLAP provide a 103 performance boost in analytics
compared to OLTP because of aggregations built from the
fact table along specific dimensions.
The view selection problem (its best to calculate cubes in
advance), leads to segmentation based on predefined (but
well-researched) groups.
Best Case Optimizations ...
 Evaluate, consolidate and optimize all Enterprise
Applications.
 Apply Data Governance principles and practices across
the board.
 Create a Master Data Management and Meta-Data
Management repository across the enterprise.
 Optimize ETL into the EDW
 Provide Best-Of-Breed analytics using OLAP cubes as
well as advanced data science solutions like SAS.
... Still Fall Short of Reality
Your customer has a physical, online, social and augmented
presence.
Your customer may visit your store or your website, or call
your help desk, or interact with a bot, or post to social
media. You customer is likely broadcasting their
geolocation through their phone. As they broadcast how
many steps they've take to your store. To their friends
whose proximity has been shared.
New types of applications and new types of data aren't really
even new anymore.
Geospatial, images and video served by mobile, social
and real-time analytic apps are table-stakes, not
differentiators.
Data volume has grown but storage and compute costs
have declined.
You pay your SAN and NAS vendors too much.
Cost-effective, elastic, rapidly-deployed architectures are
well understood and widely used.
Public and private cloud infrastructure operating
alongside traditional on-premise hardware is becoming
a standard pattern.
Where Do You Need To Be?
Segmentation Of One System
A Segmentation Of One System must enable:
 a single view of the customer
 a single truth of the supply chain
 historical to operational to predictive analytics, both
offline and realtime
Segmentation Of One System
New data systems should reflect the new data sources
with dynamic schemas, rich data structures with dynamic
attributes, scalability, and both online and offline analytic
capability.
New analytic systems need to enable a blended
architecture between your current EDW and new data
systems. This blending must occur at the source to
enable governance, security and auditability.
Common Barriers
Here are the top seven major issues “preventing
organization from providing effective digital experiences.”
 Digital experience strategy undefined [38%]
 Lack of cooperation across the organization (silos) [35%]
 Lack of people with the right skills [31%]
 Lack of time, too busy with current departmental jobs [31%]
 Rapidly changing technology solutions [24%]
 Customer experience goals and strategy not defined [24%]
 Can't measure ROI due to data/analytics challenges [24%]
Source: CustomerThink
How Do You Get There?
Technology Barriers
 Lack of people with the right skills [31%]
 Lack of time, too busy with current departmental jobs
[31%]
 Rapidly changing technology solutions [24%]
Technology Barriers
These three barriers can be addressed with the same
strategy:
Identify a platform composed of well-known solutions,
tightly integrated, at the lower-cost end of the
technology development spectrum that is most closely
aligned with the latest data interchange formats and
data types.
Technology Barriers
What is “the lower-cost end of the technology
development spectrum”?
JavaScript.
JavaScript is capable of object-oriented, imperative and
functional programming, both server-side and client-side.
As can Java and C derivatives, which have done this for
a longer time. Which is where the lower-cost comes into
play. Less time on market translates to less salary raises.
Technology Barriers
What is “latest data interchange formats and data types”?
JSON
JSON is an open standard that uses human-readable
text to transmit data using name-value pairs. JSON is
designed for data exchange while XML is designed for
document-exchange. This means that JSON can always
support new data types because it was never intended to
be a document markup language, so it doesn't have
attributes and tags.
Technology Barriers - Resolved
The clearest choice for a new data system is MongoDB.
 MongoDB stores data in BSON (binary JSON)
 The MongoDB shell is written in JavaScript and it works with
Node.js to write event-driven, scalable network programs in
server-side JavaScript. There are connectors for almost any
other programming language.
 Provides a flexible data model to store data of any structure
and dynamically modify the schema.
 Can scale up or scale out horizontally and can be deployed
in the cloud and across multiple data centers.
Technology Barriers - Resolved
Watch these three excuses ...
 Lack of people with the right skills
 Lack of time, too busy with current departmental
jobs
 Rapidly changing technology solutions
… disappear when you ask your development team if
they have the time and interest to work with MongoDB.
Organizational Barriers
 Digital experience strategy undefined [38%]
 Lack of cooperation across the organization (silos)
[35%]
 Customer experience goals and strategy not defined
[24%]
 Can't measure ROI due to data/analytics challenges
[24%]
Organizational Barriers
In a sense, these organizational barriers are a
reasonable outcome of a lack of trusted, actionable
information or
Can't measure ROI due to data/analytics challenges
[24%]
You can't manage what you can't measure.
But there are challenges to measuring.
Organizational Barriers
Lack of cooperation across the organization (silos) [35%]
Yes. Moving on ...
Unfortunately, lack of cooperation is universal and
unavoidable. Dismantling silos cannot be a prerequisite
but it will be a consequence.
You must be able to blend existing data in place with your
new data.
Organizational Barriers
Digital experience strategy undefined [38%]
Customer experience goals and strategy not defined
[24%]
Defining a strategic direction is an iterative task with no
final product.
Organizational Barriers
Can't measure ROI due to data/analytics challenges
[24%]
A Data Analytics platform requires agile, just-in-time,
straightforward access to relevant data at the source of
both the existing EDW and the new data platforms.
In order to perform real-time analytics, data cannot be
cleansed, transformed and stored before analysis. A
blended architecture is required to combine the EDW and
new data systems in real time and in place.
This new ecosystem is known as the Hybrid Data
Ecosystem, the Logical Data Warehouse and the multi-
platform Data Warehouse Environment.
Organizational Barriers - Resolved
Pentaho is the logical choice for creating a blended
architecture for analytics.
 Pentaho Data Integration connects to existing databases as well
as Hadoop, NoSQL, Analytic and specialized data sources
providing visual tools to eliminate coding and complexity.
 Business Analytics provides a code-free interface for business
users to create visual analytics, dashboards and self-service
reports. These analytics encompass EDW, Big Data, NoSQL,
IoT and more for enterprise, cloud and mobile.
 Predictive Analytics and Data Science enable powerful, state-of-
the-art machine learning algorithms, data processing tools and
sophisticated analytics uncover meaningful patterns and
correlations that are hidden with standard analysis and
reporting.
Barriers - Resolved
By creating a MongoDB-First approach to onboarding
new data sets – mobile, social, Internet of Things, you
can enable rapid adaptation to new data challenges.
By implementing a blended-architecture approach to
analytics, you can enable operational, historical and real-
time analytics across the enterprise rapidly and
accurately while maintaining proper data governance,
security and auditing requirements.
We Have Answers
We Have Answers
Is a single view of customer data a realistic goal in our
current enterprise?
Yes. There are challenges but by minimizing the amount
of change needed by owners of existing data silos,
rapidly onboarding new data sources and seamless
performing analytics on both, your chances of success
just got a lot better.
We Have Answers
How can we measure and respond to the impact of
mobile, social, cloud, big data and analytics for more
effective customer engagement?
Iteratively. In-place. In time. Without code.
We Have Answers
How can I build a team that can do this?
Be very mindful about the potential roles and
responsibilities that new architectures can require. Its
easier to deploy to the cloud. Java programmers are
more expensive then JavaScript programmers. Schemas
require data stewards. Business users should not need
developers to create and maintain their reports. If you
use exciting technologies like MongoDB and provide an
engaging developer environment, your team will come.
We Have Answers
How can we get senior leadership on board,
stakeholders engaged and silos opened?
Build on success. Projects based on MongoDB and
JavaScript work very well with iterative development
cycles. There is ramp-up time, so identify a few projects
that have visibility but a modest scope.
Implement internal social networks, tech talks and
hackathons that are open outside of of your team.
Identify stakeholders who are interested and engage
them even if they were not who you originally planned to
start working with.
How Can We Help?
Sparks Ignite
We research, evaluate, design, build & deploy
innovative information technology outcomes.
David Callaghan
Big Data Innovator
Phone (704) 241.9567
david@sparksignite.net

SegmentOfOne

  • 1.
    Implementing the Segmentation ofOne When Opportunity Meets Engagement, Sparks Ignite!
  • 2.
  • 3.
    Is a singleview of customer data a realistic goal in our current enterprise? How can we measure and respond to the impact of mobile, social, cloud, big data and analytics for more effective customer engagement? How can I build a team that can do this? How can we get senior leadership on board, stakeholders engaged and silos opened?
  • 4.
  • 5.
    Systems Of Record BothInformation Management System (IMS) and relational database management systems (RDMBS) require a well- defined structure, driven by well-known Enterprise Applications: CRM, ERP, LMS, HR, accounting, etc. Enterprise applications led to Systems of Records, which are typically  Complete  Well-known  Governed  Managed However, this ubiquitousness removes any competitive advantage since they no longer provide meaningful differentiation.
  • 6.
    Enterprise Data Warehouse MakingSystems of Record data available for enterprise analytics is largely focused on ETL: extracting, transforming and loading data from these Online Transaction Processing (OLTP) systems into centralized data marts and warehouses. Analysts can then target those Online Analytic Processing stores for complex analytic and ad-hoc queries rapidly typically using Multidimensional Expressions (MDX) and Linq. This approach is viable when the source data is typical of OLTP systems: structured (typically but not exclusively in relational formats) and at reasonable volumes.
  • 7.
    Segmentation of Groups Systemsof Record from Enterprise Applications can provide a Single Source of Truth particularly when Master Data Management and Data Governance best practices are in place. OLAP provide a 103 performance boost in analytics compared to OLTP because of aggregations built from the fact table along specific dimensions. The view selection problem (its best to calculate cubes in advance), leads to segmentation based on predefined (but well-researched) groups.
  • 8.
    Best Case Optimizations...  Evaluate, consolidate and optimize all Enterprise Applications.  Apply Data Governance principles and practices across the board.  Create a Master Data Management and Meta-Data Management repository across the enterprise.  Optimize ETL into the EDW  Provide Best-Of-Breed analytics using OLAP cubes as well as advanced data science solutions like SAS.
  • 9.
    ... Still FallShort of Reality Your customer has a physical, online, social and augmented presence. Your customer may visit your store or your website, or call your help desk, or interact with a bot, or post to social media. You customer is likely broadcasting their geolocation through their phone. As they broadcast how many steps they've take to your store. To their friends whose proximity has been shared. New types of applications and new types of data aren't really even new anymore. Geospatial, images and video served by mobile, social and real-time analytic apps are table-stakes, not differentiators.
  • 10.
    Data volume hasgrown but storage and compute costs have declined. You pay your SAN and NAS vendors too much. Cost-effective, elastic, rapidly-deployed architectures are well understood and widely used. Public and private cloud infrastructure operating alongside traditional on-premise hardware is becoming a standard pattern.
  • 11.
    Where Do YouNeed To Be?
  • 12.
    Segmentation Of OneSystem A Segmentation Of One System must enable:  a single view of the customer  a single truth of the supply chain  historical to operational to predictive analytics, both offline and realtime
  • 13.
    Segmentation Of OneSystem New data systems should reflect the new data sources with dynamic schemas, rich data structures with dynamic attributes, scalability, and both online and offline analytic capability. New analytic systems need to enable a blended architecture between your current EDW and new data systems. This blending must occur at the source to enable governance, security and auditability.
  • 14.
    Common Barriers Here arethe top seven major issues “preventing organization from providing effective digital experiences.”  Digital experience strategy undefined [38%]  Lack of cooperation across the organization (silos) [35%]  Lack of people with the right skills [31%]  Lack of time, too busy with current departmental jobs [31%]  Rapidly changing technology solutions [24%]  Customer experience goals and strategy not defined [24%]  Can't measure ROI due to data/analytics challenges [24%] Source: CustomerThink
  • 15.
    How Do YouGet There?
  • 16.
    Technology Barriers  Lackof people with the right skills [31%]  Lack of time, too busy with current departmental jobs [31%]  Rapidly changing technology solutions [24%]
  • 17.
    Technology Barriers These threebarriers can be addressed with the same strategy: Identify a platform composed of well-known solutions, tightly integrated, at the lower-cost end of the technology development spectrum that is most closely aligned with the latest data interchange formats and data types.
  • 18.
    Technology Barriers What is“the lower-cost end of the technology development spectrum”? JavaScript. JavaScript is capable of object-oriented, imperative and functional programming, both server-side and client-side. As can Java and C derivatives, which have done this for a longer time. Which is where the lower-cost comes into play. Less time on market translates to less salary raises.
  • 19.
    Technology Barriers What is“latest data interchange formats and data types”? JSON JSON is an open standard that uses human-readable text to transmit data using name-value pairs. JSON is designed for data exchange while XML is designed for document-exchange. This means that JSON can always support new data types because it was never intended to be a document markup language, so it doesn't have attributes and tags.
  • 20.
    Technology Barriers -Resolved The clearest choice for a new data system is MongoDB.  MongoDB stores data in BSON (binary JSON)  The MongoDB shell is written in JavaScript and it works with Node.js to write event-driven, scalable network programs in server-side JavaScript. There are connectors for almost any other programming language.  Provides a flexible data model to store data of any structure and dynamically modify the schema.  Can scale up or scale out horizontally and can be deployed in the cloud and across multiple data centers.
  • 21.
    Technology Barriers -Resolved Watch these three excuses ...  Lack of people with the right skills  Lack of time, too busy with current departmental jobs  Rapidly changing technology solutions … disappear when you ask your development team if they have the time and interest to work with MongoDB.
  • 22.
    Organizational Barriers  Digitalexperience strategy undefined [38%]  Lack of cooperation across the organization (silos) [35%]  Customer experience goals and strategy not defined [24%]  Can't measure ROI due to data/analytics challenges [24%]
  • 23.
    Organizational Barriers In asense, these organizational barriers are a reasonable outcome of a lack of trusted, actionable information or Can't measure ROI due to data/analytics challenges [24%] You can't manage what you can't measure. But there are challenges to measuring.
  • 24.
    Organizational Barriers Lack ofcooperation across the organization (silos) [35%] Yes. Moving on ... Unfortunately, lack of cooperation is universal and unavoidable. Dismantling silos cannot be a prerequisite but it will be a consequence. You must be able to blend existing data in place with your new data.
  • 25.
    Organizational Barriers Digital experiencestrategy undefined [38%] Customer experience goals and strategy not defined [24%] Defining a strategic direction is an iterative task with no final product.
  • 26.
    Organizational Barriers Can't measureROI due to data/analytics challenges [24%] A Data Analytics platform requires agile, just-in-time, straightforward access to relevant data at the source of both the existing EDW and the new data platforms. In order to perform real-time analytics, data cannot be cleansed, transformed and stored before analysis. A blended architecture is required to combine the EDW and new data systems in real time and in place. This new ecosystem is known as the Hybrid Data Ecosystem, the Logical Data Warehouse and the multi- platform Data Warehouse Environment.
  • 27.
    Organizational Barriers -Resolved Pentaho is the logical choice for creating a blended architecture for analytics.  Pentaho Data Integration connects to existing databases as well as Hadoop, NoSQL, Analytic and specialized data sources providing visual tools to eliminate coding and complexity.  Business Analytics provides a code-free interface for business users to create visual analytics, dashboards and self-service reports. These analytics encompass EDW, Big Data, NoSQL, IoT and more for enterprise, cloud and mobile.  Predictive Analytics and Data Science enable powerful, state-of- the-art machine learning algorithms, data processing tools and sophisticated analytics uncover meaningful patterns and correlations that are hidden with standard analysis and reporting.
  • 28.
    Barriers - Resolved Bycreating a MongoDB-First approach to onboarding new data sets – mobile, social, Internet of Things, you can enable rapid adaptation to new data challenges. By implementing a blended-architecture approach to analytics, you can enable operational, historical and real- time analytics across the enterprise rapidly and accurately while maintaining proper data governance, security and auditing requirements.
  • 29.
  • 30.
    We Have Answers Isa single view of customer data a realistic goal in our current enterprise? Yes. There are challenges but by minimizing the amount of change needed by owners of existing data silos, rapidly onboarding new data sources and seamless performing analytics on both, your chances of success just got a lot better.
  • 31.
    We Have Answers Howcan we measure and respond to the impact of mobile, social, cloud, big data and analytics for more effective customer engagement? Iteratively. In-place. In time. Without code.
  • 32.
    We Have Answers Howcan I build a team that can do this? Be very mindful about the potential roles and responsibilities that new architectures can require. Its easier to deploy to the cloud. Java programmers are more expensive then JavaScript programmers. Schemas require data stewards. Business users should not need developers to create and maintain their reports. If you use exciting technologies like MongoDB and provide an engaging developer environment, your team will come.
  • 33.
    We Have Answers Howcan we get senior leadership on board, stakeholders engaged and silos opened? Build on success. Projects based on MongoDB and JavaScript work very well with iterative development cycles. There is ramp-up time, so identify a few projects that have visibility but a modest scope. Implement internal social networks, tech talks and hackathons that are open outside of of your team. Identify stakeholders who are interested and engage them even if they were not who you originally planned to start working with.
  • 34.
  • 35.
    Sparks Ignite We research,evaluate, design, build & deploy innovative information technology outcomes. David Callaghan Big Data Innovator Phone (704) 241.9567 david@sparksignite.net