1© Cloudera, Inc. All rights reserved.
The Big Picture: Real-time Data is
Defining Intelligent Offers
Sean Anderson, Sr. Solutions Marketing Manager
Ryan Lippert, Sr. Product Marketing Manager
2© Cloudera, Inc. All rights reserved.
We empower
people to transform complex data
into clear and actionable insights
DRIVE
CUSTOMER INSIGHTS
CONNECT
PRODUCTS & SERVICES (IoT)
PROTECT
BUSINESS
3© Cloudera, Inc. All rights reserved.
DRIVE CUSTOMER INSIGHTS CONNECT PRODUCTS & SERVICES
(IoT)
PROTECT
BUSINESS
Delivering greater value through
improved customer understanding
Powering predictive analytics to increase
performance and reduce fleet downtime
Creating new revenue streams with an
advanced anti-fraud solution
Cloudera powering data-driven customers
4© Cloudera, Inc. All rights reserved.
Cloudera Use Cases
Omni-Channel optimization
Customer analysis
Sentiment analysis
Churn analysis
Market spend analysis
Next best offer
Smart promotions
Basket analysis
Network threat detection
User/entity behavioral
analysis
Logger
Merchant fraud
Connect products & services
Predictive maintenance
Remote monitoring
Supply chain optimization
Inventory optimization
Operations optimization
Spend analytics service
Drive customer insights Protect business
5© Cloudera, Inc. All rights reserved.
Powering a Variety of Use Cases…
Targeted Marketing
Smart Promotions
Recommendation Engines
Omni-Channel
Optimization
6© Cloudera, Inc. All rights reserved.
✓ Breaking down data silos
✓ Sharing data in accordance with
privacy and regulatory policies
✓ Becoming iterative, lean and
leveraging knowledge across the
business
Creating True Customer 360
7© Cloudera, Inc. All rights reserved.
Customer 360 Journey
• Marketing Systems
(Salesforce, Omniture,
CRM)
• Clickstream Data
Primary
Data
Source
• Clickstream
• NPS Systems
• Support Call Logs
• Social Feeds
Primary
Data
Source
Understand Your Customer Learn Behaviors Improve Interactions
• Shopping Cart Platforms
• Geolocation
Primary
Data
Source
8© Cloudera, Inc. All rights reserved.
The Data Journey
Collate the Data Sources Micro-Segmentation
Drive Personalized
Campaigns
Devise Micro- segments based
on combining multiple factors:
• Age
• Location
• Spending History
• Channel Preferences
• Content Preferences
• Apps Usage
• Social Influence
• Churn Score
• Lifetime Value
• Usage Patterns
• Data Usage
Drive Personalized Campaigns for
specific micro-segments
Retention campaign for high value
customers with iPhone who
recently shared a negative social
sentiment
Upsell campaign for high-data users
with family to move over to a family
bundle
Geo-Location based targeted
advertising for specific customer
micro-segments
9© Cloudera, Inc. All rights reserved.
How to Iteratively Build a True Customer 360?
Customer
Data
Source
Start with ingesting
the “best” version of
your customer profile
Find your common
identifiers across
datasets: customer
name, number, IMEI,
IMSI
IMEI
ChannelsPurchase
History
Add New Data Source
Common
Identifier
Current Source
Enrich with additional
demographic information
(purchase history or
channels) from other
systems / sources
Deliver A Use Case
Deliver a specific use case
based on the profile with
new data sets:
• Customer Lifetime value
• Next Best offer
• Omni Channel
Enrich Your Profile
• Enrich your customer
profiles with
purchase behavior
• Continue to enhance
with each new use
case
Location Clickstream
Continue to add new data sources iteratively to
enhance your customer profile with new use
cases
Call center
Social Media Apps
External
Data
New Data Sources
10© Cloudera, Inc. All rights reserved.
Three Scenarios
11© Cloudera, Inc. All rights reserved.
Three Scenarios – Event Modelling in Real Time
Events trigger changes in purchasing
preferences among customers and potential
customers.
However, most NBO frameworks are based on
historic data models. Historic models handle a
baseline of behavior/information well, but
struggle to optimize in the moment for events.
By incorporating the real-time behavior of users
back into the model on a rolling basis,
companies can capture the opportunity these
events present.
12© Cloudera, Inc. All rights reserved.
56% of all customer interactions happen
during a multi-channel, multi-event journey.
Companies that put data at the center of their
marketing and sales decisions improve their
marketing returns by 15-20% adding up to
150 to 20 billion in additional revenue.
Research states that personalized emails
improve click through rates by 14% and
conversation rates by 10%
Over 96% of organizations believe that email
personalization can improve email marketing
performance.
Three Scenarios - Email Personalization
13© Cloudera, Inc. All rights reserved.
HEB is the largest grocery chain in the state of
Texas.
When employees couldn't get to work, some
stores still operated with as few as five people
Hurricane impact projections are often not
accurate which means HEB had to plan for the
worst and leverage real-time data to make
shipping and staffing decisions.
Certain items become in high demand during a
hurricane, while other experience almost no-
demand (Frozen Foods, Flowers)
HEB leveraged real-time data to plan special
shipments which arrived before state and
federal aide.
Three Scenarios - Hurricane Harvey
14© Cloudera, Inc. All rights reserved.
The Platform
15© Cloudera, Inc. All rights reserved.
Cloudera Enterprise – The Platform for Customer 360
Location
Social
Clickstream
BI Tools
Online & Mobile Apps
Billing/
Ordering
CRM/ Profile
Marketing
Campaigns
Search
EDW
N/W
Logs
Call Center
Apps
Network
Other
Structured
Sources
Internal Systems External Sources
BI Solutions Real-Time
Apps
Search Data Science
Workbench
SQL
Machine
Learning
Systems Data
16© Cloudera, Inc. All rights reserved.
Key Enabling Capabilities
Ideal for real-time analytics on
IoT and time series data.
Simplifies Lambda architectures
for running real-time analytics on
streaming data
Leading analytic SQL engine
running natively in Hadoop. Impala
provides the fastest insights, at
high-concurrency, with the familiar
access necessary for powering BI
and analytics across the business.
Kudu: Real-Time Offers Impala: Self Service BI Data Science Workbench
Collaborative hub for enterprise
data science and an integrated
development environment for
running Python, R, & Scala with
support for Spark
17© Cloudera, Inc. All rights reserved.
• Serve real-time data at scale for real-time decision making
• Aggregate relational, NoSQL, structured & unstructured data
• Stream processing & analytics on changing operational data
• Leverage linear performance scalability and predictable TCO
• Deliver a secure, low-latency, high-concurrency experience
Extract real-time insights from big data
OPERATIONAL
DATABASE
18© Cloudera, Inc. All rights reserved.
The Underlying Driver
What drives a use case to real-time?
High Frequency
Trading
APT Detection
Fraud Detection
Predictive Maintenance
Next Best Offer
Inventory Management
Shipping/Logistic
Systems
CRM Systems
Employee Management
Strategic Planning
Real-time data management use cases are
defined by a common set of characteristics.
• Narrow time window in which to make a decision
(automated or manual)
• Opportunity for data points to change the decision,
and thus the business’s path
• Decreasing value of data over time
Not all use cases have a pressing need for
real-time data.
• Broader strategic decisions, for example, do not
require real-time data input
• Over time, decreases in HW costs and increases in
availability of real-time systems will lead most use
cases to be conducted in real-time
Real Time
Some
Latency
Acceptable
19© Cloudera, Inc. All rights reserved.
Managing Data from Customer Touchpoints
Handle real-time
data ingest from
diverse sources
Fundamentall
y Secure
Data Streams
Deployment Flexibility
Machine Learning
Capabilities
Diverse Analytical
OptionsCombine data from various sources
Customer Data Mgmt.
Hub
Scale easily & Cost
effectively
Batch or Real- time
Data Streams
A comprehensive data management platform to drive business insights from data
Data Sources
Data Storage &
Processing
Serving, Analytics &
Machine Learning
Data Ingest
Data Sources
Security, Scalability & Easy Management
20© Cloudera, Inc. All rights reserved.
The Right Storage Technology to Meet Your Use Case
Real-Time Inputs Real-Time Analytics
Input data is pushed into a
semi-static model of a well-
defined process, resulting in the
selection of an optimal strategy
given the known variables.
Input data itself becomes part of
the model, continuously
evolving (within boundaries) as
behaviors change and new
connections are identified.
21© Cloudera, Inc. All rights reserved.
Operational DB: Real-Time Architecture
Driving the Model Through Machine Learning
Kafka
Spark
Streaming
Kudu
Spark MLlib
Input Data
Addt’l
Sources
Individual Session
Full Model/Learning
Genesis
Spark
1 Event
Occurs
2
Messaging
3
Stream
Processing 4
Land in
Relational
Store
5
Apply ML
Libraries
22© Cloudera, Inc. All rights reserved.
Operational DB: Real-Time Architecture
MLlib & K-Means: Defining Microsegments via Machine Learning
Height
Weight
Height
Weight
1 2
Height
Weight
3
Height
Weight
4
L
M
S
XL
L
M
S
XS
Near
Custom
?
23© Cloudera, Inc. All rights reserved.
Operational DB: Real-Time Architecture
Driving Prediction and Optimization
Kafka
Spark
Streaming
Kudu
Spark MLlib
Input Data
Addt’l
Sources
Individual Session
1
Data
Processed
Genesis
Spark
2
Request Processed/
Kudu Queried
3
4
Results
Returned
Results
Processed
5
Processed
Data
Returned
Full Model/Learning
24© Cloudera, Inc. All rights reserved.
Operational DB: Real-Time Architecture
Driving Prediction and Optimization
Step 1: Data Processed
Apache Spark processes the data from the event (car sensors, manufacturing,
wearables, etc), which potentially involves keeping a running list of the last X
number of events
Step 2: Request Processed/Kudu Queried
A Spark application uses the data gathered in step one to query Kudu’s database
in a predefined manner to look for similar patterns defined via machine learning
Step 3: Kudu Results Returned
Kudu returns the results from the query in step 2 back to Spark to determine what
needs to be returned to the application
Step 4: Results Processed
Spark associates the results from Kudu with the information stored from the
current event to determine the next step to feed back to the application
Step 5: Processed Data Returned
The machine-generated, best possible outcome is prescribed and served to the
application
25© Cloudera, Inc. All rights reserved.
Operational DB: NBO Use Case
Prediction and Optimization
Kafka
Spark
Streaming
Kudu
Spark MLlib
Application
Addt’l
Sources
Individual Session
User Shopping
Spark
Full Model/Learning
Data Request Sent For Stream Processing
Data Cleaned/Ordered/Processed, Then
Delivered to Kudu for Modelling
Automated processes based on machine
learning enable prediction and
optimization at a new level.
Illustrative,
models will
likely have
>2
dimensions
26© Cloudera, Inc. All rights reserved.
Visit:
Solutions
Gallery
27© Cloudera, Inc. All rights reserved.
Thank you
Sean Anderson, Sr. Solutions Marketing
Manager
Ryan Lippert, Sr. Product Marketing Manager

The Big Picture: Real-time Data is Defining Intelligent Offers

  • 1.
    1© Cloudera, Inc.All rights reserved. The Big Picture: Real-time Data is Defining Intelligent Offers Sean Anderson, Sr. Solutions Marketing Manager Ryan Lippert, Sr. Product Marketing Manager
  • 2.
    2© Cloudera, Inc.All rights reserved. We empower people to transform complex data into clear and actionable insights DRIVE CUSTOMER INSIGHTS CONNECT PRODUCTS & SERVICES (IoT) PROTECT BUSINESS
  • 3.
    3© Cloudera, Inc.All rights reserved. DRIVE CUSTOMER INSIGHTS CONNECT PRODUCTS & SERVICES (IoT) PROTECT BUSINESS Delivering greater value through improved customer understanding Powering predictive analytics to increase performance and reduce fleet downtime Creating new revenue streams with an advanced anti-fraud solution Cloudera powering data-driven customers
  • 4.
    4© Cloudera, Inc.All rights reserved. Cloudera Use Cases Omni-Channel optimization Customer analysis Sentiment analysis Churn analysis Market spend analysis Next best offer Smart promotions Basket analysis Network threat detection User/entity behavioral analysis Logger Merchant fraud Connect products & services Predictive maintenance Remote monitoring Supply chain optimization Inventory optimization Operations optimization Spend analytics service Drive customer insights Protect business
  • 5.
    5© Cloudera, Inc.All rights reserved. Powering a Variety of Use Cases… Targeted Marketing Smart Promotions Recommendation Engines Omni-Channel Optimization
  • 6.
    6© Cloudera, Inc.All rights reserved. ✓ Breaking down data silos ✓ Sharing data in accordance with privacy and regulatory policies ✓ Becoming iterative, lean and leveraging knowledge across the business Creating True Customer 360
  • 7.
    7© Cloudera, Inc.All rights reserved. Customer 360 Journey • Marketing Systems (Salesforce, Omniture, CRM) • Clickstream Data Primary Data Source • Clickstream • NPS Systems • Support Call Logs • Social Feeds Primary Data Source Understand Your Customer Learn Behaviors Improve Interactions • Shopping Cart Platforms • Geolocation Primary Data Source
  • 8.
    8© Cloudera, Inc.All rights reserved. The Data Journey Collate the Data Sources Micro-Segmentation Drive Personalized Campaigns Devise Micro- segments based on combining multiple factors: • Age • Location • Spending History • Channel Preferences • Content Preferences • Apps Usage • Social Influence • Churn Score • Lifetime Value • Usage Patterns • Data Usage Drive Personalized Campaigns for specific micro-segments Retention campaign for high value customers with iPhone who recently shared a negative social sentiment Upsell campaign for high-data users with family to move over to a family bundle Geo-Location based targeted advertising for specific customer micro-segments
  • 9.
    9© Cloudera, Inc.All rights reserved. How to Iteratively Build a True Customer 360? Customer Data Source Start with ingesting the “best” version of your customer profile Find your common identifiers across datasets: customer name, number, IMEI, IMSI IMEI ChannelsPurchase History Add New Data Source Common Identifier Current Source Enrich with additional demographic information (purchase history or channels) from other systems / sources Deliver A Use Case Deliver a specific use case based on the profile with new data sets: • Customer Lifetime value • Next Best offer • Omni Channel Enrich Your Profile • Enrich your customer profiles with purchase behavior • Continue to enhance with each new use case Location Clickstream Continue to add new data sources iteratively to enhance your customer profile with new use cases Call center Social Media Apps External Data New Data Sources
  • 10.
    10© Cloudera, Inc.All rights reserved. Three Scenarios
  • 11.
    11© Cloudera, Inc.All rights reserved. Three Scenarios – Event Modelling in Real Time Events trigger changes in purchasing preferences among customers and potential customers. However, most NBO frameworks are based on historic data models. Historic models handle a baseline of behavior/information well, but struggle to optimize in the moment for events. By incorporating the real-time behavior of users back into the model on a rolling basis, companies can capture the opportunity these events present.
  • 12.
    12© Cloudera, Inc.All rights reserved. 56% of all customer interactions happen during a multi-channel, multi-event journey. Companies that put data at the center of their marketing and sales decisions improve their marketing returns by 15-20% adding up to 150 to 20 billion in additional revenue. Research states that personalized emails improve click through rates by 14% and conversation rates by 10% Over 96% of organizations believe that email personalization can improve email marketing performance. Three Scenarios - Email Personalization
  • 13.
    13© Cloudera, Inc.All rights reserved. HEB is the largest grocery chain in the state of Texas. When employees couldn't get to work, some stores still operated with as few as five people Hurricane impact projections are often not accurate which means HEB had to plan for the worst and leverage real-time data to make shipping and staffing decisions. Certain items become in high demand during a hurricane, while other experience almost no- demand (Frozen Foods, Flowers) HEB leveraged real-time data to plan special shipments which arrived before state and federal aide. Three Scenarios - Hurricane Harvey
  • 14.
    14© Cloudera, Inc.All rights reserved. The Platform
  • 15.
    15© Cloudera, Inc.All rights reserved. Cloudera Enterprise – The Platform for Customer 360 Location Social Clickstream BI Tools Online & Mobile Apps Billing/ Ordering CRM/ Profile Marketing Campaigns Search EDW N/W Logs Call Center Apps Network Other Structured Sources Internal Systems External Sources BI Solutions Real-Time Apps Search Data Science Workbench SQL Machine Learning Systems Data
  • 16.
    16© Cloudera, Inc.All rights reserved. Key Enabling Capabilities Ideal for real-time analytics on IoT and time series data. Simplifies Lambda architectures for running real-time analytics on streaming data Leading analytic SQL engine running natively in Hadoop. Impala provides the fastest insights, at high-concurrency, with the familiar access necessary for powering BI and analytics across the business. Kudu: Real-Time Offers Impala: Self Service BI Data Science Workbench Collaborative hub for enterprise data science and an integrated development environment for running Python, R, & Scala with support for Spark
  • 17.
    17© Cloudera, Inc.All rights reserved. • Serve real-time data at scale for real-time decision making • Aggregate relational, NoSQL, structured & unstructured data • Stream processing & analytics on changing operational data • Leverage linear performance scalability and predictable TCO • Deliver a secure, low-latency, high-concurrency experience Extract real-time insights from big data OPERATIONAL DATABASE
  • 18.
    18© Cloudera, Inc.All rights reserved. The Underlying Driver What drives a use case to real-time? High Frequency Trading APT Detection Fraud Detection Predictive Maintenance Next Best Offer Inventory Management Shipping/Logistic Systems CRM Systems Employee Management Strategic Planning Real-time data management use cases are defined by a common set of characteristics. • Narrow time window in which to make a decision (automated or manual) • Opportunity for data points to change the decision, and thus the business’s path • Decreasing value of data over time Not all use cases have a pressing need for real-time data. • Broader strategic decisions, for example, do not require real-time data input • Over time, decreases in HW costs and increases in availability of real-time systems will lead most use cases to be conducted in real-time Real Time Some Latency Acceptable
  • 19.
    19© Cloudera, Inc.All rights reserved. Managing Data from Customer Touchpoints Handle real-time data ingest from diverse sources Fundamentall y Secure Data Streams Deployment Flexibility Machine Learning Capabilities Diverse Analytical OptionsCombine data from various sources Customer Data Mgmt. Hub Scale easily & Cost effectively Batch or Real- time Data Streams A comprehensive data management platform to drive business insights from data Data Sources Data Storage & Processing Serving, Analytics & Machine Learning Data Ingest Data Sources Security, Scalability & Easy Management
  • 20.
    20© Cloudera, Inc.All rights reserved. The Right Storage Technology to Meet Your Use Case Real-Time Inputs Real-Time Analytics Input data is pushed into a semi-static model of a well- defined process, resulting in the selection of an optimal strategy given the known variables. Input data itself becomes part of the model, continuously evolving (within boundaries) as behaviors change and new connections are identified.
  • 21.
    21© Cloudera, Inc.All rights reserved. Operational DB: Real-Time Architecture Driving the Model Through Machine Learning Kafka Spark Streaming Kudu Spark MLlib Input Data Addt’l Sources Individual Session Full Model/Learning Genesis Spark 1 Event Occurs 2 Messaging 3 Stream Processing 4 Land in Relational Store 5 Apply ML Libraries
  • 22.
    22© Cloudera, Inc.All rights reserved. Operational DB: Real-Time Architecture MLlib & K-Means: Defining Microsegments via Machine Learning Height Weight Height Weight 1 2 Height Weight 3 Height Weight 4 L M S XL L M S XS Near Custom ?
  • 23.
    23© Cloudera, Inc.All rights reserved. Operational DB: Real-Time Architecture Driving Prediction and Optimization Kafka Spark Streaming Kudu Spark MLlib Input Data Addt’l Sources Individual Session 1 Data Processed Genesis Spark 2 Request Processed/ Kudu Queried 3 4 Results Returned Results Processed 5 Processed Data Returned Full Model/Learning
  • 24.
    24© Cloudera, Inc.All rights reserved. Operational DB: Real-Time Architecture Driving Prediction and Optimization Step 1: Data Processed Apache Spark processes the data from the event (car sensors, manufacturing, wearables, etc), which potentially involves keeping a running list of the last X number of events Step 2: Request Processed/Kudu Queried A Spark application uses the data gathered in step one to query Kudu’s database in a predefined manner to look for similar patterns defined via machine learning Step 3: Kudu Results Returned Kudu returns the results from the query in step 2 back to Spark to determine what needs to be returned to the application Step 4: Results Processed Spark associates the results from Kudu with the information stored from the current event to determine the next step to feed back to the application Step 5: Processed Data Returned The machine-generated, best possible outcome is prescribed and served to the application
  • 25.
    25© Cloudera, Inc.All rights reserved. Operational DB: NBO Use Case Prediction and Optimization Kafka Spark Streaming Kudu Spark MLlib Application Addt’l Sources Individual Session User Shopping Spark Full Model/Learning Data Request Sent For Stream Processing Data Cleaned/Ordered/Processed, Then Delivered to Kudu for Modelling Automated processes based on machine learning enable prediction and optimization at a new level. Illustrative, models will likely have >2 dimensions
  • 26.
    26© Cloudera, Inc.All rights reserved. Visit: Solutions Gallery
  • 27.
    27© Cloudera, Inc.All rights reserved. Thank you Sean Anderson, Sr. Solutions Marketing Manager Ryan Lippert, Sr. Product Marketing Manager