Transcript Big Data PowerPoint Presentation - Integral Net Information Systems

Big Data Analytics
By Anahita Saghafi
2012
Big Data - What?
• It is about data with
the following
characteristics
– Volume
– Velocity
– Variety
Volume
• 36% Average Growth of
Business data each year
• Gartner says that vendors
must revamp their data
warehouse platforms to
deal with growing data
volumes, taking advantage
of new and exciting
technologies, as more
enterprises collect greater
data volumes while
demanding real time
reporting and analytics.
Variety
All the formats and structures in which data •
arrives:
Structured In relational databases
and dimensional data warehouses –
extracted from operational source
systems such as ERP and CRM.
•
Un-structured Data in variety of
formats such as text, audio, video.
•
Semi-structured is everything in
between. Web logs in the form of
XML documents, call data records
from networks, statuses from
components in a smart grid, GPS
readings that locate a smart phone.
Velocity
78%of Big Data companies
need delivery of data
within 24 hours
• Faster speed in which
data is created/updated
and streams in
• Faster speed of access
to the intelligence
• Larger amount of new
and updated data –
faster than ever
Who is Winning?
The Aberdeen Group surveyed 247 organizations about their data
practices and assessed their responses using the following criteria:
•
•
•
•
Time to Integrate New Sources
Increase in Accessible Data
Accuracy of Data
On-time Intelligence Delivery
Aberdeen then ranked the organizations along the following scale
• TOP 20%: Best in Class
• MIDDLE 50%: Industry Average
• BOTTOM 30%: Laggards
Who is Winning?
•
•
•
•
TOP 20%: Best in Class
MIDDLE 50%: Industry Average
BOTTOM 30%: Laggards
•
•
•
21 days required to integrate new
data sources, compared with 53 days
for the Industry Average and 130
days for Laggards
32% year over year increase in
accessible business data, compared
with a 16% increase for the Industry
Average and 6% for Laggards
91% of business data is considered to
be accurate, compared with 80% for
the Industry Average and 54% for
Laggards
92% of key business information is
delivered on-time, compared with
75% for the Industry Average and
39% for Laggards
Pain Points
• 69% Said: Increasing
Demand for
Management
Information
• 67% said: New analytic
needs not well suited to
existing data warehouse
What Works?
• Executive-Level Policies
Supporting Better Data
Oversight
• Decision Culture That
Values the Use of
Supporting Data
• Formal In-House
Development of
Analytical Skill Sets
• Data Quality Tools
(Cleansing, Enrichment,
Normalization)
How BEST IN CLASS Manage BIG DATA
• Elevate Data
Management Initiatives
to the Executive Level.
• Evaluate Data Quality
Technologies.
• Improve Ability to
Measure Time to
Information
Following the Leaders
• PROCESS
61% of the Best in Class have executive-level data
oversight policies.
• ORGANIZATION
The Best in Class are 63% more likely to value
using data to support crucial decisions.
• TECHNOLOGY
The Best in Class are more likely to have a
dedicated business intelligence platform.
• KNOWLEDGE MANAGEMENT
The Best in Class are 2X likelier to centralize
creating, updating, and maintaining data
definitions.
• PERFORMANCE MANAGEMENT
Top performers are 76% more likely to capture,
monitor, and improve upon data quality issues
Benefit of Big Data Analytics in Retail
• Gain insight about the effectiveness of specific
marketing campaigns and channels
• Implement cross-channel marketing campaigns
• Use data analytics to better align overall marketing
activities with specific sales objectives and goals
• Improve the targeting of marketing offers to optimize
marketing ROI (right person, right channel, right time,
right message)
• Align marketing activities with specific sales objectives
and goals
• Improve customer profitability and value by identifying
cross-sell / up-sell opportunities to existing customers
Benefit of Big Data Analytics in Retail
• Identify new product opportunities within our current
market
• Build unique customer profiles and personas in order to
match solutions to specific customer needs
• Implement a customer experience management (CEM)
initiative
• Implement a customer analytics strategy
• Improve digital advertising yield
• Improve the brand image of our company (i.e. build
brand equity)
• Optimize marketing activities at each touch-point along
the customer lifecycle
Marketing Analytics
• Outbound Marketing Campaigns (i.e. customer segmentation,
offer targeting, etc.)
• Inbound Marketing Campaigns (i.e. calls into call centre, web
visit)
• Customer Experience Management (CEM) (aligning customer
touch points with buyer preferences / behaviour)
• Brand management / market research
• Web Experience Management (WEM) (e.g. Web analytics,
optimization, targeting, etc.)
• eCommerce merchandising (i.e. next best offer, product
recommendation, cart abandonment, etc.)
• Social media marketing and Social CRM
Customer Centric Analytics
• Increase overall response rate of marketing campaigns /
campaign optimization
• Understanding customers as individuals
• Gain a better understanding of conversion trends and
audience behaviours
• Increase accuracy of audience targeting
• Marketing Mix optimization (i.e. insight about which programs
truly make an incremental contribution)
• Reduce churn / increase loyalty
• Increase cross-sell revenue
• Increase lifetime value
• Gain insight for new product development
Knowledge Management
Requirements
• Ability to define buyer personas (based on demographic,
firmographic, and / or behavioural data) for marketing
campaigns and nurturing
• Process to dynamically optimize / customize web content
based on buyer persona / segmentation
• Ability to generate customer behavioural profile based on
real-time click-stream analysis
• Ability to optimize marketing offers or web experience based
on social activity / profile of buyer
• Prospect / buyer data captured in various channels is
aggregated and utilized in cross-channel marketing efforts
• Process to analyse cross-channel attribution
Knowledge Management
Requirements
• Ability to incorporate unstructured data into analytical
models
• Customer behaviour is used to segment and target
marketing audiences
• Access to spending history of existing customers
• Access to behavioural/inbound marketing data of
prospective customers
• Ability to capture customer sentiment data, either
structured or unstructured
• Customer profile is updated dynamically (real time or
near-real time) based on customer activity (transactional
/ behavioral data)
Performance Management
Requirements
• Ability to measure customer profitability
• Ability to incorporate campaign response / results into
analytical models
• Enterprise-wide metrics defined to measure success of
analytical models
• Ability to capture a consolidated view of all campaignrelated activities
• Process to measure the marginal lift from targeted
campaigns
• Ability to measure lifetime customer value
• Use of role-based marketing analytics dashboard (views
defined by role, i.e. campaign management to executive)
Traditional BI v Big Data Analytics
• Traditional BI explains “What happened”. Big Data
Analytics is about “What will happen”
• Traditional BI uses limited data sets, and simple
models. Big data analytics uses many diverse and
uncorrelated raw data sets, and complex predictive
models.
• Traditional BI supports causation: what happened, and
why do we think it happened? Big data analytics is
mostly about correlation: by using multiple unrelated
data sources, we've found a wonderful new insight we
can't entirely explain.
Some Examples of Big Data Analytics
Techniques
•
•
•
•
•
•
Bucket Testing with Big Data: What copy text, layout, images or colours will
improve conversion rates of the web sites
Association Rule Learning: Use Market Basket Analysis, Data for the product
bought together used to determine associations and discover interesting
relationships that can be used for marketing (customers who but bright colour
dresses also buy beauty products!)
Supervised Learning/Classification: Prediction on segment specific customer
behaviour such as buying decisions and consumption rate, using existing big data.
Data Fusion and Integration: Social Media Analysis using Natural Language
Processing, combined with real-time sales data to determine what effect a
marketing campaign has on customer sentiment and purchasing behaviour.
Neural Network: Algorithms that could be used to identify high value customers
that are at risk of leaving or identifying fraud
Network Analysis: Techniques to use network data to identify the flow od data.
Can be used for marketing by finding the influencers and bottlenecks