2. What’s it all about?
“Big Data is about the technologies and practice of handling huge data sets that conventional database management systems
cannot handle them efficiently, and sometimes cannot handle them at all. Often these data sets are fast-streaming too, meaning
practitioners don’t have lots of time to analyze them in a slow, deliberate manner, because the data just keeps coming.
Sources for Big Data include financial markets, sensors in manufacturing or logistics environments, cell towers, or traffic
cameras throughout a major metropolis. Another source is the Web, including Web server log data, social media material
(tweets, status messages, likes, follows, etc.), e-commerce transactions and site crawling output, to list just a few examples.”
(Andrew Brust from ZDNet))
In 2005, humankind created
150 exabytes of information.
In 2011, 1.200 exabytes will
be created. (The Economist)
Volume
The “V” drivers
Worldwide digital content Velocity Variety 80% of enterprise data will be
will double in 18 unstructured, spanning
months, and every 18 traditional and non traditional
months thereafter. (IDC) sources. (Gartner)
2
3. Big Data sources inside Social Business Ecosystem
Wholesale/Retail Outsourcer
My relations My relations
My world My world
Myself Myself
My relations My relations My relations
My world My world My world
Public authority
Myself Myself Myself
Customer
My relations My relations
My world Company My world
Myself Myself
Supplier Partner
3
4. Let’s get a Social CRM definition
“Social CRM is a philosophy and a business strategy, supported by a
technology platform, business rules, processes and social characteristics,
designed to engage the customer in a collaborative conversation in order to
provide mutually beneficial value in a trusted and transparent business
environment. It is the company's programmatic response to the
customer's control of the conversation.“
Paul Greenberg
CRM books author, speaker, consultant, analyst
4
5. The shift from CRM to Social CRM
CRM SOCIAL CRM
Collaborative
- Phone
- Social Network
- Email
- Micro blogging site
- Mail
- Blog
- Fax
- Forum
- Web form
- Collaborative platform
- Face2face
Operational
- Contact & Case Management
- Social media monitoring
- Trouble ticketing Management
- Unified Agent Desktop
- Marketing automation
- Enterprise Collaboration
- SFA
- Collaborative KM
- KM/BPM/ERP integration
Analytical
- VoC
- Data Mining
- Business Intelligence
5
6. The Big Data funnel for Social CRM
Real Life EXTENDED HUMAN EXPERIENCE
Touchpoint
Transactional data Traditional interaction data Web & social data Location-based data
Data streams
Information
Insight
6
7. Now we are plenty of “human” data
Customer
Myself My world My relations
- Geographic: - Information gathering: - Conversation:
Where I live How I compare Where I discuss
Where I work What I compare What I discuss about
Where I spent my holidays What drive my choice How I contribute
- Socio-demographic: What I choose - Psychographic (outspoken):
My age -Transaction: What I like
My gender What I buy What I believe
My family size How I buy What I think about
My income Where I buy What I don’t endure
My occupation When I buy - People:
My education - Usage: What people I relate with
My religion How I use Whom I’m influenced by
My nationality How much I use Who I influence
- Psychographic (formal): Where I use
My lifestyle When I use
My personality - Interaction:
My values Information need
Trouble/problem
Claim
Praise
7
9. So it’s time to really understand your people
EXPERIENCES ATTITUDES
EMOTIONS OPINIONS
9
10. How can we handle it?
“A wealth of information creates a poverty of attention and a need to allocate that attention efficiently among the
overabundance of information sources that might consume it”
Herbert Simon, Economist
"We have free and ubiquitous data, so the complimentary scarce factor is the ability to understand that data and
extract value from it.“
Hal Varian, Google's Chief Economist
Human intervention is fundamental for decision making but we need help
and support to process and understand data because of our cognitive and
time limitations
10
11. What’s in it for me?
Proactive selling Lead generation
Application
Event/Trend detection Proactive routing
Churn prediction Real-time question answering
Fraud detection Location-based marketing
Spatial information analysis Sentiment analysis
DESCRIPTIVE & PREDICTIVE ANALYTICS
Analysis
History information analysis Semantic analysis
Behavioral analysis Opinion extraction & summarization
DOMAIN
Optimization Natural Language Processing
Classification Association rules learning
Methods
Clustering & Factoring Scoring
Regression Ensemble learning
Time Series Analysis Social Network Analysis
Data
Number Free Text Tag Audio Image Video
11
12. Can we trust Analytics?
People are quite confident about numbers but are suspicious of
“unstructured data” algorithms’ output accuracy
Tools can normally reach 80% accuracy but you have to
express skepticism for >95% values (overfitting)
High accuracy doesn’t always mean more positive
business impacts
The 4 “What” on accuracy
What do you need to What scale and What accuracy measures What is the accuracy
measure to accomplish measurement will help you fit your own business impact on business?
your own business tasks? translate sentiment into needs?
business decisions?
You may want to analyze at You may prefer an explicit Most people confuse Not all inaccuracies have
document level (tweet, class or a score. Or maybe accuracy with precision. But equal business impact. You
email, etc.) or at feature level you need more mood than accuracy is a function of may focus your attention
(named entity, concept, valence. precision and recall so only to some kind of errors
topic, etc.) remember that results are and drop others depending
relevant if they can help you on your business objective
respond to a specific
business challenge.
12
13. An example for Social Customer Service
Most probable issue-related contents
Automatic routing to selected CSR
retrieval
High Churn Automatic response
High LTV (real Q&A)
Most frequent issue Most frequent
Churn Score LTV Score
(service request) issue (concept)
Churn Prediction Customer LTV Customer Claim history Concept highlighting Polarity highlighting
Behavioral Analysis / History Information Analysis Opinion extraction / Semantic Analysis / Sentiment Analysis
Transaction / Billing / Payment / Usage / Interaction Conversation / Psychographic / Relations
13
14. Great opportunities but pay attention to the issues
Liability
Data policies Security
“Sharing”
Privacy Data access obstacles
“Sharing”
incentive
Main issues Analytical
to address
culture
Change management
Technology Distributed architectures
Massive parallel processing
Talent
14