Big Data? Right Data! Chris Dobson: Director of Consultancy, Aquila Insight Email: [email protected] Web: www.aquilainsight.com Twitter: @aquilainsight @Mr_Dobbo
Big Data? Right Data!Chris Dobson: Director of Consultancy, Aquila Insight
Email: [email protected]
Web: www.aquilainsight.com
Twitter: @aquilainsight @Mr_Dobbo
Data’s always been big
2www.aquilainsight.com
But big tech is the real enabler
3www.aquilainsight.com
4www.aquilainsight.com
But big tech is the real enabler
5www.aquilainsight.com
But big tech is the real enabler
6www.aquilainsight.com
But big tech is the real enabler
7www.aquilainsight.com
But big tech is the real enabler
But first think….
8www.aquilainsight.com
What am I trying to achieve?
Is it right for the customer?
9www.aquilainsight.com
If HR did (bad) Big Data……
10www.aquilainsight.com
• Clear job spec
• Identified skill set needed
• Initial recruitment
• First interviews
Learning from HR
• Final interviews
• Job offer
• New join
• Development
11www.aquilainsight.com
• Understand the question
• Identify the what, when &
how often
• Do data discovery
• Remove the noise
Learning from HR
• Refine
• Ensure you have the right
permissions
• Use for purpose
• Measure, Adapt & Develop
Data pillars
12www.aquilainsight.com
•All data
storage and
transfer to
be done in a
secure way,
with set
procedures
and access
levels
•Clear
security
guidelines
and disaster
recovery
Secure
•Full legal
compliance
needs to be
adhered to
at all times,
from a
customers
T’s & C’s,
through to
data usage
Compliant
•Levels of
data
accuracy
should be
known, with
set levels of
tolerance for
missing data
Accurate
•Data,
analysis and
decisioning
need to be
available at
the right
time to
impact
business and
customer
decisions
Timely
•Analysis:
Enable
statistically
robust
analytics
without
impacting
performance
•Campaign:
Support
legally
compliant,
multi-variant
campaigns
with optimal
structure
Actionable
•The data
structure
should allow
for the
integration
of new data
types, and
the
derivation of
existing
variables
Scalable
•All data held
should impact
the business in
some way:
Measurement,
Analysis or
Decision
Making
Efficient
Real time multi-channel
consumer interaction
led
Optimising the digital
conversationDrive loyalty & retention via NBA
led activity
Increasing usage to drive
incremental revenue
Product UsageCustomer Development Digital Optimisation Customer Care
ConfidentialPA120/02/201222 22 Xperia Lounge PackagesOctober’14
Creating Stickiness and Value through One SonyIntegrated One Sony propositions for satisfaction, loyalty and acquisition
Logo design and packages names conceptual and TBC
PREMIUM ELITE
Enabling multi-device
experience evolution
Strategy, ATL targeting
optimisation, & competitor
tracking
Product and services design and
innovation, & support B2B goal
Finance, Product and
Application performance
monitoring and growth hacking
Performance TrackingProduct & Service Development Media Effectiveness IOT Enabler
DATA DRIVEN DECISIONING
What data can really do
www.aquilainsight.com
Commercialised consumer satisfaction modelling driving business investment, and
customer experience development
Business Problem
Sales based CRM
approach with no clear
buy in to wider
customer development
Consumer satisfaction
and NPS understood to
be important but could
not be measured,
commercialised or
drivers understood
Solution
Initial research aligned
with product & business
offers and services
CSI built and linked to
NPS & drivers modelled
Set business change
process, linking business
metrics/ROI
Move towards
behavioural CSI to
impact customer
journey development
14
Making a better customer experience
www.aquilainsight.com
Data: Structured, Research, Financial, Operational,
Behavioural
Creating business change
www.aquilainsight.com
Enabling the business to develop the customer journey through the proactive use of
complaint data
Business Problem
Reduce the volume of
complaints received by
40% per annum across
more than 200 distinct
complaint types
Customers can complain
through different
channels including
direct, branch, social
and web
Solution
Attribution: Identify
the interaction that
caused a complaint
Visualisation: ‘Set the
scene’ to understand
complaints landscape
Pre-emptive: Identify
customers that are
likely to complain
Predictive: Model the
propensity to complain,
allowing for specific
recovery journeys at a
customer level
Data: Structured & Unstructured, Complaint,
Customer Interaction, Value, Channel, Staff
Identifying issues
www.aquilainsight.com
Utilising imagery to create a predictive model for eye disease
Problem
The goal is to identify
the likelihood of the
patient suffering from
neovascularisation using
fundus colour pictures
Solution
Observe the pictures
and in particular the
vessels around the optic
nerve to look for
abnormal thin vessels
Identify the area of
interest within the
image, and create a set
of characteristics that
will explain the disease
Build model using
derived metadata. Test,
train and prove model
Data: Structured, Unstructured, Created, Imagery,
Medical
The truth is out there
17www.aquilainsight.com
• Be clear on the business question…but don’t
limit your imagination
• Ensure you have the right legal permissions,
giving customers clarity and understanding
• Invest in the right tech to meet your current
and future need, not because it’s shiny
• Look for the pattern in the stars,
you never know what you might
find!
Edinburgh Office:Argyle House, 3 Lady Lawson Street, Edinburgh EH3 9DR
London Office:18 King William Street, Monument, London, EC4N 7BP
Derby Office:Office 4, Sadler Bridge Studios, Bold Lane, Derby, DE1 3NT
T 0131 290 2300E [email protected]
/aquilainsight
@aquilainsight
www.aquilainsight.com