Young-Bean Song AnalyticsDNA September 18, 1015 Demystifying Big Data
Dec 29, 2015
“Every two days now we create as much information as we did from the dawn of civilization up until
2003” – Eric Schmidt
From chaos to something useful…
Dis
pla
y A
d
Vid
eo A
d
Searc
h
Socia
l
Em
ail
Bra
nd
web
sit
e
Ph
ysic
al
Sto
re
Pu
rch
ase
$88$134$48$32$0$64
Real-time Analytics
$30
$60
$9
$51
$93 $5
7
ROI Value
Ecom
merc
e s
tore
Targ
et
Cu
stom
ers
Devi
ces
Cookie
s
What we are covering today?
Framework for a analytics roadmap
Simplify what all this muck means
Demystify lots of buzz words and hype
Case studies & research
Connecting the dots…
Define Success
•Metrics or proxies
•By customer segments
•How do I want to see the results
•Benchmarks•What decisions will I make from the outcomes?
Methodology
•Experimental Design
•Buy-in across organization
•Prototype•Start fast, simple
Data Alignment
•What data do I need?
•Where can I get it?
•User level or macro level?
•How can I use my own data?
•Start fast, simple
Framework for an Analytics Strategy and Roadmap
Pagliacci’s Definition of Success
• Sales & Profitability• Segment by
• Volume x (Revenue & Profit) • Household or business
• Toppings to infer kids, Day of Week• What to do about it?
• Speed• Retention• Cross-sell product
recommendations
How did IKEA do it?1. Link online surfing to in-store sales2. Inputs: Onsite Behavior 3. Methodology: Multi-linear regression to
predict: In-store sales user level (95% accuracy)
4. Proxy metric that combine:1. Products added to shopping list2. Stock availability3. Local store look ups4. Ikea product searches5. Number of products viewed
Conversion Metrics
How to scale…?? Predictive >> Prescriptive
IKEA Case Study Results
Average basket size of exposed was 45% higher than non-exposed
Campaign drove 4.6 to 1 ROIYOY performance increase 91%Cost per conversion down by 51%
Myth #4 “This is easy…”
External APIs
Custom Insight
Dashboards
Da
Media
TV
Mobile
Online
Radio
Analytics Engine
PrintD
ata
Co
llect
ion
Demographic Data-Demos-Extended demos
OOH Social
Behavioral Data-In-market classification-Online behavior-Offline purchase data
Brand-level data Category level data
Economic data- By vertical
Hyg
ien
e
Weig
ht
& B
ala
nce
ETL
Data Warehouse
Measurement and Optimization
Sales Proxies
Market Factors
Advertising
Impact
Web Services
Report Generator
Re
po
rt D
istr
ibu
tion
Data Mart
Data Collection Analytics Reporting
“Over half of the business leaders today, realize they don’t have access to the insights they need to do their job.” - IBM
Case Study: How to TED Talk
@ 4:10
The Biggest Myth in Marketing:“Advertising makes people do
what I want time to do…”
REALLY GOOD MARKETINGMake us feel good, make us laugh, or at least, help us do what we already want to do,
easier…