Big data –illusion or opportunityMiha Vogelnik, Valicon
Key components of big data
Action
Analytics
Data
TECHNOLOGY
SMARTERDECISIONS
WEB
EVENT
TV
OUT
CATALOG
RADIO
WOM
WOM
WEB SEARCH
WEBSITE
SOCNET
CONTACT
WEB PRICE
MOBILE
WEB RESEARCH
WEBSITE
CONTACT
MOBILE
STORE
STORE PROMO & PUSH
STORE MOB
WEB RESEARCH
SURVEY
WOM
WEBSITE
SOCNET
0% 20% 40% 60% 80% 100%
STIMULATION
RESEARCH
PURCHASE
POSTPURCHASE
Source: Nakupna pot 2012/2013, Valicon/IPROM
OFFLINE53%ONLINE
21%
WOM26%
OFFLINE16%
ONLINE63%
WOM21%
OFFLINE85%
ONLINE15%
Potential sources of customer data
SOCIALMEDIA
INTERNALDATAPOS
TransactionsCall-centre
logsCompany
web portals
Blogs
Surveys
Winninggames
OTHER SOURCES
VARIETY
VOLUME
VELO
CITY
VERACITY
MARKETING OPTIMIZATION – LESS INVESTMENT IN „TRADITIONAL“ MARKETING WITH HIGHER UTILIZATION
8
TRADITIONAL MARKETING
1to1 MARKETING
IDENTIFICATION
General image advertising
Segmented communication
profiling
direct
TRADITIONAL MARKETING
1to1 MARKETING
IDENTIFICATION
Advanced analytics helps find meaning in vast quantity of data
Classification Associations Predictions
Grouping customer based on their common characteristics demographic, behavioral
Understanding relationship between customers in social network or product in shopping basket
Finding what differentiate buyer from not buyer, churner from not churner and make a prediction
WEB
EVENT
TV
OUT
CATALOG
RADIO
WOM
WOM
WEB SEARCH
WEBSITE
SOCNET
CONTACT
WEB PRICE
MOBILE
WEB RESEARCH
WEBSITE
CONTACT
MOBILE
STORE
STORE PROMO & PUSH
STORE MOB
WEB RESEARCH
SURVEY
WOM
WEBSITE
SOCNET
Cust_1 Cust_2 Cust_3
STIMULATION
RESEARCH
PURCHASE
POSTPURCHASE
AQUISITION
AQUISITION
PERSONALISED SELLING
CROSS-SELLING
UP-SELLING
RETENTION
WEB
EVENT
TV
OUT
CATALOG
RADIO
WOM
WOM
WEB SEARCH
WEBSITE
SOCNET
CONTACT
WEB PRICE
MOBILE
WEB RESEARCH
WEBSITE
CONTACT
MOBILE
STORE
STORE PROMO & PUSH
STORE MOB
WEB RESEARCH
SURVEY
WOM
WEBSITE
SOCNET
Cust_1 Cust_2 Cust_3
• Testing which ad/video /tweet will have better impact
• Click-stream analysis to present most suitable ad-content to display
• Lead scoring and managing in order to achieve best conversion
• Most probable next product recommendation
• Combination of most suitable products for cross-sell
• Retaining customers by predicting most probable churners
• Campaign evaluation
STIMULATION
RESEARCH
PURCHASE
POSTPURCHASE
13
Sale
s
Days/Weeks
Traditional campaigns
14
Sale
s
Days/Weeks
Traditional campaigns
Days/weeks
15
Smaller more targeted campaigns
Sale
s„Smart“ optimized campaigns
Days/weeks
16
SMART follow-up campaigns
Smaller more targeted campaigns
Sale
s„Smart“ optimized campaigns
Days/weeks
17
Increase of Sales
SMART follow-up campaigns
Smaller more targeted campaigns
Sale
s„Smart“ optimized campaigns
1. Clearly identify simple problem
2. Identify suitable data
3. Use existing data
4. Analyze data
5. Decide on action
6. Action!
7. Measure the outcome
8. Improve and do it again!
How to start with BigData?
Start with baby steps - NOW!