© 2015 SAP SE or an SAP affiliate company. All rights reserved. 1Internal
The New Simple: Predictive
Analytics for the Mainstream
Confidentlyanticipateanddrivebetterbusinessoutcomes
Montreal Advanced Analytics Workshop
April 8, 2015
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 2
It’s No Longer Sense and Respond …
Ever Faster
Decision Cycle
Analytical
Skill Gap
“Demand for deep
analytical talent in the US
could be 50 to 60%
greater than its projected
supply by 2018”
McKinsey Global Institute
1 11
Transactions
Conversations
Machines
Massive
Amount of Data
Gartner
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 3
Anticipate What Comes Next and Drive Better
Decisions… Today!
Social
Network
Customer
Data
Automobiles
Machine
Data
Smart Meter
Point of
Sale
Mobile
Structured
Data
Click Stream
Location-
based Data
Text Data
IMHO, it’s great!
RFID
68%of organizations
that used predictive analytics
realized a competitive
advantage
Ventana Research
52% use predictive
analytics to increase
profitability
55% use predictive
analytics to create new
revenue opportunities
45% use predictive
analytics for customer
services
43% use predictive
analytics for product
recommendations
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 4
“I’m building churn
models for every
region”
MS Statistics, Berkeley
Data Scientist
.01%
“I need explain to the
CEO why sales are down
in EMEA”
MBA, U of Pennsylvania
Data Analyst
~3% 97%
Business User
“The app needs to tell
me what offer to make in
real time”
BS French Literature, UC
Davis
Opportunity to Broaden Access to Predictive
5© 2015 SAP SE or an SAP affiliate company. All rights reserved.
SAP PREDICTIVE ANALYTICS
Fast  Simple  Everywhere
MAKE PREDICTIONS SIMPLE,
FAST, AND ACCURATE
Automated predictive workflow
Embedded predictive analytics in business processes and apps
ACT WITH CONFIDENCE AT THE POINT
OF DECISION
Real-time predictive on big data
SPOT OPPORTUNITIES IN
REAL-TIME
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 6
SAPPredictiveAnalytics
SAP Predictive Analytics
Data Preparation
Expert Analysis
Automated AnalysisVisualization
RecommendationScoring Social
SDK/API Model Management
Connectors
CLOUD On-Premise
Predictive Analysis
Library
Automated
Predictive Library
R-Scripts
In-Memory Processing Engine 25+ Industries  11+ LoBs
O&G,
Manufacturing
& Utilities
Public Sector
& Healthcare
Financial &
Insurance
Services
TelecommunicationsRetail &
Consumer
Products
Predictive ApplicationsSAP HANA
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 7
Imagine the Business Potential…
:-)
Brand
sentiment
360-degree
customer view
Product recommendation
Propensity to churn Real-time demand/
supply forecast
Predictive maintenance
Fraud detection
Network optimization Insider threats
Risk mitigation in
real time
Asset
tracking
Personalized care
MANUFAC-
TURING
RETAIL CPG
HEALTH
CARE
BANKING UTILITIES TELCO
PUBLIC
SECTOR
25+ industries
MARKETING
SALES
FINANCE
HR
OPERATIONS
SERVICE
IT
SUPPLY CHAIN
FRAUD/RISK
11+ LoB
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 8
Predictive Use Cases
Increase Your
Customer Reach
Show Predictive
Cause & Effect
Predict & Prevent
Customer Churn
Recommend Next
Best Product or
Service
Detect & Reduce
Fraud
Operations
Optimization
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 9
Predict & Prevent
Customer Churn
o Model: Classification
o Basic premise: Understand who, why, and when customers churn
and create programs/incentives to prevent it
o Outcome: Use historical churn analysis to assign a score/flag for
the entire current and future customer base
o Data: customer profile, purchase behavior, sentiment, social and
related links
o Improvement factors:
o Retain key customers
o Allow influencers to connect more strongly with linked profiles
o Prevent “offer jumping”
o Increase revenue
Skyrock: Unlock Big Data Sources for
MoreAccurate and Personalized
Recommendations
<600k
Reduction in customer churn rate
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 10
o Model: Regression
o Basic premise: Understand the factors, events, timeframes, and
external predictor variables which influence the target (supply and
demand, employee retention, profit and margin, more…
o Outcome: Use historical data and timed events to assign a
weighted score and probability that can be associated to all
observations in the database
o Data: varies depending on use case, but these are usually wide
data sets with many historical events
o Improvement factors:
o Immediately understand influential factors/variables on target
o Visualize predictor correlations and patterns
o Optimize necessary influential factors as a remedy
o Increase x
eBay: Enabling Early Signal Detection with
PredictiveAnalytics and SAPHANA® 97%
Confidence that a signal
is a true positive
Show Predictive
Cause & Effect
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 11
o Model: Clustering & Social Analysis
o Basic premise: Understand the factors, events, timeframes, and
external predictor variables which influence your ability to identify
and target the customers who represent the greatest opportunity for
revenue attainment
o Outcome: Use historical data and timed events to assign a
weighted “grouping” of like customers based on social, purchase,
and demographic variables to ensure more targeted marketing
o Data: customer profile, purchase history, social behavior and
linkages, location information
o Improvement factors:
o More intelligent offers for up-sell/cross-sell
o Use social populations to drive marketing activities
o Visualize geo location “areas of influence”
o Increase revenue
Mobilink: Boosting Campaign Response
Rates with SAP
PredictiveAnalytics
380%
Boost in campaign response
rates, thanks to social network
analysis
Increase Your
Customer Reach
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 12
HOW?
Before SAP Predictive
Analytics
After SAP
Predictive
Analytics
ANY/ALL DATA SOURCES,
ANSWER ANY QUESTION
IN-DATABASE FROM
START TO FINISH
AUTOMATED MODELING
& TUNING PROCESS
IN-DATABASE,
APPLICATION, & PROCESS
INTEGRATION
MODEL MANAGEMENT
& RECALIBRATION
Why SAP?
• EASE OF USE
• AUTOMATION
• IN-DATABASE
• UNDERSTANDABLE
• EMBEDDABLE
Build predictive models to help
create personalized offers more
quickly & accurately
Why SAP Predictive Analytics?
 Precise, accurate, and fast polling of
10 million observations and
800 variables
 Scalable solution to support both
short and long-term needs
Cox Communications: Supercharging Customer
Relationships
14%
More products
per customer
household
80%
Reduction in
model creation
time
28%
Reduction in
customer churn
rate
42X
Greater
throughput for
central analysts
eBay: Enabling Early Signal Detection
500Metrics analyzed
to identify
outliers
100%
Accuracy
97%
Confidence that a
signal is a true
positive
6weeks
Project duration
Separate signals from noise to
identify key changes to the
health of eBay’s marketplace
Why SAP Predictive Analytics?
 Automated signal detection selecting
the best model, increasing the
accuracy of forecasts
 Scalable system that provides real-
time insights on SAP HANA
 Decision tree logic
Unlock Big Data for accurate
predictions and personalized
recommendations on products,
friends, and content
Why SAP Predictive Analytics?
 Ability to offer relevant “friend”
recommendations
 Better understanding of individuals by
identifying communities with similar
interests, characteristics, and
behaviors
Skyrock: Unlock Big Data Sources for MoreAccurate
and Personalized Recommendations
20Relevant friend
recommendations
2X
increase in the
acceptance rate
<600k
Reduction in customer churn rate
20,000Distinct communities identified
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 16
SAPas the Market Leader
Hurwitz
Victor
“fast time to value and
ability to support
very large data sets.”
Victory
Index Report
Forrester
Leader
“SAP is a leader due to a
strong architecture
and strategy.”
Big Data
Predictive Wave
Howard Dresner
Top Vendor
“The top vendors for advanced
and predictive analytics
include SAP.”
The Wisdom
of Crowds
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 17
The SAPDifference
Complete
end-to-end
analytics solution
#1
leader in
analytics*
65,000+
analytics
customers
13,000+
partners with proven
track record of success
*Gartner, Market Share Analysis:
Business Intelligence and Analytics Software, 2013
© 2015 SAP SE or an SAP affiliate company. All rights reserved. 18
1
Next
Steps
Read customers case studies
www.sap.com/predict-and-me
Learn more and watch it in action
www.sap.com/predictive & http://scn.sap.com/docs/DOC-32651
2
Connect with you SAP representative3
19© 2015 SAP SE or an SAP affiliate company. All rights reserved.
Thank you
www.sap.com/predictive
scn.sap.com/community/predictive-analysis
#sappredictive  @sapanalytics
Contact information:
Mike Watschke
mike.watschke@sap.com
Conférence SAP Analyse Prédictive
Montréal - 08 avril 2015
Le pouvoir de prédire les conditions futures du marché et les
besoins et désirs des consommateurs est l’envie de chaque
dirigeant d'entreprise.
LE VRAI POUVOIR, C’EST LA CONNAISSANCE
COMPRENDRE L’ANALYSE PRÉDICTIVE
L‘analyse Prédictive est décrite comme un domaine d'analyse
statistique qui s'occupe d'extraire des informations à partir
des données et de l'utiliser pour prédire les tendances et
comportements futurs.
PARCOURS ANALYTIQUE DDPP
Descriptive – Que s’est-il passé?
Rapports de performances (vues sommaires des activités)
Diagnostic– Pourquoi cela s’est produit?
Analyses “AdHoc” et modèles multi-dimentionels
Prédictive – Que va t’il se passer?
Modèles d’analyses - tendances et comportements
Prescriptive – Que faire?
Modèles d’analyses – plusieurs scénarios (prédictions et résultats)
Hindsight
Insight
Foresight
Besoins d’affaires
Lorsque vous envisagez l’analyse Prédictive….
Adhésion de
l’organisation
Engagement de la
haute direction
La donnée L’expertise
Conclusion
Pour les organisations qui veulent aller vers du Prédictif,
agileDSS peut vous accompagner dans l’évaluation
organisationnelle « organisational readiness assessment »
Merci
Contact:
Nathalie Maslia
maslia.nathalie@agiledss.com
www.agiledss.com

Déjeuner Conférence - L'analyse prédictive agile avec SAP Predictive Analytics 2.0

  • 1.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 1Internal The New Simple: Predictive Analytics for the Mainstream Confidentlyanticipateanddrivebetterbusinessoutcomes Montreal Advanced Analytics Workshop April 8, 2015
  • 2.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 2 It’s No Longer Sense and Respond … Ever Faster Decision Cycle Analytical Skill Gap “Demand for deep analytical talent in the US could be 50 to 60% greater than its projected supply by 2018” McKinsey Global Institute 1 11 Transactions Conversations Machines Massive Amount of Data Gartner
  • 3.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 3 Anticipate What Comes Next and Drive Better Decisions… Today! Social Network Customer Data Automobiles Machine Data Smart Meter Point of Sale Mobile Structured Data Click Stream Location- based Data Text Data IMHO, it’s great! RFID 68%of organizations that used predictive analytics realized a competitive advantage Ventana Research 52% use predictive analytics to increase profitability 55% use predictive analytics to create new revenue opportunities 45% use predictive analytics for customer services 43% use predictive analytics for product recommendations
  • 4.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 4 “I’m building churn models for every region” MS Statistics, Berkeley Data Scientist .01% “I need explain to the CEO why sales are down in EMEA” MBA, U of Pennsylvania Data Analyst ~3% 97% Business User “The app needs to tell me what offer to make in real time” BS French Literature, UC Davis Opportunity to Broaden Access to Predictive
  • 5.
    5© 2015 SAPSE or an SAP affiliate company. All rights reserved. SAP PREDICTIVE ANALYTICS Fast  Simple  Everywhere MAKE PREDICTIONS SIMPLE, FAST, AND ACCURATE Automated predictive workflow Embedded predictive analytics in business processes and apps ACT WITH CONFIDENCE AT THE POINT OF DECISION Real-time predictive on big data SPOT OPPORTUNITIES IN REAL-TIME
  • 6.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 6 SAPPredictiveAnalytics SAP Predictive Analytics Data Preparation Expert Analysis Automated AnalysisVisualization RecommendationScoring Social SDK/API Model Management Connectors CLOUD On-Premise Predictive Analysis Library Automated Predictive Library R-Scripts In-Memory Processing Engine 25+ Industries  11+ LoBs O&G, Manufacturing & Utilities Public Sector & Healthcare Financial & Insurance Services TelecommunicationsRetail & Consumer Products Predictive ApplicationsSAP HANA
  • 7.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 7 Imagine the Business Potential… :-) Brand sentiment 360-degree customer view Product recommendation Propensity to churn Real-time demand/ supply forecast Predictive maintenance Fraud detection Network optimization Insider threats Risk mitigation in real time Asset tracking Personalized care MANUFAC- TURING RETAIL CPG HEALTH CARE BANKING UTILITIES TELCO PUBLIC SECTOR 25+ industries MARKETING SALES FINANCE HR OPERATIONS SERVICE IT SUPPLY CHAIN FRAUD/RISK 11+ LoB
  • 8.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 8 Predictive Use Cases Increase Your Customer Reach Show Predictive Cause & Effect Predict & Prevent Customer Churn Recommend Next Best Product or Service Detect & Reduce Fraud Operations Optimization
  • 9.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 9 Predict & Prevent Customer Churn o Model: Classification o Basic premise: Understand who, why, and when customers churn and create programs/incentives to prevent it o Outcome: Use historical churn analysis to assign a score/flag for the entire current and future customer base o Data: customer profile, purchase behavior, sentiment, social and related links o Improvement factors: o Retain key customers o Allow influencers to connect more strongly with linked profiles o Prevent “offer jumping” o Increase revenue Skyrock: Unlock Big Data Sources for MoreAccurate and Personalized Recommendations <600k Reduction in customer churn rate
  • 10.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 10 o Model: Regression o Basic premise: Understand the factors, events, timeframes, and external predictor variables which influence the target (supply and demand, employee retention, profit and margin, more… o Outcome: Use historical data and timed events to assign a weighted score and probability that can be associated to all observations in the database o Data: varies depending on use case, but these are usually wide data sets with many historical events o Improvement factors: o Immediately understand influential factors/variables on target o Visualize predictor correlations and patterns o Optimize necessary influential factors as a remedy o Increase x eBay: Enabling Early Signal Detection with PredictiveAnalytics and SAPHANA® 97% Confidence that a signal is a true positive Show Predictive Cause & Effect
  • 11.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 11 o Model: Clustering & Social Analysis o Basic premise: Understand the factors, events, timeframes, and external predictor variables which influence your ability to identify and target the customers who represent the greatest opportunity for revenue attainment o Outcome: Use historical data and timed events to assign a weighted “grouping” of like customers based on social, purchase, and demographic variables to ensure more targeted marketing o Data: customer profile, purchase history, social behavior and linkages, location information o Improvement factors: o More intelligent offers for up-sell/cross-sell o Use social populations to drive marketing activities o Visualize geo location “areas of influence” o Increase revenue Mobilink: Boosting Campaign Response Rates with SAP PredictiveAnalytics 380% Boost in campaign response rates, thanks to social network analysis Increase Your Customer Reach
  • 12.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 12 HOW? Before SAP Predictive Analytics After SAP Predictive Analytics ANY/ALL DATA SOURCES, ANSWER ANY QUESTION IN-DATABASE FROM START TO FINISH AUTOMATED MODELING & TUNING PROCESS IN-DATABASE, APPLICATION, & PROCESS INTEGRATION MODEL MANAGEMENT & RECALIBRATION Why SAP? • EASE OF USE • AUTOMATION • IN-DATABASE • UNDERSTANDABLE • EMBEDDABLE
  • 13.
    Build predictive modelsto help create personalized offers more quickly & accurately Why SAP Predictive Analytics?  Precise, accurate, and fast polling of 10 million observations and 800 variables  Scalable solution to support both short and long-term needs Cox Communications: Supercharging Customer Relationships 14% More products per customer household 80% Reduction in model creation time 28% Reduction in customer churn rate 42X Greater throughput for central analysts
  • 14.
    eBay: Enabling EarlySignal Detection 500Metrics analyzed to identify outliers 100% Accuracy 97% Confidence that a signal is a true positive 6weeks Project duration Separate signals from noise to identify key changes to the health of eBay’s marketplace Why SAP Predictive Analytics?  Automated signal detection selecting the best model, increasing the accuracy of forecasts  Scalable system that provides real- time insights on SAP HANA  Decision tree logic
  • 15.
    Unlock Big Datafor accurate predictions and personalized recommendations on products, friends, and content Why SAP Predictive Analytics?  Ability to offer relevant “friend” recommendations  Better understanding of individuals by identifying communities with similar interests, characteristics, and behaviors Skyrock: Unlock Big Data Sources for MoreAccurate and Personalized Recommendations 20Relevant friend recommendations 2X increase in the acceptance rate <600k Reduction in customer churn rate 20,000Distinct communities identified
  • 16.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 16 SAPas the Market Leader Hurwitz Victor “fast time to value and ability to support very large data sets.” Victory Index Report Forrester Leader “SAP is a leader due to a strong architecture and strategy.” Big Data Predictive Wave Howard Dresner Top Vendor “The top vendors for advanced and predictive analytics include SAP.” The Wisdom of Crowds
  • 17.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 17 The SAPDifference Complete end-to-end analytics solution #1 leader in analytics* 65,000+ analytics customers 13,000+ partners with proven track record of success *Gartner, Market Share Analysis: Business Intelligence and Analytics Software, 2013
  • 18.
    © 2015 SAPSE or an SAP affiliate company. All rights reserved. 18 1 Next Steps Read customers case studies www.sap.com/predict-and-me Learn more and watch it in action www.sap.com/predictive & http://scn.sap.com/docs/DOC-32651 2 Connect with you SAP representative3
  • 19.
    19© 2015 SAPSE or an SAP affiliate company. All rights reserved. Thank you www.sap.com/predictive scn.sap.com/community/predictive-analysis #sappredictive  @sapanalytics Contact information: Mike Watschke mike.watschke@sap.com
  • 20.
    Conférence SAP AnalysePrédictive Montréal - 08 avril 2015
  • 21.
    Le pouvoir deprédire les conditions futures du marché et les besoins et désirs des consommateurs est l’envie de chaque dirigeant d'entreprise. LE VRAI POUVOIR, C’EST LA CONNAISSANCE
  • 22.
    COMPRENDRE L’ANALYSE PRÉDICTIVE L‘analysePrédictive est décrite comme un domaine d'analyse statistique qui s'occupe d'extraire des informations à partir des données et de l'utiliser pour prédire les tendances et comportements futurs.
  • 23.
    PARCOURS ANALYTIQUE DDPP Descriptive– Que s’est-il passé? Rapports de performances (vues sommaires des activités) Diagnostic– Pourquoi cela s’est produit? Analyses “AdHoc” et modèles multi-dimentionels Prédictive – Que va t’il se passer? Modèles d’analyses - tendances et comportements Prescriptive – Que faire? Modèles d’analyses – plusieurs scénarios (prédictions et résultats) Hindsight Insight Foresight
  • 24.
    Besoins d’affaires Lorsque vousenvisagez l’analyse Prédictive…. Adhésion de l’organisation Engagement de la haute direction La donnée L’expertise
  • 25.
    Conclusion Pour les organisationsqui veulent aller vers du Prédictif, agileDSS peut vous accompagner dans l’évaluation organisationnelle « organisational readiness assessment »
  • 26.