#askSAP Analytics Innovations Community Call: Reimagine Analytics for the Digital Enterprise

#askSAP Analytics Innovations Community Webcast

Reimagine Predictive Analytics forthe Digital EnterpriseAugust 31, 2016

© 2016 SAP AG or an SAP affiliate company. All rights reserved. 2

Legal disclaimer

The information in this presentation is confidential and proprietary to SAP and may not be disclosed without the

permission of SAP. This presentation is not subject to your license agreement or any other service or subscription

agreement with SAP. SAP has no obligation to pursue any course of business outlined in this document or any

related presentation, or to develop or release any functionality mentioned therein. This document, or any related

presentation and SAP's strategy and possible future developments, products and or platforms directions and

functionality are all subject to change and may be changed by SAP at any time for any reason without notice. The

information in this document is not a commitment, promise or legal obligation to deliver any material, code or

functionality. This document is provided without a warranty of any kind, either express or implied, including but not

limited to, the implied warranties of merchantability, fitness for a particular purpose, or non-infringement. This

document is for informational purposes and may not be incorporated into a contract. SAP assumes no

responsibility for errors or omissions in this document, except if such damages were caused by SAP´s willful

misconduct or gross negligence.

All forward-looking statements are subject to various risks and uncertainties that could cause actual results to

differ materially from expectations. Readers are cautioned not to place undue reliance on these forward-looking

statements, which speak only as of their dates, and they should not be relied upon in making purchasing

decisions.

SAP Analytics Innovations: Community Call Series

• Quarterly series for the Analytics community hosted by SAP Analytics

• An opportunity for you to direct the discussion, get your questions answered,

and end the session with some useful advice

• Live and interactive 90 minutes

• Connect on topics before, during, and after the call via twitter using #askSAP


Ashish MorzariaGlobal GTM Director,

Advanced Analytics

@AshishMorzaria

Greg MyersSAP Mentor

@gpmyers

Today’s Speakers

Richard MooneyLead Product Manager

for Advanced Analytics

@richardjmooney

https://twitter.com/AshishMorzaria

https://twitter.com/gpmyers

https://twitter.com/richardjmooney

INTRODUCTION TO

SAP BusinessObjects Predictive Analytics

Product and Use Cases


Everything we touch… Every good we purchase…

In the New Digital Economy, Everything is Digitized and Tracked

Every transaction we conduct…


Customers Operational Margin Growth

How do you personalize each

interaction across all channels?

How do you improve your performance across

thousands of processes and decisions?

How do you create new products,

services, and business models?

The Digital Economy To Your Advantage…


Early Adopters Are Winning

In the next 10 years, 40% of the S&P 500 will no longer

exist if they do not keep up with these technology trends*

+9%Revenue

creation

+26%Market

valuation

+12%Impact on

profitability

* “The Digital Advantage: how digital leaders outperform their peers in every industry”: CapGemini and MIT Sloan

Those Embracing Digital Transformation are Outperforming


The Power of Predictive

Unlocks Big Value:the need for Predictive

68%of organizations using predictive analytics

realized competitive advantages.

60%of fraudulent transactions have stopped

using predictive.

28%reduction in customer churn rate with predictive.

• Use historical data to predict behaviors or outcomes

• Answer “what-if” questions

• Ensure employees have what they need to make

optimized decisions

• Fully leverage customer relationships with better insight

• Make meaningful sense of Big Data


SAP

BusinessObjects

Predictive Analytics

Data PreparationCreate meaningful and

reusable data sets

Automated AnalyticsReduce time and skills required

to create accurate models with

repeatable workflow

With Big DataUse Hadoop data with automated

techniques directly in Spark

Ultimate Flexibility

for AlgorithmsUse off-the-shelf algorithms or

bring specialized ones – such

as R functions

Accurate Results in Days, Not WeeksFor everyone: perfect for Analysts AND Data Scientists

Native in-memory SolutionSAP HANA optimized for on-the-fly

predictive data processing


Native In-Memory Predictive Analytics


SAP HANAReal-time in-memory predictive analytics platform

R Scripts

Execution of R scripts via

high-performing parallelized

vector based connection;

R scripts embedded as part

of overall query plan

Application Function

Library (AFL)

Application Function Library

(AFL) framework allows SAP,

partner, and customers to

develop, deploy, load, and

leverage their own advanced

analytic custom functions in

SAP HANA

Custom Open Source

R-Server

SAP HANA

Other Native

Libraries

© SAP AG or an SAP affiliate company. All rights reserved. 13

SAP HANAReal-time in-memory predictive analytics platform

R Scripts

Execution of R scripts via

high-performing parallelized

vector based connection;

R scripts embedded as part

of overall query plan

Application Function

Library (AFL)

Application Function Library

(AFL) framework allows SAP,

partner, and customers to

develop, deploy, load, and

leverage their own advanced

analytic custom functions in

SAP HANA

Custom Open Source

Accelerated predictive

analysis and scoring with

native in-database

algorithms

Predictive Analysis

Library (PAL)

SAP

Predictive

Analysis

Library

Automated

Predictive Library

(APL)

The predictive analysis

capabilities of SAP’s

Predictive automated

analytics engine

(formerly KXEN) in

SAP HANA

Automated

Predictive

LibraryR-Server

SAP HANA

Other Native

Libraries

APL: Automated Algorithms

Native implementation of automated predictive

algorithms: Regression

Clustering

Forecasting

Recommendation

Social Network Analysis

No data extraction required

Fully accessible from “Automated” and “Expert”

interfaces

PAL: Data Scientist Algorithms

Aims to supply most commonly used data

science algorithms (80/20 rule) natively

90+ natively coded algorithms (C++)

Freely mixable with APL algorithms

No data extraction required

R: Open Source Data Scientist Algorithms

8500+ algorithms available

Full support for custom coding

Requires data extraction (externalized process

to HANA)

Fully integrated development when using SAP

PA Suite license


Traditional Analytics Versus In-Memory Predictive Analytics

Predictive

Analysis

Library

Automated

Predictive

Library

R-Server

SAP HANA

Other Native

Libraries

• Create and apply models on very large datasets

within SAP HANA or in a Hadoop storage

transparently connected to SAP HANA

• Real-time predictions recommendations: integrate

predictive models into processes

• Native integration with SAP HANA for ERP and BW,

to provide in-applications predictive modeling

1. Copy data from transactional and external sources

2. Extract data from storage, convert & clean for analytics

3. Download analytical results & load into predictive

analytics application

4. Transfer predictive scoring results into database

SAP BusinessObjects


vs.


Support for SAP HANA Smart Data Streaming

• Automated Analytics now supports HANA Spark

Data Streaming

• Generates CCL Code which can be deployed to

HANA SDS

• Smart Data Streaming Use Cases

o IOT Data for Predictive Maintenance and Quality

o Clickstream analysis for Marketing

o Connected Retail

HANA Smart Data

Streaming


Automated

Modeller


Predictive Analytics on Big Data


Existing Process: entire dataset is transferredConnectivity = SQL only

FULL (Big Data)

dataset is transferred

for processingDataset BIG Datasets Dataset

Big Data

SQL Engines

(Spark SQL,Hive)010001100100

100101001011

100010010101

010011110101

010001100100

100101001011

100010010101

010011110101

010001100100

100101001011

100010010101

010011110101

Traditional Predictive Analytics

Data

Warehouse

RDBMS

Data platform…• power not being

leveraged properly

• just transfers data

Modeler..• Pulls in data, processes,

• Pulls in more data, processes…


Traditional

Application

Leverage Hadoop + Spark = big data store + application platform

Processing on a single server

Data Transfer

CPU/Memory scales dynamically

Processing on 100’s-1000’s of nodes

Hive QLSQL

Database

Native Application

Limited CPU/Memory


With Native Spark Modelling, processing closer to data in Hadoop

FULL training dataset

is transferred

No dataset transfer required!

Data platform…

• runs the Spark application

• processing close to dataNative Spark Connectivity


Native Spark Modelling


• controls the process



Execute automated predictive models directly on Hadoop

using the Apache Spark engine

• Push the data intensive modeling workload to Native Spark -

Classification and Regression models supported

• Model Lifecycle management on Hadoop with RETRAIN and APPLY

• User structure and custom cutting strategy supported on Native Spark

• Real Time Scoring via Spark Streaming API

Benefits

• No data transfer – heavy lifting operations brought close to data

• Faster response times – 7 to 10 times performance gains

• Higher scalability – scale your training process with wider and data more

models

• Better utilization of CPUs – in distributed Hadoop environment

• Abstraction – Analysts can work with Big Data seamlessly

HDFS

(Hadoop Distributed File System)

Hive

(SQL)Spark SQL

Model Lifecycle Manager (Factory)

Scorer

Predictive Analytics Data Manager

In-DB

scoring

(Spark /Hive

QL)

Analytics

Dataset

Definition

Layer

Advanced

Analytics

Execution

LayerSpark

Streaming

(Java

Export)

Modeler -

Training

Native

Spark


Traditional Big Data vs. Big Data with SAP BusinessObjects


Next?

Who?

How?

Big Data Analytics SAP BusinessObjects Predictive Analytics

Code Wizard Based Approach with GUI for End-Users

Big Data Developers

Ideal Tool for use by both a Data Scientist and a

Business Analyst OR Citizen Data Scientist

Data Scientists

Manually Deployed &

Monitored

Automated Deployment & Monitoring using

Predictive Factory


Bringing The Gift of Predictive Insightto Business Intelligence


Descriptive (Business Intelligence) vs Predictive Analytics

Business Intelligence Predictive Analytics

• Who are my most valuable customers? • Who will be my most valuable customers next month?

• Who could become my most valuable customer and why?

• What are my most important products? • What will be my most important products?

• What products could become my most valuable products?

• What are my most successful promotions? • What promotions should I run?

• What promotions could be a good idea to run in the future?

• When did customer X visit my store last? • What is the chance of customer X visiting in the next 2

weeks?

• What were the most profitable products for

customers in my loyalty program?

• What products should I focus on to increase my profit from

customers in my loyalty program?


Smarter BI that goes beyond visual

analysis into insights that cannot hide

Predictive dashboards that

prescribe and can trigger actions

Reports that include reasons and

recommendations on next steps

Move from Descriptive to Predictive BI


Model deployed using

In-Database-Apply

Customer Database

Hancock, John M 38 D Y 4.2 N Y

Doe, Jane F 45 M Y 9.4 N N

Red, Simply F 18 S N 2.1 N Y

SQL Dataset w/ Scoring

Business Users can get on-the-fly

scoring without even knowing they

are using predictive algorithms

BI Artifact

(or even just a dataset)

SAP BI (3.x/4.x)

Embedded into any application

Cloud Applications (SaaS/PaaS/IaaS)

SQL

(Or any other application)

Embedding Predictive Analytics into BI Workflows

https://cloud.saplumira.com/explorer?itemId=53044C6E5B755FF7E10000000A4E423F&type=HANALYTIC&initialMode=View

https://cloud.saplumira.com/explorer?itemId=53044C6E5B755FF7E10000000A4E423F&type=HANALYTIC&initialMode=View

26© 2016 SAP AG or an SAP affiliate company. All rights reserved.

Hancock, John M 38 D Y 4.2 N ?

Doe, Jane F 45 M Y 9.4 N ?

Red, Simply F 18 S N 2.1 N ?

Model

NEW Data

(Current Customers)


Doe, Jane F 45 M Y 9.4 N N




Targeted List

(CR)

Significantly increase ROI through dataset reduction:

• Lower campaign costs by targeting those most likely to leave

• Increase response rate by targeting even more specifically on other attributes

• Increase C-Sat by not hassling loyal customers

Name Gender Age Marital Recent Activity C-Sat Renewed

Before

Predicted

Churn

Customer not expected to

churn, so don’t bother them!

Analysis

(WEBI / Lumira)

Batch scoring

#askSAP Q&A


Scale to large numbers

of Predictive Models

© 2016 SAP SE or an SAP affiliate company. All rights reserved. 29Public

Sales andMarketing Operations

Fraudand Risk

Financeand HR

OtherSectors

• Churn Reduction

• Customer Acquisition

• Lead Scoring

• Product Recommendation

• Campaign Optimization

• Customer Segmentation

• Next Best Offer/Action

• Predictive Maintenance

• Load Forecasting

• Inventory/Demand

Optimization

• Product Recommendation

• Price Optimization

• Manufacturing Process Opt.

• Quality Management

• Yield Management

• Fraud and Abuse Detection

• Claim Analysis

• Collection and Delinquency

• Credit Scoring

• Operational Risk Modeling

• Crime Threat

• Revenue and Loss Analysis

• Cash Flow and Forecasting

• Budgeting Simulation

• Profitability and

Margin Analysis

• Financial Risk Modeling

• Employee Retention

Modeling

• Succession Planning

• Life Sciences

• Health Care

• Media

• High Education

• Public Sector /

Social Sciences

• Construction and Mining

• Travel and Hospitality

• Big Data and IoT

Solve Real Business ProblemsBy Optimizing Resources and Improving Margins


Predictive Process P

rob

lem

Ide

ntified

Bu

sin

ess

Re

su

ltsIdentify

Relevant

Variables

Aggregate

Prepare Data

Derived

Features &

Encode

Variables

Develop

ModelsDebrief models

Write Code for

Database

Execution


Value of SAP’s Predictive Automation

What SAP BusinessObjects Predictive Analytics does for automation:

Data Manager:

• Generate SQL for

• HANA

• Hadoop:

o HIVE, SparkSQL

• All major databases

Auto-algorithms:

Make this section obsolete

Auto-algorithms:

Numbers, strings, dates

Categorical, continuous,

textual

Date parts

Composite variables

(example: position from

latitude and longitude)

Auto-algorithms:

Classification,

regression, clustering,

times series, key

influencers

Link analysis,

recommendations

HANA (APL)

Hadoop (Scala)

Auto-algorithms:

All descriptive statistics

available

Key influencers,

decision trees,

segments, optimal

binning and banding

Communities

In-Database Apply:

Automated SQL

generation

Optimized with data

manager

Hadoop:

HIVE, SparkSQL,

Streaming (Java)

Pro

ble

m

Ide

ntified

Bu

sin

ess

Re

su

ltsIdentify

Relevant

Variables

Aggregate

Prepare Data

Derived

Features &

Encode

Variables

Develop

ModelsDebrief models

Write Code for

Database

Execution


The Predictive Factory

• Manage the lifecycle of predictive models created in


• Automatically retrain, apply, test for deviation and

forecast your models

• Robustly embed predictive analytics at scale in

business processes

Key benefits

• Manage thousands of models easily and robustly

• Automate model refresh and application

• No scripting needed

• Multi-User, collaborative experience


Predictive Factory Features

Segmented Modelling• Take a dataset with thousands of

segments. e.g. Retail outlets, market

segments, geographies, products,

machines ….

• Build a model for one segment using

Automated Modeler. Import the Model

into Predictive Factory

• Segment the model in Predictive Factory

to build models for every other segment

with the same model parameters and

configuration

• Scalable to thousands of segments

• Supports Time Series in 3.0

External Commands• Run Data Preparation using external tools

• Run external, non PA Predictive Models

Sales

EMEANorth America

Product 1

Q1 Forecast

Q2 Forecast

Product 2

Q1 Forecast

Q2 Forecast

Product 3

APAC MEE

Build thousands of models

in a single operation


SAP BusinessObjects Predictive Analytics 3.0Simplify Next Generation UI

Streamlined Predictive User

Experience and Workflow• Modern design principles based on

Fiori UX and HTML 5 for a

completely reimagined user

experience

• Personalized, responsive and

simple user experience across

devices and deployment options

• In-app notifications

• X-Ray support for In-App

Contextual Help to ease first time

user experience

DemoDemo































The Difference

Before SAP

BusinessObjects


After SAP

BusinessObjects


Answer any/all questions with

any/all data sources –

No limits!

In-database automated dataset

production -

No data movement!

Automated modeling and tuning

process -

Focus on accurate results, not

algorithms or code!

Native in-database and

application/process deployment -

Embed and consume for immediate

results!

On-going model management and

recalibration -

No rework necessary!

Days


Predictive Analytics and SAP Applications


Value for Business Users

• Take advantage of predictive analytics

and machine learning without Data

Science expertise

• Discover new insights in your data,

improving your business process

powered by predictive

Automated, Guided and Trusted Experience

Guided Analysis designed for Business Users,

featuring the power of Exploratory Analytics

New Discoveries

We guide you on your journey to find the answer

to your questions

Guided Machine

Discovery as Part of

SAP BusinessObjects Cloud


• Perform an embedded predictive

forecast in their planning model

• Predictive forecast runs a time series

algorithm on historic data in order to

predict future values considering trend,

cycles and/ or fluctuation.

• It can be leveraged to aid the planning

process using a data-driven approach.

Predictive Forecast as Part

of SAP BusinessObjects

Cloud


Detect fraud earlier to reduce financial loss

o Leverage the power and speed of

SAP HANA

o Integration into business processes

o Alert notification and management

Improve the accuracy of detection at less cost

o Minimize false positives with real-time

simulations

o Ability to handle ultra-high volumes

of data by leveraging SAP HANA

Predict & Prevent and deter fraud situations

o Detection based on rules and predictive

analytics to adapt to changing fraud patterns

SAP Fraud Management

with Predictive Analytics

#askSAP Q&A


Customer Case Study

Stella Predictive Analytics

• SAP BusinessObjects Predictive Analytics for Automated Analytics and rapid prototyping of our models

• Forward engineered into SAP HANA for real-time predictions using native, logistical regression model

• This approach allowed for identification of key predictors that more heavily influence a behavioral health outcome

• Run as a pilot to rapidly prototypethe concepts 8

Weeks for Pilot

99%Prediction Accuracy

“This tool will allow me to completely redesign the clinical

process and provide the right amount of care at the right

time. ” – Executive Director of Mental Health Provider

Stella User Experience

• Seamless UX integration• Allows for up to the

minute prediction on incoming jail records

• Flags important predictive factors for clinician

• Enables real time decision support for accurate resource allocation

Stella

This pilot allowed SAP Partner, EV Technologies, to assist Harris Logic through a successful SAP HANA pilot and later, into a cloud based architecture.

Phase 1 – Pilot – Stella 3.0 – Q1 2016• Develop use cases organized by cost, time to

deliver, and return on investment• Executed a migration of the needed JAVA

application components to SAP HANA• Successfully modelled the first two predictive

models and integrated into the pilot application – high utilizers and propensity to recidivize

Phase 2 – Stella 3.0 – June 2016• Full implementation running SAP HANA and

SAP BusinessObjects on AWS • Transitioned all pilot code to next-generation

Stella 3.0

Phase 3 – Stella 3.x+ - Q3 2016• Selected as strategic partner for the new 18

month roadmap• Developed use cases for remaining SAP

HANA capabilities including Text Analysis and the Spatial Engine

• Prioritized remaining use cases into release schedule

Questions?Eric [email protected]

75

#askSAP Final Q&A


Online Resources

Key links

Roadmaps on SAP Service Marketplace http://service.sap.com/saproadmaps

SAP Community Network http://scn.sap.com/

Predictive Analytics Community http://scn.sap.com/community/predictive-analytics

30 days Trial Download https://www.sap.com/trypredictive

SAP BusinessObjects Predictive Analyticshttp://sap.com/predictive

Where to go to provide product feedback and ideas

SAP Idea Place https://ideas.sap.com

Predictive Idea Place https://ideas.sap.com/PredictiveAnalytics

Influence programs http://service.sap.com/influence

Sign up to our newsletter http://scn.sap.com/docs/DOC-66912

http://service.sap.com/saproadmaps

http://scn.sap.com/

http://scn.sap.com/community/predictive-analytics

https://www.sap.com/trypredictive

http://sap.com/predictive

https://ideas.sap.com/

https://ideas.sap.com/PredictiveAnalytics

http://service.sap.com/influence

http://scn.sap.com/docs/DOC-66912

© SAP AG or an SAP affiliate company. All rights reserved.

Thank Youwww.sap.com/predictive

www.sap.com/scn-predictive

#sappredictive @sapanalytics

http://www.sap.com/predictive

http://www.sap.com/scn-predictive


© 2016 SAP AG or an SAP affiliate company. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or for any purpose without the express permission of SAP AG or an SAP affiliate company.

SAP and other SAP products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of SAP AG (or an SAP affiliate

company) in Germany and other countries. Please see http://global12.sap.com/corporate-en/legal/copyright/index.epx for additional trademark information and notices.

Some software products marketed by SAP AG and its distributors contain proprietary software components of other software vendors.

National product specifications may vary.

These materials are provided by SAP AG or an SAP affiliate company for informational purposes only, without representation or warranty of any kind, and SAP AG or its

affiliated companies shall not be liable for errors or omissions with respect to the materials. The only warranties for SAP AG or SAP affiliate company products and services

are those that are set forth in the express warranty statements accompanying such products and services, if any. Nothing herein should be construed as constituting an

additional warranty.

In particular, SAP AG or its affiliated companies have no obligation to pursue any course of business outlined in this document or any related presentation, or to develop or

release any functionality mentioned therein. This document, or any related presentation, and SAP AG’s or its affiliated companies’ strategy and possible future

developments, products, and/or platform directions and functionality are all subject to change and may be changed by SAP AG or its affiliated companies at any time for

any reason without notice. The information in this document is not a commitment, promise, or legal obligation to deliver any material, code, or functionality. All forward-

looking statements are subject to various risks and uncertainties that could cause actual results to differ materially from expectations. Readers are cautioned not to place

undue reliance on these forward-looking statements, which speak only as of their dates, and they should not be relied upon in making purchasing decisions.

http://global12.sap.com/corporate-en/legal/copyright/index.epx

Technology

#askSAP Analytics Innovations Community Call: Reimagine Analytics for the Digital Enterprise