36
Copyright © 2012, SAS Institute Inc. All rights reserved. WELCOME TO SAS FOR MARKETING 17 FEBRUARY 2016

SAS for Marketing - Analytics, Business Intelligence …€¢Malcolm Lightbody –Product Management, Customer Analytics •Developing all analytics (including attribution) capabilities

Embed Size (px)

Citation preview

Copyr i g ht © 2012, SAS Ins t i tu t e Inc . A l l r ights reser ve d .

WELCOME TO SAS FOR MARKETING

17 FEBRUARY 2016

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

AGENDA

Time Session

09.30 Registration & refreshments

10.00 Start

10.15 ADVANCED Attribution – What and How

Rule based attribution and algorithmic (advanced) attribution

Understand how SAS Enterprise Guide and SAS Enterprise Miner can be used

11.15 Break

11.45 Understanding Machine Learning in Marketing

What is Machine Learning?

Where can it be applied in Marketing?

How can it be applied in Marketing?

The next steps for your organisation.

12.45 Lunch

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED ATTRIBUTION – WHAT AND HOW

SAS CENTRE OF EXCELLENCE

Simon Waller and Iain BrownSAS

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

WHO ARE WE?

• Simon Waller – Global Practice Customer Intelligence

• Here 2 years ago talking about Attribution.

• Currently working on an attribution project for a customer in the UK.

• Dr Iain Brown – Analytics Centre Of Excellence

• Expert in analytics

• Understands the analytics behind advanced attribution

• Malcolm Lightbody – Product Management, Customer Analytics

• Developing all analytics (including attribution) capabilities in SAS CI

Solutions

• Insights into attribution project in USA

• Keen to hear from the field about attribution experiences and requirements!!!

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

AGENDA

• What it is

• 2 slide refresher on Attribution

• What is Advanced Marketing Attribution

• What you will need

• The environment required for Attribution

• Beyond Attribution

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

WHAT IS ADVANCED ATTRIBUTION

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ATTRIBUTION REFRESHER

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ATTRIBUTION REFRESHER - MODELS

Single Source

Attribution

First “Click”

Last “Click”

Fractional

Attribution (Rules

Based)

Fractional Attribution

(Algorithmic or

Probabilistic

Attribution)

Copyr i g ht © 2012, SAS Ins t i tu t e Inc . A l l r ights reser ve d .

ALGORITHMIC ATTRIBUTION

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

DIGITAL ATTRIBUTION SCENARIO

How can we measure

attribution across all

marketing interactions?

Find out which interactions have

highest conversion rate using an

algorithmic approach.

Business

Executive

Marketing /

Customer Service

CustomersBusiness Manager

Business Analyst

Business

Manager

(Marketing /

Customer

Service)

Analytics

Modeler

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

DIGITAL ATTRIBUTION SCENARIO WITH RPM

Business

Executive

Marketing /

Customer Service

CustomersBusiness Manager

Business Analyst

Business

Manager

(Marketing /

Customer

Service)

Analytics

Modeler

Business Analyst RPM

How can we measure

attribution across all

marketing interactions?

Find out which interactions have

highest conversion rate using an

algorithmic approach.

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

THE ANALYTICS LIFECYCLE

IDENTIFY /

FORMULATE

PROBLEM

DATA

PREPARATION

DATA

EXPLORATION

TRANSFORM

& SELECT

BUILD

MODEL

VALIDATE

MODEL

DEPLOY

MODEL

EVALUATE /

MONITOR

RESULTSDomain Expert

Makes Decisions

Evaluates Processes and ROI

BUSINESS

MANAGER

Model Validation

Model Deployment

Model Monitoring

Data Preparation

IT SYSTEMS /

MANAGEMENT

Data Exploration

Data Visualization

Report Creation

BUSINESS

ANALYST

Exploratory Analysis

Descriptive Segmentation

Predictive Modeling

DATA MINER /

STATISTICIAN

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

BUILDING ALGORITHMIC ATTRIBUTION MODELS

1. Data access – SAS/ACCESS, SAS Enterprise Guide / SAS Data

Integration Studio

2. Visual data exploration to access data quality issues – SAS Visual

Analytics

3. Reshaping the data for analytical modelling (recoding, transformations,

joins, summarisation and transpositions) – SAS Enterprise Guide / SAS

Data Integration Studio

4. Algorithmic model development – SAS Enterprise Miner (HPDM)

5. Model Scoring – SAS Enterprise Guide / SAS Data Integration Studio

6. Marketing channel performance dash boarding – SAS Visual Analytics

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

EXAMPLE

• 17 digital marketing channels

• 1,000 predictors

• ~24,000,000 digital visitor journeys

• Rate conversion less than 1 %

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

DATA-DRIVEN ATTRIBUTION

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

CHAMPION/CHALLENGER MODEL SELECTION

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

MODEL FLOW – APPLIED MACHINE LEARNING

• Model selection criteria:

• Misclassification on Validation set

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

LIFT CHART

• Identify group of responders more likely to convert compared to randomly

targetting entire marketable population

• Profile segment

• Find similar customers

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

MARKETING CHANNEL WEIGHTING – ODDS RATIO

• Attributes above line have a positive impact in increasing the probability of a

visitor conversion.

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

MODEL DEPLOYMENT

• Optimised score code generation (including data transformations)

• SAS code deployment through a scheduled job

• C or Java code deployment into applications (SAS Server/Client)

• PMML based scoring (SAS Server/Client)

• Scoring code and batch job built and executed in SAS Data Integration

• Model scoring task in SAS Enterprise Guide

• Scoring code can be scheduled for execution in parallel (grid computing)

• In-database functions (IBM DB2, Teradata, Netezza, EMC-Greenplum)

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

DEMONSTRATION

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

ADVANCED

MARKETING

ATTRIBUTION

ALGORITHMIC ATTRIBUTIONBENEFITS

• Leverage broad analytic data preparation options to build better attribution

models and get more reliable results

• Modern set of analytic algorithms to increase confidence in your results

• Reduce time to results with automated model scoring and deployment

• Consistent validation, management and monitoring of “best” models to

improve performance.

• Provides

• Increased accuracy

• More precise/flexible and controlled

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

WHAT YOU WILL NEED

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

WHAT YOU WILL NEED

1. Scalable (Onsite or Cloud)

• Big Data processing

• Data management

2. High quality multichannel data

• Accurate Customer Path Data

• Both offline and online data

3. Analysing the data

• Advanced Analytics Capabilities

• Opportunities outside attribution

4. Future Proof

• Do more as your expertise grows

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

1. SCALABLE

• From Desktop to Enterprise

• Big Data processing

• Data management

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

2. HIGH QUALITY MULTICHANNEL DATA

• Accurate Customer Path Data

• Data models open and attribution-ready

• Joining offline and online data

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

2. HIGH QUALITY MULTICHANNEL DATA

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

2. HIGH QUALITY MULTICHANNEL DATA

TV

Social

Search

1

2

3

4

5

6

Mailing£

£

£

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

2. HIGH QUALITY MULTICHANNEL DATA

Joining offline and online data

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

3. ANALYSING THE DATA

Advanced Analytics

Advanced Analytics

Actioning the data

Data Requirements

Flexible reporting

Using the data to feed opportunities outside

attribution

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

OPPORTUNITIES BEYOND ATTRIBUTION

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

OPPORTUNITIES OUTSIDE ATTRIBUTION

1

2

3

4

5

6

£

£

£

Prediction Window

£

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .

SAS Customer Decision Hub

Ru

les

CUSTOMER DECISION HUB

in Real-Time

across all channels

based on pre-defined rules

deciding about the optimized offer.

Central Decision Making unit:

4. FUTURE PROOF THE SAS® CUSTOMER DECISION HUB

http://www.sas.com/en_us/insights/articles/marketing/customer-decision-hub.html

Copyr i g ht © SAS Ins t i tu t e Inc . A l l r ights reser ved .sas.com

Thank you very much!