master data science for marketing...customer insights are the gold buried within your data. 4/3/2018...
TRANSCRIPT
4/3/2018
1
Consumer Marketing 2018
Brandon Purcell
Master Data Science For Marketing
2© 2017 FORRESTER. REPRODUCTION PROHIBITED.
Customer insights are the gold buried
within your data
4/3/2018
2
3© 2017 FORRESTER. REPRODUCTION PROHIBITED.
“We are drowning in data and starving for insight.”
— Global Bank
4© 2017 Forrester Research, Inc. Reproduction Prohibited
Key phases of the insights lifecycle
Insights
Action
Data
4/3/2018
3
5© 2017 Forrester Research, Inc. Reproduction Prohibited
Key phases of the insights lifecycle
Insights
Action
Data
Data-to-insights gap
Insights-to-action gapContinuous
improvement gap
6© 2017 Forrester Research, Inc. Reproduction Prohibited
The data-to-insights gap is most prevalent
24%
26%
28%
40%
43%
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%
Hiring talent to manage measurement and analytics
Providing real-time insights to the business
Getting buy-in from business stakeholders on thevalue of measurement and analytics
Accessing data from a variety of sources
Ensuring data quality from a variety of sources
Please rank the top three challenges that prevent your organization from making use of measurement and analytics.
Data-to-insights gap
Insights-to-action gap
4/3/2018
4
7© 2017 Forrester Research, Inc. Reproduction Prohibited
Customer analytic techniques bridge the data-to-insights gap
8© 2017 FORRESTER. REPRODUCTION PROHIBITED.
Now let’s explore the menu
4/3/2018
5
9© 2017 Forrester Research, Inc. Reproduction Prohibited
Methods that drive acquisition…
Source: February 2014 “TechRadar™: Customer Analytics Methods, Q1 2014”
Acquisition analytics
methods:
• Behavioral customer
segmentation
• Customer lifetime
value analysis
• Customer lookalike
targeting
10© 2017 Forrester Research, Inc. Reproduction Prohibited
Methods that increase retention and loyalty…
Source: February 2014 “TechRadar™: Customer Analytics Methods, Q1 2014”
Retention and loyalty
analytics methods:
• Customer propensity
analysis
• Churn and attrition
analysis
• Social network
analysis
4/3/2018
6
11© 2017 Forrester Research, Inc. Reproduction Prohibited
Methods that drive personalization…
Source: February 2014 “TechRadar™: Customer Analytics Methods, Q1 2014”
Personalization
analytics methods:
• Next best action
• Recommendation
analysis
• Cross-sell and upsell
analysis
12© 2017 Forrester Research, Inc. Reproduction Prohibited
Identify dependencies between methods
Source: February 2014 “TechRadar™: Customer Analytics Methods, Q1 2014”
4/3/2018
7
13© 2017 Forrester Research, Inc. Reproduction Prohibited
Some methods enjoy more success than others
14© 2017 Forrester Research, Inc. Reproduction Prohibited
Most customer analytics techniques rely on
machine learning
4/3/2018
8
© 2017 Forrester Research, Inc. Reproduction Prohibited 15
Supervised vs unsupervised machine learning
Supervised learning Unsupervised learning
Purpose To predict / classify To explore / understand
Training data Labelled (knows the
“answer”)
Not labelled (no “right
answer”)
Accuracy Measurable Qualitatively evaluated
Use cases for
marketing
Predict which customers
are likely to respond /
churn / buy
Behavioral customer
segmentation
© 2017 Forrester Research, Inc. Reproduction Prohibited 16
An algorithm detects patterns in data to create a model
4/3/2018
9
17© 2017 FORRESTER. REPRODUCTION PROHIBITED.
Supervised learning – Will I play tennis?
Day Outlook Temp. Humidity Wind Play Tennis?
D1 Sunny Hot High Weak No
D2 Sunny Hot High Strong No
D3 Overcast Hot High Weak Yes
D4 Rain Mild High Weak Yes
D5 Rain Cool Normal Weak Yes
D6 Rain Cool Normal Strong No
D7 Overcast Cool Normal Weak Yes
D8 Sunny Mild High Weak No
D9 Sunny Cold Normal Weak Yes
D10 Rain Mild Normal Strong Yes
D11 Sunny Mild Normal Strong Yes
D12 Overcast Mild High Strong Yes
D13 Overcast Hot Normal Weak Yes
D14 Rain Mild High Strong No
18© 2017 FORRESTER. REPRODUCTION PROHIBITED.
Decision tree for “Play tennis?”
Outlook
Sunny Overcast Rain
Humidity
High Normal
Wind
Strong Weak
No Yes
Yes
YesNo
4/3/2018
10
19© 2017 Forrester Research, Inc. Reproduction Prohibited
Data type Definition Examples
Nu
meri
cal Continuous All numerical values, including
fractions
Purchase amount
Discrete Only integer values Website visits
Cate
gori
cal
Nominal Qualitative data without order Region
Ordinal Ordered qualitative data Level of education
Binary Qualitative data with 2 values Offer click
Time and date Time and date data Timestamp
String Text data Email content
From an algorithmic perspective, there are only a few types of data
© 2017 Forrester Research, Inc. Reproduction Prohibited 20
Meet the algorithms!
Use case:
Segmentation
Use case:
Product
recommendation
Use case:
Forecasting
Use case:
Churn / propensity
analysis
4/3/2018
11
21© 2017 FORRESTER. REPRODUCTION PROHIBITED.
Don’t be confused by the confusion matrix
22© 2017 FORRESTER. REPRODUCTION PROHIBITED.
A cumulative gains curve shows a model’s expected lift
4/3/2018
12
23© 2017 Forrester Research, Inc. Reproduction Prohibited
Machine learning results in new outputs requiring different operational processes
24© 2017 Forrester Research, Inc. Reproduction Prohibited
➢ What is the business objective of the project?
➢ Who is the project owner? Who are the relevant
stakeholders?
➢ Are there risks / constraints we need to take into
account?
➢ What would an ideal solution look like in action?
➢ How will we measure the success of this project?
Data science starts with answering key business questions
4/3/2018
13
Thank youBrandon Purcell
+1 510.926.2694
26© 2018 Forrester Research, Inc. Reproduction Prohibited
Meet the Panel
Richard P. WatsonDivision Vice President Client Experience, ADP
Newcombe ClarkGlobal Director, Rapid Learning, AIG