@KennethLim
NOT
ALL
DATA
IS
CREATED
EQUAL
If Big Data was useful,
we would have called it
Useful Data
Donald Rumsfeld:
“ There are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns -- the ones we don't know we don't know.
” 7
Finding Data
known knowns
known unknowns
unknown unknowns
8
Data Application
Analysis & Testing
Discovery
We analyze data…
not because
we want to report it
not because
we can
and certainly not because
#YOLO
but because we want to
improve the outcomes
We must learn
how our goals are impacted
by understanding the
relationships within the data
Not all data
is created equal
Goal
Performance
Process
Behavior Circumstances
Data Hierarchy
16
Circumstances: variables that can influence Behavior, e.g. Seasonality
Performance: what you need to reach your goal, e.g. € 2M Profit
Process: a key figure that impacts Performance, e.g. Profit Margin per Product
Behavior: individual actions within the Process, e.g. Products Bought and Price Paid
Goal
Data Hierarchy
17
: what you ultimate want to achieve, e.g. 10% Financial Growth
Online Shop Example
Known Knowns Known Unknowns
Goal Profit
Performance Revenue Costs
Process Revenue per Customer Revenue per Order Revenue per Email Campaign
Behavior Website Visits Products Bought Orders Made Amount Paid Discounts Applied Abandons Emails Opened Email Links Clicked
Customer Online Times
Circumstances Holidays Gifts Email Campaigns
Birthdays Anniversaries
18
Circumstances
Understanding Relationships within Data
19
Goal
Performance
Process
Behavior
Profit
Revenue
Emails Opened
Email Links Clicked
Email Campaign
Amount Paid
Revenue per Email Campaign
Discounts Applied
Abandons Customer
Online Times
Email Campaign Process
20
1. Email Received
9. Email Order
Revenue (in €)
4. Order
Placed?
5. Abandoned?
7. Discount Applied?
8. Email Order
Discount (in €)
6. Email Order
Abandon (in €)
2. Email Opened
3. Email Link Clicked
Yes Yes
Yes
No No
Measuring Revenue per Email Campaign
21
Total Email Order Revenue
Total Number of Emails Sent
Total Unique Email Opens
Total Number of Emails Sent
Total Email Order Revenue
Total Email Order Revenue + Total Email Order Discounts + Total Email Order Abandons
*
Revenue per Email Campaign =
Adjusted Revenue per Email Campaign =
But wait…
there’s more!
Circumstances
Improving Revenue per Email Campaign
23
Goal
Performance
Process
Behavior
Profit
Revenue
Emails Opened
Email Links Clicked
Email Campaign
Amount Paid
Revenue per Email Campaign
Discounts Applied
Abandons Customer
Online Times
We can obtain data
by asking
We can obtain data
by taking
We can obtain data
by testing
We should obtain
data by asking,
taking & testing
Improving Revenue per Email Campaign
28
Total Unique Email Opens
Total Number of Emails Sent
Total Email Order Revenue
Total Email Order Revenue + Total Email Order Discounts + Total Email Order Abandons
*
Adjusted Revenue per Email Campaign =
An Evolving Approach to Data
1. Understand the impact of and the relationships within the data
2. Collect the data that is important
3. Analyze the outcomes
4. Optimize the approach
29
The story is never
about the data itself
Final Thoughts
• Always look to improve the outcome
• Establish a firm understanding of the relationships within your data
• Challenge the unknown
• The story is never about the data itself
31