Sunteți pe pagina 1din 13

Agenda

1 Background and Objective

2 Week 1 of project

3 Proposed Approach

4 Challenges

5 Detailed Project Plan

© Absolutdata 2014 Proprietary and Confidential 1


Background and Objective
 Casino is one of the more profitable business of the cruise lines. However, there is no systematic
and automated way of forecasting the casino spends in any voyage. This leads to sub optimal
decisions that potentially leads in lost opportunity for Casino business

 Casino business aims to understand the key drivers and dynamics of casino penetration and
spend with a scientific & data driven approach

 Additionally, attempt is made towards making the business planning process more robust using
the right blend of statistically proven drivers and business judgment

 Finally, aim is to built a user friendly planning tool that takes gives management scientifically
driven predictions and levers to take right action at right time

© Absolutdata 2014 Proprietary and Confidential 2


Factors Influencing Customer Casino Penetration and Casino Spend

Factors Sample Variables

% Germans

Customer
profile % Honeymooners

Port of call(s)
Itinerary
details No. of days at sea
What impacts
Casino Spend
Christmas time period
Seasonality
Summer vacation

% of land based
Marketing acquisitions customers
Levers
VIP sailings

© Absolutdata 2014 Proprietary and Confidential 3


Proposed 3 Step Approach
Exploring Data in multiple aspects & in multitude of dimensions:
Exploratory Data
Analysis (EDA) &
 Itinerary Details  Casino Specification  Other Services (Spa)
Driver Study
 Marketing Effectiveness  Guest and Family  Ship Specific

Casino Penetration and Spend specific predication capabilities at a


Drivers
Driven Model voyage level with the help of two type of models:
i. Casino Penetration Models
ii. Casino Spend Models

Taking the planning process to a next level with:


Planning
Tool i. Scientifically driven prediction of penetration and spend
ii. Using past data and business expectations
iii. In the form of a user friendly tool across portfolios

© Absolutdata 2014 Proprietary and Confidential 4


Data Sources
 3 years of past data would be required for analysis from the following possible sources – Casino Data, CRM Data
Warehouse and Marketing Data across the 5 Brands in question – Carnival, Princess, P & O AU, Costa, Holland
America

Casino Data CRM Data Warehouse Marketing Data

 Casino spend data  Customer details  Acquisition Efforts

– Slots spend, table spend – Gender, age, martial status, city etc – Redemption of certificates
 House hold details from land based casino
– Coin in and Coin out partnerships
– No. of people, income etc
– Total casino visits  On Board Details  Retention Efforts

– Voyage currency – Number of game tables, number of slot – Number of customers utilizing
machines etc. direct mail offers
 Customer profile
 Booking and sail – VIP sailings
– Loyalty membership, loyalty
points earned – Price paid, number of travelers
– Type of cabin
– Other demographics
– Itinerary details
– Port of embarkation/ D-embarkation
– Number of days at sea
– Ship details
– Age
– Casino employees
© Absolutdata 2014 Proprietary and Confidential 5
Step 1 – Exploratory analysis
 Carnival Casino team to get back around the key areas/ question they would like to be probed
and which will stress more in the EDA by 11th July 2014

Are free cabin offers


better than price
How does time on promotions
sea impact spend ?
Do additional slot
machines lead to similar
rise in spending

Exploratory Data
Are guest of same Analysis &
age group spend
similarly Driver Study
irrespective to
ethnicity Are customers coming
through land based casino
partners better?

Should groups with all “boys” Can off seasons be


need to given more incentives better managed

* Illustrative Examples; Subject to relevant variables being available


© Absolutdata 2014 Proprietary and Confidential 6
Step 2 – Driver Models
Carnival Guests Casino Guests

Penetration Model

Casino Guests Casino Spenders

Spend Model

i. Models will be build at a Voyage level

ii. These models will utilize the drivers and variables studied in EDA and weighing each
of the drivers in sync with past behaviour

© Absolutdata 2014 Proprietary and Confidential 7


Step 2 – Driver Model Details (Contd.)

Data Dependent Independent


Type Model Technique
Required Variable(s) Variable (s)

• Itinerary • Number of sea


• Ship/ itinerary details days
details • Type of ship • Overall guest
Voyage Level - Percentage of • Ship brand characteristics • Regression(Hierar
• Brand level
Casino Separate people • Seasonality (demographic chical Bayesian
transactional
estimates at spending in • Services profile) in the regression)
Penetration Brand/ type of
data
casino during • Onboard voyage • Fixed and random
• Data activity
ship level that Voyage effects model
warehouse • Sales data
data • Ports covered

• Itinerary • Number of sea


• Total Spend in details days
• Ship/ itinerary
Casino, • Type of ship • Overall guest
details • Regression
Voyage Level - • Casino spend/ • Ship brand characteristics
• Brand level (Hierarchical
Casino Separate day/ voyage, • Seasonality (demographic
transactional Bayesian
estimates at • Percentage • Services profile) in the
Spends Brand/ type of
data
casino spend • Onboard voyage
regression)
• Data activity • Fixed and random
ship level as overall
warehouse • Sales data effects model
voyage
data • Ports covered
revenue

© Absolutdata 2014 Proprietary and Confidential 8


Step 3 - Planning Tool

1. The above shown tool (draft structure) will enable the user to achieve estimated penetration
and sales based on driver driven model;
2. The user will be able to provide input around key drivers and examine the forecast with the help
of a comprehensive MIS

© Absolutdata 2014 Proprietary and Confidential 9


1 Background and Objective

2 Week 1 of project

3 Proposed Approach

4 Challenges

5 Detailed Project Plan

© Absolutdata 2014 Proprietary and Confidential 10


Challenges

 Central to EDA and modelling is availability of data – both casino specific as well as brands
covering demographic, spend, solicitations, voyage and ship specific information.

 Second key piece is to synchronize and standardize the data at guest and voyage level across time
and across businesses. For example: Identifying the same guest across data casino and brands
data

 The structure of data and the process with which the information is captured under each variable
might differ across portfolios

 Brand / Portfolio level difference pertaining to the business itself. For example – different
currencies used by different businesses and exchange rate conversions may not reflect the true
cost differences for the guests

© Absolutdata 2014 Proprietary and Confidential 11


1 Background and Objective

2 Week 1 of project

3 Proposed Approach

4 Challenges

5 Detailed Project Plan

© Absolutdata 2014 Proprietary and Confidential 12


Project Plan
Segmentation Activity 1st Month 2nd Month 3rd Month
Steps 07/07 07/14 07/21 07/28 08/04 08/11 08/18 08/25 08/31 09/07 09/14 09/21
Initial hypotheses & Final Sample
1
Plan
2 Data Setup
Data pulling, structuring and
3
Cleaning
4 Exploratory Data Analysis
Milestone 1
5 Casino Penetration modelling
Milestone 2
6 Casino Spend Modelling
Milestone 3
7 Optimization Tool

8 Workshops and Presentations

Plan Data Setup Data structuring and cleaning Exploratory Data Analysis

Milestone 3 Milestone 2
Milestone 1
Optimization Tool Penetration Modeling
Presentations Casino Spend Modeling

© Absolutdata 2014 Proprietary and Confidential 13

S-ar putea să vă placă și