Sep 30, 2016 - Get insight from data variables and apply to CRM ... business growth since 2010/11 ... And to further max
The 9th RBForum: - HKJC’s Loyalty Management
Joice Lam Executive Manager, Customer Management 30 Sept 2016
Customer and Marketing Division
Reward / Incentive
Target
Loyalty 4 million customers
Privileges / Service
Cost
2
Customer and Marketing Division
Principles for HKJC’s loyalty design and management 1. Use of customer data (big data modelling) • To better serve the customer needs Properly segment across tiers • To create a flexible structure so we can segment based on loyalty factors Cater to customers’ desire for choices • To create a gift redemption platform offering a broad set of rewards • Meanwhile, we know our customer more through their choices.
3
Customer and Marketing Division
Get insight from data variables and apply to CRM • 400 variables are collected from different sources
Betting History
Gift Reward Choice
Betting Pattern
Customer Survey
Personal Particular
Customer Preference
Personal Information
Customer Feedback
4
Customer and Marketing Division
…and identify program components which is effective CRM Program Objective
Selection metrics
Avg. Response Rate
Wake up
•
Customer who becomes inactive after the summer break
~30%
Reduce lapse / sustain active rate
•
Customer who participate less recently
~40%
Go Racing Invitation (for interest)
•
Regular race-goers
~20%
Go Racing (for retention)
•
Customer who consistently participate in racing with similar involvement
~20%
Go Racing (for reactivation)
•
Customer who has no racing participation over 1 season
~5%
Absenteeism
•
Customer with predictable absence in coming meetings
~5%
This is one of the drivers to sustain business growth since 2010/11 5
Customer and Marketing Division
Only big data modelling can handle trillion of data set
Customer Preference
Personal Information
Betting Pattern
1.6 trillion of data
6
Customer and Marketing Division
And to further maximize our capability in data analytics
… to proactive • Customer database From reactive… • • Customer database • Observed behavior (e.g., switching to competitors)
Predictive behavior (e.g., will switch) via mathematical models
• Pre-emptive CRM actions before churn
• CRM actions
7
Customer and Marketing Division
We automate the process of analytics with predictive ability via machine learning models 1 Build customer database
Database 2 million+ customers
2 Identify relevant variables
3 Develop predictive models on customers
400 data variables E.g., Wagering behavior Personal information Social activities Race information Weather information
8
Customer and Marketing Division
Not only improves accuracy in targeting but predict the incremental business offered by loyalty Models
Description
1.
Lapse (or Decline) Model
• Predict customers likelihood of dropping / stopping in next race meeting • Reduce lapse using pre-emptive campaigns
2.
New Customer Potential
• Predict if an individual new customer can convert into a regular customer
3.
Customer Trends
• Develop customers segmentation using unsupervised clustering • Plan customer programs by refined segmentation
4.
Bet-type Adoption
• Predict an individual customers’ bet-type active rate
9
Customer and Marketing Division
With big data models, targeting would be much easier with a few clicks on desktop PC
3
1
2
3
4
1
2
10
Customer and Marketing Division
With big data modeling, pre-emptive actions can be made before loyalty declines Decline model:
Growth model:
High chance to decline
High chance to grow 11
Customer and Marketing Division
Different validation results showed that our targeting accuracy has improved by 20-40% 90 Segment Conventional Method 80
+ 22%
+ 37%
+ 38%
70
Big Data Model
+ 17%
+ 20%
60 50 40 30 20 10 0
Chance to be Regular (New Customer Model)
Decline Results (Growth/Decline Model)
Growth Results (Growth/Decline Model)
Participation Potential (Bet Type Model)
Preference Potential (Bet Type Model) 12
Customer and Marketing Division
We will continue to invest in the big data domain to further improve CRM Textual mining Campaign effectiveness modeling Integration into CRM system
•
Integrate customer potential scores into CRM system to improve day-to-day experience
•
Predict how to optimize response rate of a campaign
•
Utilize campaign resources and automate some marketing decisions potentially
•
Mine customer voices from various touch points (e.g., Retail, CS hotline, social media etc.) via sentimental analysis modeling
•
Append insights data to existing customer profile or customer intelligence (e.g., preference for gifts, other hobbies/ lifestyles)
13
Customer and Marketing Division
Principles for HKJC’s loyalty design and management 1. Use of customer data (big data modelling) • To better serve the customer needs 2. Properly segment across tiers • To create a flexible structure so we can segment based on loyalty factors Cater to customers’ desire for choices • To create a gift redemption platform offering a broad set of rewards • Meanwhile, we know our customer more through their choices.
14
Customer and Marketing Division
We created a 4-tier flexible structure to segment based on customers’ spending and interest • Similar to all existing loyalty programs, easy to understand
Betting = spending
Racing = interest
•
Frequency of using wagering account
•
Frequency of visit to racecourse
•
Sum of betting using wagering account
•
Frequency of purchase at gift shop
•
The length of time as a betting account customer
(Upcoming factors to be rewarded) •
Participation of Marketing game
•
Shared or Liked a Facebook post
15
Customer and Marketing Division
The point based loyalty system also assist customers to set out upgrade goal For leisure: - Racing familiarization program For racing fans: - Racecourse invitation For lifestyle customer: - Happy Wednesday event
For HKJC Priority customer: - Priority access to betting channel - Exclusive content - Exclusive event - Reward
Black
(For future) New tier
Gold Silver 120,000 points
Red Mass customer
30,000 points 10,000 points 6,000 points
16
Customer and Marketing Division
Principles for HKJC’s loyalty design and management 1. Use of customer data (big data modelling) • To better serve the customer needs 2. Properly segment across tiers • To create a flexible structure so we can segment based on loyalty factors 3. Cater to customers’ desire for choices • To create a gift redemption platform offering a broad set of rewards • To know our customers further from their choice of rewards.
17
Customer and Marketing Division
We provide flexibility in customers’ choices of reward
6
12
22
42
PRIORITY Counter Priority Counter at OCBBs & Racecourse
PRIORITY Access Telebet Operator Services Priority Customer Care
PRIORITY Racecourse Privileges Priority Venues Booking Access to Priority Card VIP Venues
PRIORITY Offers Priority Birthday Treats Priority Festival Gifts
Priority Reservation Services for Overseas Races 18
Customer and Marketing Division
Tailored racing experience is made for racing fans, while exclusive betting services for betting customers. Racing Fans prefer: • Racing event with Jockey / Trainer • Oversea tour to witness International races • Racecourse invitation • Gift with racing elements
Betting Customers prefer: • Priority betting services • Exclusive racing content • Free racing newspaper • Gift with high value
Annual Champion Awards
19
Customer and Marketing Division
The cardholder base is now 2.4 times the size in 2011
No. of P-card customers
S1112
S1213
S1314
S1415
S1516
The customer base is grown by 2.4 times from 2011 to 2016.
S1617
S1718 est
20
Customer and Marketing Division
There has been a continual growth in terms of Racing turnover (even under the current economy downturn) Total racing turnover (in HK$ billion) 120
100
105
107
2013/14
2014/15
2015/16
94
80 64 60
104
67
68
2007/08
2008/09
82
84
2010/11
2011/12
72
59
40
20
0 2005/06
2006/07
2009/10
2012/13
Economy recession 21
Customer and Marketing Division
The future development of HKJC’s Loyalty program Mobility!
22
Customer and Marketing Division
The world is going mobile, loyalty has to go mobile • We believe the loyalty card is on phone, customers will engage deeper. • Customers expect rewards to be delivered more real time so that they can use it more frequently. • We can maintain a closer connection with customers (eg. survey, voting for best Jockeys/Horses). • Mobile also enables us to open up the platform for future expansion, (e.g expose to social media)
23
Customer and Marketing Division
HKJC’s loyalty management process is continual and aiming at creating long term relationship
Use of customer data
Customer Relationship Cater to customer’s desire for choices
Segment across tiers
24
Customer and Marketing Division
Thank you
25