Banyan - Sutra Project Workshop 4 Materials

Banyan - Sutra Project Workshop 4 Session

Objectives

Cross Sell and Upsell and Increase Lifetime value of customer

Question that needs to be answered: What campaign should i run to achieve above

Business lines in scope:

  • CM (Hotel)
  • MYK (Hotel)
  • PB (Hotel)
  • Banyan Tree Essentials (SPA Retail)

Marketing Campaign

Themes:

  1. Visit New Property
  2. Visit Another Property
  3. Come Back to Hotel One More Time
  4. Visit Shopify
  5. Come back to Shopify One More Time
  6. Upsell Purchases

Data Ingestion, Cleaning & Prep

  • Data Ingestion from Oracle On Premise DB, Oracle Cloud, Shopify and Google Analytics [Tool: VAL API Integration, VAL Database Integration]
  • Deduplicate Customer List as Universal Customer List [Tool: VAL Dedup]
  • Data Prepping [Tool: VAL Function Fields, VAL SQL]

Visit Shopify

Objective: Increase engagement with BT customers through retail outside their stay at BT

Target Audience: BT customers who have not spent at Shopify

Campaign ROI

Step 1: Determine overlap of customers across engagement points

Step 2: Determine customer spent of overlap engagement

Step 3: Calculate potential ROI [Tool: VAL Rules Mining & ROI]

image
Source
Target
# Potential Customers
Target Avg LT Spent
Confidence
Additional Revenue
Max Potential Revenue
Return on Investment
Target Effectiveness
support
support_confidence
['CM']
['MYK', 'PB']
14473
3,287
0.02%
9,990
47,572,860
9
1
0.01%
0.00%
['MYK', 'PB']
['CM']
34413
1,523
30.00%
15,721,515
52,405,049
15,721
1
0.01%
0.00%
['PB']
['MYK', 'CM']
169
2,737
1.78%
8,211
462,585
7
1
0.01%
0.00%
['MYK', 'CM']
['PB']
47237
875
0.20%
83,053
41,320,015
82
1
0.01%
0.00%
['CM', 'PB']
['MYK']
14632
3,296
30.00%
14,467,887
48,226,292
14,467
1
0.01%
0.00%
['MYK']
['CM', 'PB']
34254
1,518
0.01%
4,679
51,985,559
4
1
0.01%
0.00%
['MYK']
['SHOP', 'CM']
34254
1,069
0.01%
3,296
36,623,663
2
1
0.01%
0.00%
['SHOP', 'CM']
['MYK']
20304
3,296
20.00%
13,384,180
66,920,901
13,383
1
0.01%
0.00%
['PB']
['CM']
169
1,523
5.92%
15,228
257,358
14
1
0.02%
0.00%
['CM']
['PB']
14473
875
0.07%
8,735
12,660,088
8
1
0.02%
0.00%
['CM']
['MYK']
14473
3,296
10.30%
4,910,945
47,702,236
4,910
1
2.81%
0.29%
['MYK']
['CM']
34254
1,523
4.35%
2,269,087
52,162,920
2,268
1
2.81%
0.12%
['PB']
['MYK']
169
3,296
5.92%
32,959
557,015
32
1
0.02%
0.00%
['MYK']
['PB']
34254
875
0.03%
8,689
29,963,287
8
1
0.02%
0.00%
['MYK', 'SHOP']
['CM']
39968
1,523
2.27%
1,383,447
60,864,354
1,382
1
0.01%
0.00%
['CM']
['MYK', 'SHOP']
14473
2,726
0.02%
8,285
39,454,166
7
1
0.01%
0.00%
['SHOP']
['MYK']
5846
3,296
2.26%
435,074
19,268,104
434
1
0.25%
0.01%
['MYK']
['SHOP']
34254
112
0.39%
14,715
3,822,153
14
1
0.25%
0.00%
['MYK', 'CM']
['SHOP']
47237
112
0.20%
10,594
5,270,830
10
1
0.01%
0.00%
['SHOP']
['MYK', 'CM']
5846
2,737
0.05%
8,161
16,001,616
7
1
0.01%
0.00%
['CM']
['SHOP']
14473
112
0.10%
1,680
1,614,936
1
1
0.03%
0.00%
['SHOP']
['CM']
5846
1,523
0.26%
22,879
8,902,447
22
1
0.03%
0.00%

Dashboard - 00_Segments (thinkval.io)

Campaign Flow

image

Campaign Organisation

image

Dataset

Customer Filter

  • Customer last engagement across Shopify and Hotel post 1 Jan 2017
  • Remove Purged customer
  • Remove Negative spend
  • Remove Banyan Staff using banyan.com email

Transaction Filter

  • Remove Cancelled Reservation
  • Transaction till 31 Dec 2022 for all hotels and Shopify
  • Remove Pseudo Transaction identified by room

Customer Segments

Step 1: Identify spent behaviour of customer at BT hotel that goes to Shopify

Purpose

  • Used to identify spent profile of customer who have spent at Shopify
  • Used to identify similar spent profile as above who have not spent at Shopify
  • For customers who have not spent time at BT hotel, use this to identify what they have engaged with and use images or text to personalise content

Steps

  • Identify hotel spent cluster [Tools - VAL Cluster]
  • Identify shopify spent cluster [Tools - VAL Dashboard, VAL Function Fields or VAL Cluster]
    • Show printscreen of histogram of spent in VAL + Rules Fields Config
  • Identify RFM of hotel [Tools - RFM Analysis]
  • Identify RFM of shopify [Tools - VAL RFM]

Findings: [Tools - VAL Dashboard]

  • Population of customers in different clusters split by different dimensions (Hotel Spent)
  • Population of customers in different clusters split by different dimensions (Shopify Spent)
  • Population of customers in different clusters split by different dimensions (Hotel and Shopify Spent)
  • Behaviour of existing customer spent in Hotel whom have already spent in Shopify

      • Dashboard - Shopify Parallel plots (thinkval.io)

      image
    1. People who spent in Shopify don’t spend a lot in hotel.
    2. People who spent in Shopify don’t stay much in hotel

List of customers with all attributes combined with analytics data from above

Workspace - Person Spend (thinkval.io)

image

Step 2: Calculate recency of visit to existing establishments and which establishment

Purpose

  • To identify customers that have fresher memory of BT hotel experience (0 - 3 years)
  • Use images of most recent visit to increase familiarity to experience

Steps:

  • Filter base on the dimensions and metrics available in universal customer list [Tools - VAL Workspace]

Step 3: Identify customers with children or no children

Purpose

  • Use different images or wordings to increase personalisation of content

Steps:

  • Filter base on the dimensions and metrics available in universal customer list [Tools - VAL Workspace]

image

Step 4: Identify what existing customer of BT hotel that goes to Shopify purchase

Purpose

  • Use to identify bundle to create for sales or image of product to use to increase chances of sales

Steps

  • Perform product basket analysis to identify most bought product combination [Tools - VAL Basket Association]
Based on customer who have visited Shopify and Hotel
Based on customer who have visited Shopify and Hotel
Based on customer who have visited Shopify in America
Based on customer who have visited Shopify in America

Step 5: Identify what time customers buy on Shopify and which country they are from

Purpose

  • Use to identify timing for marketing email to be sent out for each customer

Steps

  • Identify timing for conversion on Shopify [Tools - VAL Dashboard]
image
image

Step 6: Validate hypothesis and identify other potential relationship

Purpose

  • Get different perspective of relationship to identify unidentified ones or validate existing one

Steps

  • Perform correlation analysis of all customer attributes (demo, spent, stay, rfm, etc) [Tools - VAL Correlations]

Findings

  • Day Count from Last Shopify Spent is negatively related to Day Count Last Resort Visit
image
❗
- Visit Shopify - What did you last do at BT hotel - Family or Non-family message - Time to send to you so that you will purchase

Campaign Data Driven Components Idea

image
image

Campaign Ideas

image
image
image
image