2. 4 P’S OF MARKETING
Goods
Services
Online content
Email
SMS
Mobile App
Social media …
Email campaigns
Facebook ads
Google display ads
adverts
Discounts
Additional promotion
Additional features
Free premium period
Product
Promotion
Place
Price
3. INTENT OF MARKETING
Acquire new customers
Retain and increase value ( to and from ) customers
Bring back Churned customers
6. MARKET RESEARCH – SURVEY
FOOD DELIVERY SERVICE
Person Free Delivery Time to Deliver Tasty options Discounts
1Not important Important Imporant Very Important
2Very Important Not important Very Important Not at all important
3Important Imporant Imporant Imporant
4Very Important Imporant Imporant Very Important
5Not at all important Very Important Very Important Not at all important
Attribute Level Part-worth Utility Attribute Utility Range Attribute Importance
Free Delivery
Important 30
Very Important 60 40 27%
Not Important 20
Not at all Important 10
Time to
Deliver
Important 90
Very Important 50 90 33%
Not Important 20
Not at all Important 0
Tasty
Options
Important 20
Very Important 10 20 10%
Not Important 5
Not at all Important 0
Discounts
Important 30
Very Important 20 50 30%
Not Important 20
Not at all Important 10
Customer Survey Conjoint Analysis
33%
30%
10%
27%
Chart Title
Time to Deliver
Discount
Tasty Options
Free Delivery
7. MARKET RESEARCH
KNOW YOUR PRODUCT, CUSTOMER AND MARKET
Region Sales Packaging Quality Price Range
1 23478 2 1
2 23489 1 4
3 54525 3 2
4 45778 4 3
5 23456 5 4
Does the price range and the packaging quality influence the sales
One Way ANOVA – Statistical Significance
of influence
Significant different – Concentrate on
packaging
9. CROSS-SELL/UPSELL
Add to Cart
Page/Abandon
Cart campaigns
Upsell Apriori( Market Basket )
AlgorithmIntentWherePeople who bought
also bought
Product Detail
Page/Post
Purchase
campaigns
Cross
Sell
Collaborative Filtering
Singular Value Decomposition
Product Detail
Page/Post
Purchase
campaigns
Cross
Sell
Product-Customer
Collaborative Filtering
Singular Value Decomposition
10. COLLABORATIVE FILTERING
Customer 1 2 3 4
a 0.2 0.4 0.7 0.9
b 0.9 0.3 0.4 0.8
c 0.7 0.2 0.6 0.1
d 0.2 0 1 0.2
Product 1 2 3 4
a 0.2 0.4 0.7 0.9
b 0.9 0.3 0.4 0.8
c 0.7 0.2 0.6 0.1
d 0.2 0 1 0.2
• Matches the Implicit (Clicks and purchases ) and Explicit rating of customers
• Marches the Attribute values of products
• Identify similarity of customers and Products
• Recommend similar products or products rated high by similar customers
11. SINGULAR VALUE DECOMPOSITION
• Recommendation based on similarity of customers/products based on
Implicit ratings like clicks, purchases
• Matrix Factorization: Decomposes customer-product implicit/explicit rating
matrix into 3 matrices
• Identifies latent features of Customers and products
• Compute similarity of customers and products and recommend
Customer Prod1 Prod2 Prod3 Prod4
a 1 0 0 1
b 0 1 1 1
c 1 0 0 0
d 0 0 0 1
Customer-Product
Implicit Rating
Customer Matrix Product Matrix
12. APRIORI
• Identifies frequent products in the same order ( basket )
• Association Rule mining
• Uncovers association between items
• Support – Number of times item A and B are bought together
• Confidence – Number of transactions with Item A
13. ENGAGEMENT RFM/PURCHASE RFM
▪ Not all customers are same
▪ Different Strategy for different segment of customers
▪ Purchase RFM – RFM Score computed from purchases
▪ Engagement RFM – RFM Score computed from clickstreams
▪ Recency – A score that indicates retention
▪ Frequency – A score that indicates Engagement
▪ Monetary – A Score that indicates the Life time value
Win back
Campaign
Welcome
seriesDon’t worry now
Sneak peak
campaigns
14. WE KNOW WHAT YOU DID LAST SUMMER ☺
• Java tag based cookie tracking for customer acquisition and customer tracking
• No. 1 in digital customer acquisition
• Surpassed AdWords and search digital spend for acquisition
• Personalize customer experience with click history
• Consistent content and messages across channels and devices
15. CHURN PREDICTION
▪ Churn Analytics
Customer acquisition is 3 times more costlier than retention
Identifies customers with Churn risk with the purchase and behavioral data
Past churned customers with similar patterns
Survival analysis, Classification to identify potential customers with Churn Risk
Marketing automation tool runs Churn save series for customers who enter this life cycle
Logistic Regression Survival Analysis
16. MODERN MARKETERS TOOLKIT
▪ Tag manager ( Google tags )
▪ Third party DMP ( Bluekai, Axiom , Audience manager )
▪ Big data processing ( Hive, Hbase, Mongo, Cassandra )
▪ Cloud Computing ( AWS, GCP, Azure )
▪ Marketing automation ( Responsys, Clevertap, Mailchimp, Pardot, Marketo )
▪ Analytics ( Google Analytics, Webanalytics, Omniture )
17. LIST VS MARKETING AUTOMATION
List Marketing Marketing Automation
Manual selection of customer segments Customers enter the series as they become eligible
More dependency on the campaign development
team
More dependency on Data Engineers and Data
science Teams
Dependent on the availability of team No dependency on team, runs 24x 7
Potential Compliance risk. No central repository of
responses, opts outs for compliance
All the responses, opt outs are central
No data connectedness of Channels Total data connectedness of behavior and
responses across channels