This document discusses using Hadoop for big data analytics. It begins by explaining how big data is increasing in volume, variety, and velocity. It then contrasts the traditional approach of structured SQL databases with Hadoop's ability to handle unstructured data at scale in an open and affordable way. The document provides examples of using Hadoop for customer innovation through analytics on user behavior data. It outlines some early adopter success drivers for Hadoop projects, including IT partnerships with business units and embracing analytics hubs for sharing and reusing work. Finally, it demonstrates a use case for analyzing customer data on Hadoop to gain insights and improve the customer experience.
Gen AI in Business - Global Trends Report 2024.pdf
Why Hadoop is the New Infrastructure for the CMO?
1.
2.
3. Business Value of Hadoop
Adoption Success Criteria
“Jumpstarting” Your Hadoop
Big Data on Hadoop
4. > OMG! – The Opportunity - What Do We Do With All This Data?
Volume, Variety and Velocity all increasing rapidly
Source: IBM, Oct 2012
5. Big Data on Hadoop
A Needed Technology Disruption
Told Us What Happened & Made Us More Efficient
BI Using
Data Cube
Analysis
Structured, Sampled
Transitional, Closed,
Expensive
RDBS & EDW
SQL
Driven by Vendors
Infrastructure for Sales,
Finance & Operations
CFO
COO
CRO
Supply Chain
ERP
Sales Ops
6. Big Data on Hadoop
Technology Disruption
What Can We Make Happen?
“Database of Intent”
Analytics Using
Total Fidelity
Analysis
Unstructured
Behavioral,
Open, Affordable
HADOOP
HQL
BI Using
Data Cube
Analysis
Structured, Sampled
Transitional,
Closed, Expensive
RDBS & EDW
SQL
Driven by Open SourceDriven by Vendors
7. Big Data on Hadoop…What is it Good For?
Customer Innovation – Optimize the Customer Engagement Cycle
BUY USE
SUPPORT
Customer
Innovation
Personalized
Cross&Upsell
Channel
Analytics
Superior
Products
Telemetry
Analytics
Proactive
Support
Response
Analytics
13. #1. IT Partnership With LOB (Marketing)
Find Use Case
Identify Budget
Form project teams
Partner on a small POC
Educate
Big Data on Hadoop
Success Driver #1
14. “By 2016 the CMO will have
more budget than the CIO”
- Gartner Group
Big Data on Hadoop
15. #2. Use Second-Generation Big Data Analytics
Support both power analysts & business users
Work collaboratively on projects
Empower Interactive Visualizations
Facilitate the new “Big Data” workflow
Gamify
Big Data on Hadoop
Success Driver #2
16. #3. Embrace Analytics Hubs…They’re Coming!
Sharing
Re-Use
Enforce Standards
Analytics “Assembly”
Big Data on Hadoop
Adoption Driver #3
17. Go Native & Get Total Fidelity for
Analytics on Hadoop
Very rich
Not sampled
No data replication or movement
Low complexity and TCO
Big Data on Hadoop
19. Data
Warehouse
OLTP to OLAP
Mapping
Analyst
BI Using Data Cube Analysis
Analysts Worked with Transformed, Aggregated, Sampled Data
Ordering App
Financial App
Master Data
Staging
OLAP
Reports
BI Using
Data Cube
Analysis
Structured, Sampled
Transitional,
Closed, Expensive
RDBS & EDW
SQL
Driven by Vendors
20. Analysts Access All the Data With a New Data WorkFlow
Iterative & Adaptive
Big Data Analytics on Hadoop Using Total Fidelity Analytics
Application
AnalyticsData
Unstructured
Behavioral,
Open, Affordable
HADOOP
HQL
Driven by Open Source
Analyst
Analytics Using
Total Fidelity
Analysis
21. 1. How do we increase traffic to our site?
2. How do we make customers stay on the site, come
into the store and engage with our offerings?
3. How do we convert browsers to buyers?
4. What else can promote knowing their profile?
5. How can make them come back and shop for more?
What kind of customer innovation can we drive?
22. 1. How often do they visit, what did they buy, how
much did they spend?
2. What did they view, how long did they stay on the
site? What did they click on? What did they rate?
What did their friends buy?
3. How do we offer the most relevant product for
service and invite them for the right campaign.
What insights do we need?
To better understand my customer we need a more
granular segmentation (micro-segmentation)
23. The Signals of Big Data on Hadoop
Logs - Search terms, page views,
useragent , Geo, IP, duration, size...
Campaigns, keywords,
channels, SEO, Display, Affiliates
Reviews - SKU, date, who,
comment, rating, location
Products-
SKU, categories, bundles, descripti
on, prize
Orders – SKU, prize, purchased
with, Shipping date, status
Profiles – Names, location, gender,
demographics, reach, interests, influence
24. 1. Describe and prepare the data
2. Perform initial analysis on raw data
3. A number of insights can be derived without further
analysis (RFM)
4. Identify feature set and extract training data for
modeling
5. Create the model in any model authoring tool
6. Score the model in Hadoop
7. Use the insight to improve business value
Steps in a typical analysis
25. Complexity of data compared to transactional world
{"frequentlyPurchasedWith": [], "color": "Black", "skutype": "parent", "productTemplate":
"Computer_Accessory", "salesRankMediumTerm": "3039", "shortDescription": "Compatible with Windows 8 and
RT and Android 3.0 tablets; Bluetooth technology; convertible stand/carrying case", "includedItemList":
[{"includedItem": "Logitech Tablet Keyboard for Windows 8 and RT and Android 3.0+ Tablets"}, {"includedItem":
"4 AAA batteries"}, {"includedItem": "Owner's manual"}], "subclassId": 2409, "sku": 6541967, "width": "12.3"",
"subclass": "BLUETOOTH KEYBOARDS", "source": "BoxStore", "modelNumber": "920-004569", "digital": false,
"department": "COMPUTERS", "type": "HardGood", "productId": 1218752781558, "description": "None",
"technologyCode": "None", "longDescription": "This Logitech 920-004569 keyboard features a low-profile, 65-key
design for easy, comfortable typing on your Windows 8 or RT or Android 3.0 tablet. The convertible stand allows
comfortable viewing and provides on-the-go protection for the keyboard.", "categoryPath": [{"name": "Box
Store", "id": "cat00000"}, {"name": "Computers & Tablets", "id": "abcat0500000"}, {"name": "iPad, Tablets & E-
Readers", "id": "pcmcat209000050006"}, {"name": "Tablet Accessories", "id": "pcmcat231800050009"}, {"name":
"Tablet Docks, Keyboards & Stands", "id": "pcmcat242000050003"}], "manufacturer": "Logitech", "classId": 492,
"upc": "097855090973", "regularPrice": 69.99, "class_1": "TABLET ACCESSORIES", "relatedProducts": [{"sku":
4974041}, {"sku": 1306578835}, {"sku": 4640745}, {"sku": 9610542}, {"sku": 8785729}, {"sku": 6640676}]}
{"frequentlyPurchasedWith": [], "color": "Gray", "skutype": "parent", "productTemplate": "Computer_Accessory",
"salesRankMediumTerm": "None", "shortDescription": "Compatible with BlackBerry Playbook tablets; wool
construction; TPU plastic cradle; elastic band; metallic clip; functions as a stand; play-through design",
"includedItemList": [{"includedItem": "DICOTA TabBook Case for BlackBerry Playbook Tablets"}], "subclassId":
2404, "sku": 6738835, "width": "5.5"", "subclass": "SO TABLET ACCY", "source": "BoxStore", "modelNumber":
"D30203", "digital": false, "department": "COMPUTERS", "type": "HardGood", "productId": 1218789793935,
"description": "None", "technologyCode": "None", "longDescription": "This DICOTA TabBook D30203 case helps
keep your BlackBerry Playbook tablet safe from hazards, with wool construction and a TPU plastic cradle for
durability and an elastic band to keep your tablet snug and secure in the case.", "categoryPath": [{"name": "Box
Store", "id": "cat00000"}, {"name": "Computers & Tablets", "id": "abcat0500000"}, {"name": "iPad, Tablets & E-
Readers", "id": "pcmcat209000050006"}, {"name": "Tablet Accessories", "id": "pcmcat231800050009"}, {"name":
"Tablet Cases, Covers & Sleeves", "id": "pcmcat242000050002"}], "manufacturer": "DICOTA", "classId": 492,
"upc": "7332752000964", "regularPrice": 49.99, "class_1": "TABLET ACCESSORIES", "relatedProducts": [{"sku":
26. How big your data can grow? Number of unique visitors per day