Using	Geode	to	Enhance	Customer	Experience
2Copyright © Capgemini 2015. All Rights Reserved
What	problem	are	we	trying	to	solve?	
Every	retailer’s	problem	to	solve…	
Personalized Service
Mobile
check-
out
Optimized Staffing,
Tasks and Training
Real-time Inventory
Tracking
Assortment and Price
Differentiation
Endless Aisle
Facilities
Control
Fault
Detection
Dynamic
Layout
Platform Data SecurityAnalytics
Virtual Wall
Dynamic Labels
Smart Digital Signage
Customer Devices
END POINT CAPABILITY PROCESS CORE PROCESS END POINT CAPABILITY
Knowledgeable Sales
Associates
Heating and Lighting
Robotics
POC and Mobility
Clienteling Apps
Customer
Employee
Physical
Store
Product
Customer Profiling
Traffic, Purchase Patterns
In-store Interactions
Personalized
Recommendations
What	retailer’s	“Loyalty	2.0”	might	be	
grappling	with…		
1. Omni-Channel,
Seamless Commerce
2. Insights & Data
3. Marketing Resource
Management
4. Applied Innovation
5. Unified View of the
Customer
Next	GeneraFon	Loyalty	Programs	requirements…	
The	holy	grail	of		
customer	loyalty	
Delivering insights at
the point of action in
“human time”
3Copyright © Capgemini 2015. All Rights Reserved
Some	key	dimensions	of	the	problem…	
§  Retailers	have	many	Big	and	Fast	Data	challenges	...	
	
§  Data	volumes	–	number	of	companies	storing	100+	TB	is	growing	…	
…	but	they	are	analyzing	only	a	frac%on	of	their	data	assets	
	
	
	
	
	
	
	
	
	
§  Most	retailers	have	10’s	to	100’s	of	systems	churning	out	data	in	real	Fme	
-	plus	external	data	sources	(social	media,	weather,	demographic,	spaFal)	
that	they	should	be	using	…	
…	but	they	struggle	to	store	and	integrate	data,	and	manage	data	complexity		
	
>	today,	BATCH	rules	
0
50
100
Structured Semi-Structured Unstructured
Not Analyzed Analyzed
Volume
Velocity
Variety
4Copyright © Capgemini 2015. All Rights Reserved
Just	imagine:	urban	living	in	2025	and	the	requirements	of	the	
retail	environment.	
Consumer	engagement	
Taking	part	in	a	dialogue	with	consumers,	jusFfying	
their	trust	in	the	industry	
	
Transparency	
Keeping	consumers	informed	about	products’	key	
a`ributes,	ingredients,	nutrients	and	provenance	as	
well	as	their	environmental	and	societal	impacts	
	
The	last	mile	of	distribu%on	–	both	to	the	retail	
store	and	to	the	consumer	
Reconsidering	the	assumpFon	that	it	is	an	area	where	
companies	operate	independently	of	each	other,	and	
exploring	opportuniFes	to	collaborate	to	improve	
speed,	efficiency	and	consumer	saFsfacFon	
Enabled	by	Modularized	Technology	
Business	agility	and	rapid	collaboraFon	require	the	
transformaFon	from	rigid	and	purpose-built	IT	
structures	into	mulF-use,	component-based	technology	
capabiliFes	which	allow	for	easy	assembly	or	
disassembly	according	to	business	needs.
5Copyright © Capgemini 2015. All Rights Reserved
Everything	changes	before	tomorrow	
Click Call Read
Click
Location
Push
Engage
Lifestyle
Location
Demand
?
6Copyright © Capgemini 2015. All Rights Reserved
Why	tomorrow	is	too	late	
I’m at the store
in Charlotte
Human-Time
Social Analytics
Store
Master
Customer
Master
Human-Time
Business
Analytics
Customer
Transactions
Product
Master
Store ABC
Jane Doe
Jane
Hey Shovel
fans: Offer
Code!
Great: On our
way!
Jane’s Friends
7Copyright © Capgemini 2015. All Rights Reserved
Geode Cluster
How	we	are	using	Geode	for	this	
Data Lake
Geode Geode Geode Geode Geode Geode
Write Back Batch Analytics Services
Reference
Data
Enterprise
History
Cache
Fast
Speed
Consumer
Data
Human-Time analytics
Bulk Enterprise Data
8Copyright © Capgemini 2015. All Rights Reserved
Advantages	of	a	no-RDBMS	architecture	
§ Geode/Gemfire handles the transactional integrity
§ Data Lake stores the history
§ Developer model is much more suited to Java
developers than RDBMS
§ Enables multiple other engines to be added to the
Batch part
§ Neo4J for network/graph analytics
§ Cost
9Copyright © Capgemini 2015. All Rights Reserved
What	could	tomorrow	look	like?	
The Present Tomorrow?
Retailers send consumers
coupon books generated on a
monthly basis.
Retailers figure out consumer shopping patterns, and deliver offers the day before the
consumer plans to shop. No coupons – the consumer has an electronic wallet that pops
up in-store.
Customers get printed coupons
when they go through the check
out line.
Consumers scan products as they shop, interact with “recommender” systems they like,
and skip the check out line. Apple Pay.
-  Recipe Help
“It looks like you’re making chili – but you didn’t pick up any kidney beans – did you
forget?”
-  Healthy Choices
“I see you just selected potato chips – would baby carrots be a better choice?”
-  Market Share Shifting
“There’s an extra 20% off if you choose Doritos instead of Lays today.”
-  Managing Perishable Store Inventory
“I know you like lettuce – there’s special pricing if you buy today.”
Hint: Store ordered too much lettuce, it’s going to spoil if not sold soon.
Customers push their carts up
one aisle and down the other.
“I see that it’s been several weeks. Have you been busy? Since you are one of our best
customers, would it help if we were to suggest some items to tide you over, and deliver
them by Uber right now?”
“The weather forecasters are predicting a blizzard. Do you need some extra supplies?
Do you need bottled water for baby formula? We can deliver…”
10Copyright © Capgemini 2015. All Rights Reserved
Enterprise	Big	Data	Reference	Architecture	–		
Business	Data	Lake	
Hadoop – Foundational Elements
REST
HTTP/S
Stream
SPARK-ML (Formerly Mahout)
ClassificationClustering
Collaborative
Filtering
Processing & Computation
In-Memory
Process
SPARK R
SPARK
YARN
Business Data Lake Zones
Interactive
Analytics
Visualization
& Reporting
Analytical
Tools,
Simulation &
Languages
Enterprise
Solutions
(API – XML –
Json)
Localized
Data Sources
Deep Learning
Machine Learning, Evolutionary/Genetic Programming, Complex Event Processing & Enterprise Automation
Tachyon
Load & Refine
Scoop
Flume
Hive
NFS
WebHDFS
Pig
Speed Layer Zone 0 Zone 1 Zone
5
Zone 2 Zone 3 Zone 4
Cleanse
d
Ideation
SandboxEnriched
Detailed
Modeled
Aggregate
ModeledStagin
g
Metadata
Security
Knox Ranger Kerberos
Provision, Monitor,
Manage
Ambari
Zoo
Keeper
Schedul
e
Oozie
Deployment ChoiceLinux Windows On-Premises Cloud
Hue
NO SQL- Slider
HIVE - TEZ
Phoenix
HBase
Solr (Search)
Landin
g
SOURCES
Geolocation
IT Systems
Sensor & Machine
Server Logs
Web & Social
Clickstream
Unstructured
Apache
Geode
The	new	digital	divide	
•  The	new	digital	divide:	the	gap	between	consumer’s	digital	
behaviors	and	expectaFons	–	in	contrast	to	the	readiness	and	
ability	of	retailers	to	deliver	on	the	desired	experiences.	
	
	
–  Percentage	of	shoppers	100%	connected	–	growing	quickly	
–  ExpectaFons	–	evolving	rapidly	
–  “Showrooming”	–	e.g.	fear	that	digital	drives	consumers	online	(it’s	a	myth)	
–  Micro-characterisFcs	–	geographic,	demographic,	ethnic,	social,	age,	gender	
–  Ability	of	retailers	to	harness	technology	that	–	
•  Permits	applicaFons	to	be	built	quickly,	dismantled	quickly	
•  Delivers	consumer	behavior-influencing	acFons	in	real	Fme	
•  Drills	down	to	individual	consumers	
•  Integrates	data	–	from	inside	and	outside	the	enterprise
13Copyright © Capgemini 2015. All Rights Reserved
….and	how	could	we	do	this?	
Unleash Data and Insights
as-a-service
Make Insight-driven
Value a Crucial
Business KPI
Empower your People
with Insights at the
Point of Action
Develop an Enterprise Data
Science Culture
Master Governance,
Security and Privacy of your
Data Assets
Enable your Data
Landscape for the Flood
coming from Connected
People and Things
Embark on the Journey
to Insights within your
Business and
Technology Context
1 2 3
7654
Join the Apache Geode Community!
•  Check out: http://geode.incubator.apache.org
•  Subscribe: user-subscribe@geode.incubator.apache.org
•  Download: http://geode.incubator.apache.org/releases/

#GeodeSummit - Using Geode as Operational Data Services for Real Time Mobile Experience

  • 1.
  • 2.
    2Copyright © Capgemini2015. All Rights Reserved What problem are we trying to solve? Every retailer’s problem to solve… Personalized Service Mobile check- out Optimized Staffing, Tasks and Training Real-time Inventory Tracking Assortment and Price Differentiation Endless Aisle Facilities Control Fault Detection Dynamic Layout Platform Data SecurityAnalytics Virtual Wall Dynamic Labels Smart Digital Signage Customer Devices END POINT CAPABILITY PROCESS CORE PROCESS END POINT CAPABILITY Knowledgeable Sales Associates Heating and Lighting Robotics POC and Mobility Clienteling Apps Customer Employee Physical Store Product Customer Profiling Traffic, Purchase Patterns In-store Interactions Personalized Recommendations What retailer’s “Loyalty 2.0” might be grappling with… 1. Omni-Channel, Seamless Commerce 2. Insights & Data 3. Marketing Resource Management 4. Applied Innovation 5. Unified View of the Customer Next GeneraFon Loyalty Programs requirements… The holy grail of customer loyalty Delivering insights at the point of action in “human time”
  • 3.
    3Copyright © Capgemini2015. All Rights Reserved Some key dimensions of the problem… §  Retailers have many Big and Fast Data challenges ... §  Data volumes – number of companies storing 100+ TB is growing … … but they are analyzing only a frac%on of their data assets §  Most retailers have 10’s to 100’s of systems churning out data in real Fme - plus external data sources (social media, weather, demographic, spaFal) that they should be using … … but they struggle to store and integrate data, and manage data complexity > today, BATCH rules 0 50 100 Structured Semi-Structured Unstructured Not Analyzed Analyzed Volume Velocity Variety
  • 4.
    4Copyright © Capgemini2015. All Rights Reserved Just imagine: urban living in 2025 and the requirements of the retail environment. Consumer engagement Taking part in a dialogue with consumers, jusFfying their trust in the industry Transparency Keeping consumers informed about products’ key a`ributes, ingredients, nutrients and provenance as well as their environmental and societal impacts The last mile of distribu%on – both to the retail store and to the consumer Reconsidering the assumpFon that it is an area where companies operate independently of each other, and exploring opportuniFes to collaborate to improve speed, efficiency and consumer saFsfacFon Enabled by Modularized Technology Business agility and rapid collaboraFon require the transformaFon from rigid and purpose-built IT structures into mulF-use, component-based technology capabiliFes which allow for easy assembly or disassembly according to business needs.
  • 5.
    5Copyright © Capgemini2015. All Rights Reserved Everything changes before tomorrow Click Call Read Click Location Push Engage Lifestyle Location Demand ?
  • 6.
    6Copyright © Capgemini2015. All Rights Reserved Why tomorrow is too late I’m at the store in Charlotte Human-Time Social Analytics Store Master Customer Master Human-Time Business Analytics Customer Transactions Product Master Store ABC Jane Doe Jane Hey Shovel fans: Offer Code! Great: On our way! Jane’s Friends
  • 7.
    7Copyright © Capgemini2015. All Rights Reserved Geode Cluster How we are using Geode for this Data Lake Geode Geode Geode Geode Geode Geode Write Back Batch Analytics Services Reference Data Enterprise History Cache Fast Speed Consumer Data Human-Time analytics Bulk Enterprise Data
  • 8.
    8Copyright © Capgemini2015. All Rights Reserved Advantages of a no-RDBMS architecture § Geode/Gemfire handles the transactional integrity § Data Lake stores the history § Developer model is much more suited to Java developers than RDBMS § Enables multiple other engines to be added to the Batch part § Neo4J for network/graph analytics § Cost
  • 9.
    9Copyright © Capgemini2015. All Rights Reserved What could tomorrow look like? The Present Tomorrow? Retailers send consumers coupon books generated on a monthly basis. Retailers figure out consumer shopping patterns, and deliver offers the day before the consumer plans to shop. No coupons – the consumer has an electronic wallet that pops up in-store. Customers get printed coupons when they go through the check out line. Consumers scan products as they shop, interact with “recommender” systems they like, and skip the check out line. Apple Pay. -  Recipe Help “It looks like you’re making chili – but you didn’t pick up any kidney beans – did you forget?” -  Healthy Choices “I see you just selected potato chips – would baby carrots be a better choice?” -  Market Share Shifting “There’s an extra 20% off if you choose Doritos instead of Lays today.” -  Managing Perishable Store Inventory “I know you like lettuce – there’s special pricing if you buy today.” Hint: Store ordered too much lettuce, it’s going to spoil if not sold soon. Customers push their carts up one aisle and down the other. “I see that it’s been several weeks. Have you been busy? Since you are one of our best customers, would it help if we were to suggest some items to tide you over, and deliver them by Uber right now?” “The weather forecasters are predicting a blizzard. Do you need some extra supplies? Do you need bottled water for baby formula? We can deliver…”
  • 10.
    10Copyright © Capgemini2015. All Rights Reserved Enterprise Big Data Reference Architecture – Business Data Lake Hadoop – Foundational Elements REST HTTP/S Stream SPARK-ML (Formerly Mahout) ClassificationClustering Collaborative Filtering Processing & Computation In-Memory Process SPARK R SPARK YARN Business Data Lake Zones Interactive Analytics Visualization & Reporting Analytical Tools, Simulation & Languages Enterprise Solutions (API – XML – Json) Localized Data Sources Deep Learning Machine Learning, Evolutionary/Genetic Programming, Complex Event Processing & Enterprise Automation Tachyon Load & Refine Scoop Flume Hive NFS WebHDFS Pig Speed Layer Zone 0 Zone 1 Zone 5 Zone 2 Zone 3 Zone 4 Cleanse d Ideation SandboxEnriched Detailed Modeled Aggregate ModeledStagin g Metadata Security Knox Ranger Kerberos Provision, Monitor, Manage Ambari Zoo Keeper Schedul e Oozie Deployment ChoiceLinux Windows On-Premises Cloud Hue NO SQL- Slider HIVE - TEZ Phoenix HBase Solr (Search) Landin g SOURCES Geolocation IT Systems Sensor & Machine Server Logs Web & Social Clickstream Unstructured Apache Geode
  • 12.
    The new digital divide •  The new digital divide: the gap between consumer’s digital behaviors and expectaFons – in contrast to the readiness and ability of retailers to deliver on the desired experiences. –  Percentage of shoppers 100% connected – growing quickly – ExpectaFons – evolving rapidly –  “Showrooming” – e.g. fear that digital drives consumers online (it’s a myth) –  Micro-characterisFcs – geographic, demographic, ethnic, social, age, gender –  Ability of retailers to harness technology that – •  Permits applicaFons to be built quickly, dismantled quickly •  Delivers consumer behavior-influencing acFons in real Fme •  Drills down to individual consumers •  Integrates data – from inside and outside the enterprise
  • 13.
    13Copyright © Capgemini2015. All Rights Reserved ….and how could we do this? Unleash Data and Insights as-a-service Make Insight-driven Value a Crucial Business KPI Empower your People with Insights at the Point of Action Develop an Enterprise Data Science Culture Master Governance, Security and Privacy of your Data Assets Enable your Data Landscape for the Flood coming from Connected People and Things Embark on the Journey to Insights within your Business and Technology Context 1 2 3 7654
  • 14.
    Join the ApacheGeode Community! •  Check out: http://geode.incubator.apache.org •  Subscribe: user-subscribe@geode.incubator.apache.org •  Download: http://geode.incubator.apache.org/releases/