4. Smarter Way. Faster Growth.SM
4
People Presence Performance
• Record speed to shelf
• World-class sell-in rates
• Industry-recognized in-store
and out-of-store programs
• 40,000+ employees
• Highly trained & experienced
• World-class talent
• 1,100+ customer buying points
• 550,000+ US retail outlets served
• 29,000,000+ in-store hours/yr
• 7,000,000+ out-of-store hours/yr
GROWING BRANDS FOR OVER 110 YEARS
HEADQUARTERED IN PLANO, TX WITH OFFICES THROUGHOUT THE U.S., CANADA,
MEXICO, AUSTRALIA, AND NEW ZEALAND
5. At the Intersection of CPG and Retail Data
5
CROSSMARK must be able to natively communicate with Manufacturers,
Retailers and Consumers the way they see the world.
CROSSMARK works with CPG
and Retailer firms to help brands
grow faster and more profitably. Consumers
Retailers Manufacturers
• Digital, social, experiential, affinity, panel,
consumer incentives
• Hundreds of thousands of direct engagements
• Millions of loyalty/affinity card members
Helping manufacturers activate their
products with retailers and
consumers.
• Planning, forecasts, sales to
retailer, promotion, digital
Helping retailers manage their
store activities.
• Instore activation and execution
data
• POS, assortment, planograms,
beacons, loyalty, digital.
• Billions of direct observation
data points at retail per year
Manufacturers
Consumers
Retailers
6. 6
How do you define a place?
Blood Mountain as part of Georgia
Appalachian Trail, via Google Maps
USGS Benchmark Coin
Appalachian Trail Section Map from
“2015 Northbound The AT Guide”
Blood Mountain, GA via Google Maps
34.74002,-83.93743
7. How our Customers Define a Place
7
There are inconsistencies in definitions within the same industry,
and even the same company on the definition of a place.
Consumer Packaged Goods
Place may be a reflection of
supply chain focus (ERP), including
Ship to’s, Sold to’s, distribution
centers, and wholesalers.
For Marketing and Sales, “Place”
may be an industry definition of a
location from Nielsen TDLinx or
VIP.
Retailers
Place may be individual stores,
regions, divisions or market areas.
Retailers may also look at place
from a source of supply, including
their own distribution centers.
8. When is a Place not just a Place?
8
The Superbowl isn’t
just a 4 hour event on
television
Or what happens in
the stadium
10. When a Place is More Than a Place
10
The right Place definition enables us to monitor & measure the interaction with brands at key events
before, during and after the event.
And integrate with our internal knowledge & execution data to identify unique opportunities.
11. How CROSSMARK Defines a Place
11
A Place is not just where we sell. It is anywhere we engage in the activation of brands. We must
be able to monitor, target, plan and activate against one definition of a Place.
Location Area
A Location is a specific place registered for an activity.
An area is a set of geocoordinates that defines a space,
path or area. The benefit of an area is it allows us to
identify all Locations and Sublocations within or near the
area, without having to name each Location.
The purpose of the Area is to enable better targeting and
planning.
Retail outlets
Residences/domiciles
Business locations
Stadiums
Amusement parks
Parks
Mass transit locations
Government offices
Standard circumference around a location (KML)
Effective boundary of a large Location (e.g. amusement
park, stadium)
Street festivals
A path such as a trail or parade route
Predefined Areas could include countries, states,
MSAs/DMAs, counties, census blocs
WhatitisExamples
13. The Situation
13
CROSSMARK had seen it’s technology ecosystem grow organically over 15 years,
consisting of primarily internally built, internally hosted solutions.
Our new CIO started the digital transformation of the organization with a push to a
cloud-first approach centralized around Salesforce and the Force.com platform.
In addition to transforming the application architecture, CROSSMARK still had traditional
MDM challenges:
• “Master data management” was a new formal practice
• Extensive point-to-point integrations
• Multiple versions of the truth; inefficient “mastering” processes
14. Point-to-Point Integrations of Master Data
14
SalesTrak
Employees
D2
SalesTrak Rep-
Supervisor
Relationship
D4
Service Now
Employees
D18
Concur Users
D4
Active
Directory
Internal
Accounts
D6
Office 365
Accounts
D7
Kenexa
Employees
D20
Kenexa
Applicants
D22
Import Lawson
Employees into
SalesTrak
P1
Import Lawson
Employees into
Kenexa
P6
Import Hiring
Managers into
Kenexa
P5
Kenexa Hiring
Managers
D21
Import Lawson
Employees into
KMS
P7
KMS
Employee
Master
D24
KMS New
Hires
D23
Import Hired
Applicants into
KMS
P31
Service Now
Division
Managers
D19
Update
Employee-
Supervisor
Relationship for
Retail Reps
P3
SalesTrak Field
Management
Interface
I12
Retail Supervisors
Change Rep Territory
Assignment
P2
Kenexa Applicant
Tracking
Interface
I6
Export New Hires Who
Have completed their
Onboarding
P32
New Hire File
D24
Import New Hire File
into Lawson
P33
Fill open
positions with
new hires or
existing
employees
P32
Lawson
Employee
Change Log
D12
Export
Employee
Changes to
Change Log
Table
P36Personnel Action
Form
I13
Submit a PAF for
Employee
Changes to HR
P35
MindLeaders
I9
Putnam
I10
Benefit Focus
I11
Mercury
Fulfillment
I14
TALX
I8
Add Employee ID
Into Kenexa
P34
Division
Manager Table
D11
Extract List of
Division
Managers to
Service Now
Employee ID-AD
Cross-Reference
Table
D26
Extract Employee
ID & Active
Directory
Account
Active Directory
Changes
P40
Employee
Database View
Lawson to
Concur
Employee Export
P4
Address Change
Batch file
Direct Deposit &
Emergency
Contact Updates
AMS
Employees
D10
CROSSView
Planner Interface
I10
List of emps who need Food
Safety certifications to
MindLeaders
P10
List of Employees
to Putnam
P11
List of Employees
to Benefits Focus
P12
List of Employees to
Mercury Fulfillment
P16
Employee data and ACA
data to TALX
P19
Mercury
Fulfillment
I14
Cigna
I1
List of Employees to
Mercury Fulfillment
P16
(Future) Employee data
to Cigna
P29
ComData
I2
Chase Pay Cards
I3
Employee data to
ComData
P21
Send list of employees who
will be paid on a payment
card
P23
List of all
Employees to
AMS
P13
List of Employees
to Service Now
P17
Manually Update
Division Manager
List
Send list of employees who
will be paid on a payment
card
P23
Timesheet
Employee Data
D8
Send list of employees who
will be paid on a payment
card
P23
Timesheet
Manager data
D9
Send list of employees who
will be paid on a payment
card
P23
Timesheet
Manager data
D9
Employee data to Workers'
Comp Insurance provider
P27
Workers Comp
Insurance Provider
I5
Employee Self
Serve
I15
KMS User
Interface
I7
New Hire Fills
out KMS
Paperwork
BlueVue
I16
Employee data to
BlueVue
P25
Gigwalk Entity
Integration
D17
New Employee
Mailbox Creation
P36
Employee Data
to Gigwalk
P28
Gigwalk
Customers
I17
Hyperion Employee
D15
Employee data to
Hyperion
P26
Employee Client
Permissions
D5
Manually
Manage List of
Employee Client
Permissions
P37
CAP User Create
Application
I19
Create CAP (Au & NZ)
users in Active
Directory
P41
Create CAP (Australia &
NZ) employees in
SalesTrak
P42
Lawson
Employees
D1
The previous “master data” approach was a series of
point-to-point-to-point integrations.
• Hosted entirely internally with limited ability to
integrate with external solutions.
• Often conflicting data from multiple sources creates
multiple sources of truth.
• Multiple silo’s of management for the same domains
• Added cost and complexity to development and
support
• Increased operational costs due to incorrect master
data.
15. Complex Authoring Process(es)
15
Receive list with no TDLinx
IDs
Size?
Preferred
Retailer?
Compare with Store
Assignment Spreadsheet
Is complete?
Compare with STNG
Match Locations
Location #
Concatenation
Banners
Group match
No conforming
No enrichment
Opportunity to check
active status
Is inactive &
needs to be
active?
New Store? Ticket to create TDLinx IDs
Ticket to change status
YesNo
NoYes
Yes
No
Yes
Small
IT Operations Receives
Ticket
No
TDLinx Match?
Location DB
TDLinx Request
Surrogate ID
Yes
Receive updated TDLinx ID
Nielsen receives and
processes request
Large
Match Locations
Location #
Concatenation
Banners
Group match
No conforming
No enrichment
Opportunity to check
active status
Functional Teams IT Operations Nielsen
16. High Upfront Costs of Traditional MDM Solutions
16
Software
purchase
On-premise server
purchase
Traditional MDM solutions
require a large upfront
CapEx investment
Realized value
+
+
Ongoing support
Ongoing support
Ongoing support
Year 0 Year 1 Year 2
18. MDM Priorities
18
CROSSMARK’s key priorities for Master Data Management: Place.
01 Drive increased business agility. Define a Place once in a way that enables us to plan, manage and engage
with CPG firms, Retailers and Consumers.
03
Elastic Scalability. The solution must allow CROSSMARK to efficiently scale as the company grows.
02
04 Cloud-first architecture. The MDM solution must support integration with Salesforce, Force.com and other
cloud providers, as well as internal platforms.
Enhanced Authoring. The solution must reduce the operating time needed to author master data while
maintaining data quality.
05
Accelerate Timelines. The MDM solution needed to help drive Digital Transformation, despite starting 14
months after the inception.
19. Project Approach
19
• Leveraged cloud based platform to accelerate the realization of value.
• Custom Force.com application for the creation of Place records.
• Dell Boomi MDM for the mastering and management of Places.
• Dell Boomi for integration between Force.com, Boomi MDM and other systems.
• Loquat for address cleansing and enrichment.
• Google Location API for additional enrichment.
• Place data sourced from multiple locations.
• Initially Agile approach, later switched to Kanban.
20. Development & Implementation Timeline
20
Total time from initial kickoff to production was 4 months.
Cleansing
Using the Boomi MDM process and other
tools, cleansed 800,000 place records.
Design
Modeled existing & future needs for Place
data.
Develop
Developed authoring UX and
Boomi integrations, including
advanced matching
algorithms.
Training
Developers attended both Dell
Boomi MDM and Force.com
training.
Finish
START
Data cleansing & preparation
Scoping
Integration
Design
Develop
Migration
21. Cloud MDM Powers Our Modern Architecture
21
Enterprise Workforce
Management
Managing a workforce of 35,000
in the US, Canada, Mexico, New
Zealand and Australia
Enterprise Reporting
& Analytics
CROSSMARK leverages cloud based
reporting to power tens of
thousands of both internal
employees and our customers.
Retailer Solution
CROSSMARK has built several
proprietary industry solutions on
the Force.com platform, including
our widely recognized Retailer
Management Solution.
Internal ERP
CROSSMARK still hosts billing,
payroll and finance applications
onsite.
Support Systems
CROSSMARK leverages many
cloud based support platforms
to run it’s business.
MDM Authoring Platform
Using the Force.com platform and
Dell Boomi MDM, CROSSMARK
built a location authoring platform.
22. Multisource Model of Place Data
22
Place
One record for a placeCROSSMARK Source
Syndicated Source
Retailer Source
Client Source
CROSSMARK receives Place data from vendors and customers that are
often duplicative but not linked.
The Dell Boomi MDM platform provides us the ability to develop
custom models that span our clients, vendors and our definition of a
domain.
When a Place from a different source is identified, the system
establishes a cross-reference to that source identifier.
This enables CROSSMARK to master a Place once, while being able to
communicate natively to our customers with their identifiers.
23. Force.com Mastering Application
23
Individual Entry
Enables data entrants to enter
or modify individual domain
records.
Bulk Upload
Provides data entrants the
ability to directly upload up to
10,000 places via a file.
24. Dell Boomi MDM Stewarding Interface
24
Quarantine Summary
Dell Boomi MDM provides a consolidated view for data stewards
to manage records in the quarantine.
Quarantine Detail
Dell Boomi MDM provides prebuilt views for data stewards to
view and depose quarantined records.
25. Leveraging API’s & Algorithms Radically Speed Mastering
25
Receive list with no TDLinx
IDs
Match Locations Process
Location #
Concatenation
Banners
Group match
No conforming
No enrichment
Opportunity to check
active status
Ticket to create TDLinx IDs
and/or change status
Yes
IT Operations Receives
Ticket
TDLinx Match?
Location DB
TDLinx Request
Surrogate ID
Yes
Receive updated TDLinx ID
Nielsen receives and
processes request
Check against Location DB
Functional Teams IT Operations Nielsen
New Authoring Process
We decreased our location mastering time from a maximum of
4 weeks (50,000 places) of all manual effort, to no more than a
week.
• We have one process for the creation and mastering of
Places.
• The process uses Loquat service to enrich and standardize
relevant location records across 170 countries.
• We integrate with additional API’s to further enrich records.
• We developed advanced algorithms and fuzzy matching on
Boomi MDM to manage previously manual tasks.
• We have decreased the amount of manual touching of
records.
• Most records are processed automatically within seconds.
26. Cloud-based MDM Efficiently Scales to Size of the Business
26
Software
license
Boomi MDM enables
CROSSMARK to quickly
scale out cloud based
AtomSphere nodes to meet
increased demand as our
business continues to grow
Realized value
Year 0 Year 1 Year 2
Software
license
Software
license
27. MDM: Place Maturity
27
Level 1
Initial
Level 2
Developing
Level 3
Defined
Level 4
Managed
Level 5
Optimizing
Vision None Bottom-up initiatives only
Top-down, but limited in
vision
Unifying vision for cross-
enterprise MDM
Key enabler of business
success
Strategy None No enterprise strategy
Focus on domains or use
cases
Consistent and integrated
approach
Ongoing investment and
improvement
Governance No one has responsibility
IT-led; lacking business
involvement
Domain-level governance
Cross-enterprise
multidomain governance
Well-established cross-
enterprise governance
Organization No data stewards
Developing the culture of
data stewardship
Potentially strong team,
but limited in scope
Centralization or
federation of data
steward groups
Well-established cross-
enterprise stewardship
Processes
Silo-based – applications
or functions
Starting to think in terms
of the data life cycle
Focus on data life cycle at
domain level
Best practice shared
across the enterprise
Continue to optimize the
life cycle
Technology
Few or no data quality
tools
Data quality tools, but no
MDM solutions
One or more
disconnected MDM
solutions
Integrated and consistent
set of capabilities
Integrated and consistent
set of capabilities
Metrics
No metrics scheme for
master data
Starting to develop DQ
metrics
Successful use of metrics
at the domain level
Using metrics to measure
success cross-domain
Basis of management and
investment
Current state
28. Place Master Data at Use
28
Using Place Master Data to Drive Operational Activities
2
4
1
3
5
Reporting & Analysis
Providing CROSSMARK, CPG’s and Retailers
direct access to view activities at retail.
Finance
GL, Financial Planning & Analysis
tools for managing the profitability of
our business.
Billing & Payroll
Internal system that collects work data and
mileage, bills clients, and compensates
employees.
Workforce Management
Tool that optimizes the work and collects
details for 35,000 field workers.
Retailer Solution
CROSSMARK system used by retailers
used to manage work within their
stores.
29. 29
Our MDM Path Forward
Start
Q4 2015
Go Live
CROSSMARK has 800,000 places in
production with most systems cutover to the
new source.
Q3 2015
CROSSMARK has been live on Dell Boomi MDM since August 2015.
30. 30
Services Domain
CROSSMARK will be deploying Services
domain using similar mastering architecture.
Q1 2016
Q2 2016
Products Domain
Implementing consumer products domain
with ordering and consumer product
attributes.
Analytic MDM Pilot
Integrating Dell Boomi MDM with Big Data
efforts to unlock additional value, including
generation of dynamic polygon places and
hyperlocal targeting.
31. 31
To be Continued…..
Additional Domains
We have identified 9 domains in total for mastering,
including Employees, Clients, Customers, Products,
Services, Places and more.
32. 32
Key Learnings
Far faster development than we expected
We had working prototypes of the end-to-end process in place within 1 month.
Algorithm development was the most complex, with 2 months of iterations.
Performance
Most of our performance issues have been specific to the Salesforce platform. We’re
still learning how to optimize the entire environment.
API Integration is very powerful
The ability to integrate to external API providers is extremely powerful. We think it
will enable the next waves of analytically derived master data.
A flexible platform with a few opportunities for growth
We are not constrained by standard definitions of domains. Where there are
limitations we were able to overcome by building them within the Force.com
authoring application.
Dell Has Listened
We’re working with Dell on continued development in data stewardship and profiling
capabilities. They have quickly embraced our suggestions and integrated them in
their roadmap
This is the 5 day precipitation map of the Southeast United States. Hurricane Joaquin has dropped heavy rains all over the region, including 20” in Charleston on Saturday night alone.
I’m grateful to be off the trail, dry and with you all today.
A place is a point on a map, part of a path, boundary, an elevation and more.
Limiting your definition of a place to traditional operational views constrains the ultimate value you get from master data.
CROSSMARK did not presently have a Master Data Management capability in place.
Data was authored in multiple locations, but not mastered.
Conflicting standards existed for domains
Data governance was led exclusively through IT
Data quality tools were primitive
There were no metrics around data quality
There were 4 separate highly manual processes for creation of Place data.
Authoring and mastering of Place data took up to 4 weeks.
CROSSMARK did not presently have a Master Data Management capability in place.
Data was authored in multiple locations, but not mastered.
Conflicting standards existed for domains
Data governance was led exclusively through IT
Data quality tools were primitive
There were no metrics around data quality
High upfront costs required of traditional MDM solutions make it challenging to financially justify efforts.
Also becomes challenging to scale out to serve greater users of the solution.
High upfront costs required of traditional MDM solutions make it challenging to financially justify efforts.
Also becomes challenging to scale out to serve greater users of the solution.
Speed time to value with crowd-sourced data modeling. Domain data modeling is crucial but time-consuming. Save hours of trial-and-error modeling with Boomi MDM. It inventories all data models in use, referencing the collective intelligence of its multi-tenant resources, and suggests an optimal solution which you can then customize as needed.
Leverage the power of Dell Boomi MDM cloud managed approach to span organizational boundaries across departments, domains and applications.
Leverage bidirectional, real-time data flows to unleash the potential of your constantly changing data. Dell Boomi MDM captures, aggregates, and publishes ongoing changes to the applications that run your business.
Unify master data across domains to eliminate MDM silos. Unlike domain-specific master data solutions, Boomi MDM is domain-agnostic, supporting data from any domain, whether customer, product, location or other data.