living in a multiorg world
TRANSCRIPT
Living in a MultiOrg World
Objectives of this sessionDiscuss the reality of why we are hereConsiderations around multi/single org Living in a multi-org world
modelspro’s cons of eachactions in assessing
Discussion/experience share
Assumptions – why you’re here.
You have multiple instances with common business processYou have multiple instances each used for totally different thingsYou have multiple instances, and have no idea what is common or differentYou hear a lot about companies slipping into this and want to prepare
Other items
LANGUAGE
GEO
COMPLIANCE
LEGACY CUSTOMIZATIONS
LEGAL
Why are you Multi?Salesforce seed strategy
Why are you Multi?Salesforce seed strategyMerges and acquisition
Why are you Multi?Salesforce seed strategyMerges and acquisition Intentional strategy
Variance in business processFundamental separations in functions/business
Why are you Multi?Salesforce seed strategyMerges and acquisition Intentional strategyLegacy technical constraints
Why are you Multi?Salesforce seed strategyMerges and acquisition Intentional strategyLegacy technical constraintsNo idea
Before we get startedConsider what SFDC is to you.
Customer Relationship ManagementCommon understanding of customerOne version of truthConsolidate the perspectives
CUSTOMER
Marketing
Sales
Product
Services
Support
People
Operations
Finance
Before we get started• Consider what SFDC is to you.
• Operational Platform – capture X functional process
Linking the organization internally. Collaboration. Place I go to find/execute workCa
mpa
ign
Lead Oppo
rtuni
ty Orde
r
Proje
ct
Purc
hase
Or
der
Reso
urce
Effor
t
Log
Invoic
e
Case
Before we get started• Consider what SFDC is to you.
• Development platform (App Cloud) – host custom applications
What have 2000 projects taught us
No matter how unique the business. Core processes can be normalized.
Humans are hoarders. In life, data and metadata.
Users operate in short time windows. It’s not that they forget. It’s just not worth it to them.
Systems drive efficiency at the company level. This can feel like inefficiency at the user level.
If we had a fresh start…If we were to start again with all that we know
What would we do?Establish the ideal state.Process into requirementsMacro needs
Don’t default to single org. What is reasonable? Single view of customer.
SecurityCommon business processAnalytics
TCO
But we don’t have a fresh start Jenkins!
What do the sponsors tell you?
How are the systems being used?Conduct user / admin interviews
User storiesDay in the life interviewsBusiness process reviewRole/ Business Function/Process/Statement/Validation
What does the data tell you?Conduct detailed data audit of orgs
What objects are in use.What fieldsWhat date stamps/ageWhen were they usedWhat reports are being usedWhat data is being entered
What does the data tell you.
What does the meta data tell you?Conduct detailed audit of the event logs– user interaction logs
WhoClicked whatWhen
Large data volumes – warning Analytics overlay requiredWhat does the data tell you.
Simple summarythink forensic analysis
Two philosophies
SPEED QUALITY
COST
NATURAL SELECTION DOMINATES BUSINESS ENGINEERING
One framework
NATURAL SELECTION DOMINATES
BUSINESS ENGINEERING
How do we get the best of both?
One framework
NATURAL SELECTION DOMINATES
BUSINESS ENGINEERING
How do we get the best of both?
Intentional experimentsRollback strategySharing intentionsLearning repositoryDocumented resultsLearnings/Go forward
Successful experimentsNon-negotiable elementsTrue governance
What are the options
SEPARATED – BI SummaryAllow for true agility at the “org” level. Provide a set of services for integration and require BU’s to maintain their dictionaries and provide data for rollup/executive reporting
++ +++
BI Summary Layer
Consider ERP integrations?
Cover up. Why not fix the source.
PARTIAL INTEGRATEDAKA matrix strategy, each org subscribes to object and data definition standards, communicating projects and data dictionaries, allowing for point to point integrations between orgs to drive business process.
+
+
S2
+
+
S1
+
Samples shown for S1 and S2 only.
Integrations become a nightmare. Anonymous captors.
MASTER DATABusiness units or even departments will maintain their own instances, but be forced to tie all common data to a separate consolidated environment. Could be MDM or a Master Data Instance of Salesforce.
MASTER
Account MasterContact MasterProduct MasterPrice MasterReference Master
Leverages Data.com use DUNS as a common key
Does not consider business process.
MASTER META DATABusiness units or even departments will maintain their own instances, but be forced to maintain common metadata definitions, data structures, governance standards. Common stack for all orgs. Tip- common sandbox for ALL previous to rollout will drive governance.
MASTERMETA
Account Base ObjectContact Base ObjectProduct Base ObjectWorkflowTriggers3rd party apps
Does not consider common data. No 360.
CLONEDAll orgs are “birthed” from a common – evolving master implementation. While they will be expected to diverge over time, the base is consistent. In cases where base functionality has to be changed, the parent is changed and change has to be pushed down. Similar to Master Meta, only less rigid.
++++
Source Org
Internal Appexchang
e
What about the data?
Great. Nothing works. Awesome
So what is the right approachWhat should be considered table-stakes
Regardless of the Multi Approach Common data definitionsMaster Meta-Data
Maintain a data dictionaryEstablish a x-functional group to maintain with representationForce standards on field labelling
Regardless of the Multi Approach Common reference dataCommon Master Data
Who at any one time “owns” the attribute (system/person)Do they understand/know thisStandard integration “services/apps” Canonical view – pure standards at the central sourceMethod of nominating the “golden record” attributes
Regardless of the Multi Approach Common governance/stewardshipCommon maintenance
What are the standards X all orgsHow are they being maintainedHow are we assessing them – metadata discoveryWhat are the functional roles of the people to manage
Regardless of the Multi Approach Common technology stack where possibleCloned
Avoid additional complexity introduced through new appsStandardize at the app level
Regardless of the Multi Approach Common rollout/QA and metadata managementCommon methodology
Sandbox and rollout strategyQA processesProject documentation and requirements managementCleanup/retirement of data/app strategy
Regardless of the Multi Approach Reflect performance from the top of the hierarchyCommon benchmarking
Provide installable packages or access to enterprise reportsDisplay differences and performance gapsShow orgs what is seen at the exec level – drive accountability
If you cannot recall anything else – next slide counts
TAKEAWAYSSize of business does not indicate Multi/Single fit
Common business process trumps common system
Common apps across all orgs – allow intentional experiments
Common reference data – dictionary
Common org management (governance/training/qa/deployment/sb)
Discussion startersHow will the new capabilities around Lighting affect this?What is the worst thing we can do in running our SFDC environments?Where do I start?How do we find the balance between democratic/dictator?What are the indicators of bad things ahead?
Thanks!