How to develop Data_Strategy_Webinar_Final.pptx

DATA STRATEGY how to develope a effective data strategy

1. Secrets to Creating an Effective Data Strategy Tips from Industry Insiders By Wayne Eckerson June 2024

2. Speakers

3. What is a data strategy? Why do you need a data strategy? How do you implement a data strategy? Agenda

4. Data Strategy Resources

5. I. What is a data strategy?

6. Data Strategy Components “Road Markers” “Operating Model” “Data Architecture” “Data Governance” “Change Management”

7. Data Strategy = Business Plan Business Plan Mission, Vision, Values Goals and objectives Critical success factors Key stakeholders Target customers Success metrics Roadmap and priorities Initiatives and tasks Budget and plans Mike Masciandaro Former Head of BI Dow Chemical RUN DATA & ANALYTICS AS A BUSINESS

8. II. Why do you need a data strategy?

9. “If you don’t know where you’re going, any road will get you there.” Why a Data Strategy? 1. Get out of a quagmire…. Data silos Data bottlenecks Data inconsistency Poor data quality Low adoption 2. Modernize your data environment Improve performance Reduce overhead costs Address technical debt Prioritize data initiatives 3. Better serve the business Align with strategic initiatives Increase agility Reduce risk Improve customer satisfaction

10. New data-driven CEO or CIO Business transformation Modernize data architecture Resolve disagreement Validate existing strategy Create actionable plan Political Issues CYA Drivers for Creating a Data Strategy

11. Communicate with executives Show how data helps the business Get buy in - secure funding Communicate with department heads Show how data supports their area Get buy in – avoid sabotage Communicate with your team Get everyone on the same page Gain direction and focus Benefits of a Data Strategy A data strategy is a communications vehicle

12. Annually As part of annual strategy & budget presentation After a major change Merger or acquisition New CEO or CIO or CDO New corporate strategy Technology paradigm shift When to refresh your strategy?

13. III. How do you implement a data strategy?

14. Current State Assessment DATA AND GOVERNANCE FOUNDATION A B B C XXXX is struggling to deliver ad hoc analysis and advanced analytics because it has a weak foundation in data management and data governance. Client Sample

15. High-Level Data Strategy 02 04 Activate Data Governance Optimize the Operating Model Modernize Data Architecture 03 Enhance Data Literacy Improve Self-Service 05 Initiatives Elevate strategic importance of data Consolidate data & analytics teams Allocate time & resources to build data foundation Fill in staffing gaps Initiatives Execute data lakehouse design Identify MVP use cases Implement master data management Establish architecture oversight Migration and conversion strategy Initiatives Establish Data Governance Foundation with Dashboard Activate Data Governance Mature Data Management & Integrate in all Cloud Projects Initiatives Define framework Execute framework Monitor and evolve Initiatives Classify users Implement data catalog & glossary Implement analytics workbench Rationalize BI tools and functionality 01 CURRENT STATE FUTURE STATE Apply Change Management Practices 06 Client Sample

16. Roadmap and Budget FY 2021 FY 2022 FY 2023 Theme Deploy Consolidate Reorganize Data Architecture Deploy new data prep, cloud DW and data integration tools Annual Costs Data prep = $ Cloud data platform = $ Data integration = $ Cloud data platform consultant = $ Data platform architect = $ Data Lineage = $ Data prep = $ Cloud data platform = $ Data integration = $ Head of Data Analytics (CDO) = $ Data engineer = $ New technologies = $ New FTEs + 2 temporary consultants +2 FTEs +4 FTEs Data Analytics Address high priority use cases Analyze Canvas usage data Launch SWAT Teams to rapidly build capabilities for departments Operating Model DW consultants for on-boarding, administration, development Data scientists, analytics services, data literacy staff Head of Data Analytics (CDO), data engineer Launch Enterprise Data Analytics team and program Consolidate data and organizational silos Est. Expenditures Data Governance Establish DG Council and functional coverage Expand new platform & tools Expand new platform & tools $37k OpEx $705k CapEx $605k OpEx $1,084k OpEx Data prep = $ Cloud data platform = $ Data integration =$ Head of Data Analytics = $ Data engineer = $ 2 data scientists = $ Analytics services manager = $ Data literacy coordinator = $ New technologies = $ FY 2024 Extend Implement advanced Analytics & Data Literacy Use case driven AI projects) Continuous monitoring Create Data Literacy Curriculum Data Science Platform Business Monitoring System 0 FTEs AutoML tool - $ Business monitoring - $ Client Sample

17. Executive Summary: The Ask Poor quality data Limited data access Minimal data governance Inflexible systems Trustworthy, accessible data A more agile organization Steady stream of data products Better decision making CURRENT STATE JUNE 2023 FUTURE STATE JUNE 2026 BUDGET Data Strategy High costs, risks, and staff dissatisfaction due to poor quality data and systems. Unleash time and energy of staff by eliminating “data drudgery”. We need: 1. Operating dollars: Year 1: $1.145M Year 2: $740k Year 3: $790k 2. Organizational commitment to data 3. Staff time and resources Client Sample

18. 2. Standardize Data Architecture Current State Future State Description: Data Architecture as data models that align with the consumer’s need, and that defines the flow of data to populate these models, has degraded and needs refactoring. The Data Architecture should be easy to work with, fast to update, and quick to enhance. XXXX has created a spider web of data that contradicts this, and untangling the web is necessary for future success. Challenges. Multi-directional data flows provide immediate relief but make maintenance, future enhancements, and other support more difficult Data zones not defined with no clear location for users to access data Data is not modeled and is provided to users based on their level of expertise and utilization patterns The Data Warehouse is difficult to use and not modeled according to best practices Master data is not formally implemented Data is slow to move and transform Data Warehouse does not have elasticity needed in a modern environment (due to it being traditionally licensed on-prem) Business users Have a common well-known location to access their data that is precisely curated for their needs, and easy to understand and use Most of their solutions are achieved with with no coding, and be accomplished via ‘drag and drop’ user interfaces Faster implementations for MVPs with cloud technology Data analysts & data scientists : Data analysts and data scientists leverage the data lake for discovery, machine learning, and statistical analysis Have sandboxes for development, making data private until promoted for enterprise visibility when they can onboard and publish approved data to production Utilize the modern features of the cloud for discovery such as scaling up and out, parallelism, and unlimited storage Implement CI/CD/CT and participate in governance process Data engineers : Use the Data Lake as the initial landing zone for all data Create data refinements and transformations using the data lake as the source, or a data set that originated from the data lake Use tools “built for the cloud” Initiatives To Close the Gap Implement a data services architecture Created refined zones of data Re-engineer DW and DL data flows Create DW cloud migration plan Establish oversight for data architecture Extend MDM capabilities Client Sample

19. Initiative #2b: Implement a data services architecture 1. Be agile in creating new data services framework that establish a foundation to grow without re-engineering 2. Start implementation with an MVP that will succeed in a 8-12 week period 3. Use cloud technology for agility and easy expansion 4. Ensure architects have strong grasp of modeling techniques and choices Change Management Strategy Tasks Description Capabilities Required (People, Process, Tools) Duration/ Milestones/ Dependencies Costs 1. Purchase and deploy software to use for building data services Provide a platform that would allow XXXX to build a data services framework, while also being a data virtualization engine that would add immediate benefit. A data lakehouse product utilizing data lake storage, is a logical fit for XXXX to build its data service on. Data Architects and engineers would need to be dedicated to this effort to stand it up initially. Early Year 1 2 weeks 2 FTE Lakehouse Product 2. Deploy data services software as the virtualization engine over the enterprise This will reduce data copying and provide immediate relief to users experiencing delays due to lack of access. Data Architects working with Enterprise Architects Early Year 1 6 weeks 2 FTE Lakehouse Product 3. Engineer data access for all data consumption Every type of usage at XXXX should have the appropriate level of access, including reporting, analytics, applications, mobile, and others. Data Engineers build interfaces and refined data sets as needed Year 1 12 weeks 2 FTE Lakehouse Product 4 . U se the data lake as the place to access all raw data and refined data sets Transition reporting, BI, analytics, and application consumption to this data framework for consumption. Data Architect, Data Engineers, Project Managers, Data Consumers Year 1 12 months 2 FTE Lakehouse Product Goal: Create an architecture providing XXXX consumers standardized access to enterprise data, in real-time if needed, through a data services “backbone” Client Sample