Top Essential Actions for Governing Power BI - Melissa Coates Coates Data Strategies - Coates Data ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Top Essential Actions for Governing Power BI Melissa Coates Coates Data Strategies March 25, 2021 Download slides: CoatesDS.com/Presentations
Melissa Coates Data architect | Technical trainer | Consultant Specialist in Power BI governance & administration Microsoft Data Platform MVP Owner of Coates Data Strategies @SQLChick | @CoatesDS Creator of Power BI Deployment & Governance Workshop Power BI Data Culture & Architecture Security & Adoption & Governance Center of Excellence Decisions Data Administration Protection
Top Essential Actions for Governing Power BI Agenda Data Governance Top 10 Final Overview Actions Q&A Things You Can Do Right Now
Top Essential Actions for Governing Power BI 1. Deeply understand the centralized & decentralized BI approaches in use 2. Focus on data trustworthiness 3. Define your guiding principles 4. Implement short, clear, data policies for the most important items 5. Set accountability expectations with clear roles & responsibilities 6. Invest in a data-centric audit & protection strategy 7. Be transparent about system administration decisions 8. Analyze & act on data for adoption, auditing & training 9. Foster your internal community, support, training & resources 10. Actively cultivate your data culture
Where Are You At? I’m assuming that… You are using Power BI currently & have experienced some successes You are familiar with the basics of deploying Power BI content You are aware there are areas of improvement with how you oversee Power BI
When Does Data Governance Come In? Source: https://docs.microsoft.com/en-us/power-bi/guidance/powerbi-migration-overview
Two Paths Building out the governance groundwork Handling the most pressing analytical needs Focus on continual, iterative, progress on both paths.
What Do We Mean by Data Governance? Data Data Platform & System Strategy Architecture Administration Data Policies, Risk Management Standards & Processes & Compliance Data Master Data Governance Business Process Management Management Change Data Internal Management Quality Audit Data Data Data Business Applications Databases Integration Lakes Warehouse Intelligence
Defining Data Governance “ A system of decision rights and accountabilities for information-related processes, executed according to agreed-upon models which describe who can take what actions with what information, and when, under what circumstances, using what methods. ” Data Governance Institute
Data Is Governed When It Is… Trustworthy Secured The right data The right people Managed & Documented & Audited Understood The right way The right reason →More info: CoatesDS.com/blog/what-is-governed-data
Why Do Users Care About Data Governance? Is the data I Who owns this want available? data? Can I trust this data that I am using? What is the Am I in system of record compliance for for this data? what I want to do?
Effective governance can dramatically improve user experience –if– user enablement is a primary objective (within the requisite guardrails)
Tip 1: Deeply understand the centralized & decentralized BI approaches in use
Business Intelligence Approaches Top- Blended Bottom- Down Up Enterprise Managed Business-Led BI Self-Service BI Self-Service BI Centralized Decentralized Data Central IT/BI/COE Business authors ownership : Report Business authors Central IT/BI/COE ownership:
Business Intelligence Approaches Top- Blended Bottom- Down Up Enterprise Managed Business-Led BI Self-Service BI Self-Service BI Centralized Decentralized Most organizations have all 3 approaches actively happening, whether you’re aware of it or not
Things Assess your current state: know what’s you can working & what’s not. do right now Make a specific plan for iterative, continual, improvement.
Things Determine which groups can – and should – you can manage their own content. do right now What exactly does that look like?
Things Decide how to handle co-ownership you can scenarios, when data & reports are owned do right now or managed by multiple teams.
Tip 2: Focus on data trustworthiness
Data Is Governed When It Is… Trustworthy content is a key pillar of governance: Trustworthy Secured The right data The right people Managed & Documented & Audited Understood The right way The right reason →More info: CoatesDS.com/blog/what-is-governed-data
What Do We Mean by Trustworthy Content? Content quality is assured Approved, common definitions and KPIs are used Owner, SME, and steward responsibilities are clear Data provenance & lineage are clear Data duplication is minimized with reuse Data & load processes are managed & secured Awareness of content which Clear expectations about is discoverable & findable responsible data use →More info: Improving Trustworthiness of Power BI Content at CoatesDS.com/presentations
Decoupling Data & Reports A shared dataset is intended for reuse across multiple reports Also known as: Shared dataset: Sales Data Golden dataset & thin Live Connection Analyze In Excel reports YTD Sales Product Channel Hub & spoke Sales Sales by method Reports: Revenue Analysis Product →More info: CoatesDS.com/blog/5-tips-for-separating-power-bi-datasets-and-reports
“Discipline at the core & flexibility at the edge ” https://docs.microsoft.com/en-us/power-bi/guidance/center-of-excellence- microsoft-business-intelligence-transformation#discipline-at-the-core
Decoupling Data Preparation from the Model Dataflows are intended for reuse of data preparation across datasets Power Query Customer Master Dataflow(s) Online Dataflow: Customer Sales Shared Sales Data History Programs Datasets: Channel Product Sales YTD Sales Customer Sales Sales Program Reports: Revenue Retention Summary Analysis Effectiveness
Taking Things Farther with Composite Models Data sources DirectQuery for Power BI dataset Additional Additional data sources data sources Live Connection →More info: https://powerbi.microsoft.com/en-us/blog/directquery-for-power-bi-datasets-and- azure-analysis-services-preview/
What Can Be Endorsed? Certified The author wants to signify the content is trustworthy Promoted The author wants to amplify that the content is available Data: Reports: Set of Reports Power BI Report & Dashboards: Dataset App Dataflow Paginated Report →More info: Improving Trustworthiness of Power BI Content at CoatesDS.com/presentations →More info: https://docs.microsoft.com/en-us/power-bi/collaborate-share/service-endorsement-overview
Things Use shared datasets when data is you can useful for > 1 analysis to reduce risk do right of inconsistencies, and lessen now compliance concerns.
Things Decouple your dataset creation & report you can creation work. do right now Teach content authors new habits for how to work so they can take advantage of data reusability.
Things Use certified & promoted endorsements. you can do right Create a specific review process for now certified data to ensure that certification = trustworthiness.
Tip 3: Define your guiding principles
What is the Right Balance? Risk Management Enablement Compliance Flexibility Security Data Sharing Privacy Standardization Data Integrity
Finding the Right Governance Focus Offense strategy Defense strategy Revenue growth, Risk reduction, cost reduction, govt & industry regulations, decision-making data privacy, improvements compliance laws Some tolerance for Greater investment “multiple versions of in “single version of the truth” the truth” Some chaos accepted to Stability over chaos encourage speed & is highly valued innovation
Example guiding principle: Everyone is a data steward Responsibility for the governed and secure delivery of data rests equally among content owners, subject matter experts, and COE members.
Example guiding principle: Our most critical data elements are most actively managed during the information lifecycle Not all data has the same value. Our critical data elements are managed and monitored to a greater degree.
Things Figure out if you are playing offense or you can defense when it comes to governance do right decisions. now In other words, figure out your why.
Things Get consensus on your top few guiding you can principles with your executive sponsor – do right before any data policies are crafted. now
Tip 4: Implement short, clear, data policies for the most important items
Inputs & Outputs of a Data Policy Organizational Objective High-level strategic vision Guiding Principle A statement of belief Policy What needs to be done and why Support Process or Controls & details for how Procedure a policy is implemented Auditing & Assessments A measurable level Standard of attainment
A Simplified Example Improve decision-making throughout Organizational Objective the organization Critical data elements (CDEs) are Guiding Principle identified, managed, and audited Datasets containing CDEs have data definitions, Policy sensitivity labels & are certified Process or Data certification process Procedure Data definition procedure 100% of certified datasets have an owner Standard 85% of certified datasets are reviewed 2x/yr
Data Certification Process At a minimum, include: Data Certification Report Certification Data source & lineage review Report content review Data model tech review Slightly lighter data review: Security review Data source & lineage review Data model tech review Data accuracy validation Security review Documentation review Data accuracy validation Documentation review
Things Determine what existing decisions you you can should align with & reuse from your data do right governance board, change management now board, etc. What differs for a self-service BI platform like Power BI?
Things Make sure the data policies can be easily you can searched & referenced by your internal do right Power BI users. now Use a consistent organizational approach so users aren’t surprised.
Things Pick your most important thing to start with. you can Examples: do right now • Data certification process • Data ownership policy • Data handling policy (aka data usage & distribution policy) (aka data classification policy)
Tip 5: Set accountability expectations with clear roles & responsibilities
Business Intelligence Approaches Top- Blended Bottom- Down Up Enterprise Managed Business-Led BI Self-Service BI Self-Service BI Centralized Decentralized Your approaches in use, by whom, and how, will have a big effect how the roles & responsibilities will be defined
Define Roles & Responsibilities Does everyone IT & Power BI Ancillary understand: Roles Roles • What are normal Data Governance expectations? Roles BI & • What is required? Business Developer User Roles • What is prohibited? Roles • Where is flexibility allowed?
Content owners & subject matter experts in the business units form the foundation of a well-governed system
Governance Checks & Balances Executive Leadership Level 3: Audit + Compliance Internal audit and external audit Level 2: Data Governance Team + Center of Excellence Define policies, standards, and procedures Support business units with questions, issues, and training Level 1: Business Units + Ancillary Teams Content authoring, publishing, and security Data management, data stewardship and validation The foundation of a well-governed system
Things Get clarity on who your Power BI you can executive sponsor is, and how much do right authority this person has for now implementing policies throughout the organization.
Things Define your relevant roles & responsibilities. you can do right Include them within formal HR job now descriptions whenever practical.
Things Communicate & teach content creators you can what their role is related to safeguarding do right data & the responsible use of data. now
Tip 6: Invest in a data-centric audit & protection strategy
There’s a Lot of Power BI Content Everywhere Source data in Content published Content exported databases, apps to the Power BI from the Power BI & data lakes cloud service cloud service Datasets PBIX files exported Files stored in file servers, OneDrive, Exports to SharePoint, laptops Dataflows PowerPoint & PDF Power BI Desktop & Data exports Paginated Rpt files Reports E-mail subscription Excel workbooks images & Dashboards attachments Source data files Content Workbooks embedded in External tools files other services
Data-Centric Audit & Protection (DCAP) DCAP strategy: an approach to information protection that focuses on protecting the data itself Know Protect Govern Identify where data is Define data protection Audit the use of sensitive located & classify it by policies for how & where data, verify changes to sensitivity level data is managed to permissions & when user prevent data loss behavior is not typical
Sensitivity Labels A Know > Protect > Govern strategy begins with the classification of content
Sensitivity Labels Drive Policies & Recommendations Sensitivity Labels User Data Storage Data Loss Preferred Content Access Allowed Prevention Distribution Internal Online Locate App, workspace, sensitive info sharing External Offline/download Conditional Block access Email Encryption access Anonymization Unmanaged devices **Many things can be automated; others are educational
Data Catalog A Know > Protect > Govern strategy includes a data catalog as a key component. You can’t effectively govern data if you don’t know what you have.
Things Establish a classification taxonomy. you can do right Define sensitivity labels in Microsoft 365 now which are: • Simple & few (4-7 in total) • Intuitive & can be localized • Hierarchical • Generic enough to last a long time • Broadly applicable across the organization
Things When creating a data handling policy, focus you can on exactly what you need to prevent for do right each sensitivity label. now Automate what you can with data loss prevention (DLP) policies in Microsoft Cloud App Security and/or Microsoft 365 Compliance Center.
Things Look into the Azure Purview public preview. you can • Self service users: catalog for search, lineage, do right now glossary, classification • Administrators: data map of what data exists, where, lineage & who uses it • Compliance: where sensitive data is located & who uses it
Tip 7: Be transparent about system administration decisions
Sources of Confusion & Irritation for Users Why can’t I start Why can’t I a Pro trial? Why can’t I create a export data? workspace? Why can’t I certify a dataset? Why can’t I Why can’t I use share to this custom Teams? Why can’t I visual? install a gateway?
Many Settings Affect the User Experience Tenant Settings Gateways & Data Sources Premium & PPU Settings Admin Organizational Visuals Licensing & Portal Free Trials Azure Connections Workspaces Software Featured Content Installations →More info: What You Need to Know to Administer Power BI at CoatesDS.com/presentations
What we want to avoid is a Power BI administrator deciding *solely on their own* what should & shouldn’t be allowed
Things Review each tenant setting to ensure it is you can purposeful for: do right • Behaviors you want to encourage now • What you need to deny
Things Document settings for the broader internal you can Power BI community, including which group do right applies to each setting. now This documentation should be easily located in your internal Power BI community “hub.”
Things Review who is a Power BI administrator. you can Reduce the number of permanent do right administrators if you have more than a few. now If there are people who legitimately need access occasionally, implement that with Azure Active Directory PIM (Privileged Access Management).
Tip 8: Analyze & act on data for adoption, auditing & training
Why Activity Analysis is Critical Critical content What content is most frequently used? Is it adequately supported? Change tracking What changes occur, when, and by whom? Internal and external auditing Are you able to satisfy requests from auditors?
Why Activity Analysis is Critical Monitoring adoption efforts Can we analyze not only usage stats, but that the system is being used consistently and optimally/as it was intended? Data trustworthiness levels How many certified vs. non-certified datasets? How many datasets support > 1 report? License usage Who is (and is not) using Power BI, at what frequency?
Why Activity Analysis is Critical Understanding usage patterns How are users *really* using Power BI? Finding training opportunities Is training actively made available to new users, or to encourage specific behaviors? Suspicious usage patterns Are any concerning activities occurring?
End-To-End Power BI Auditing Solution M365 Power BI MSFT Graph Gateway Audit Log REST APIs REST APIs Servers Power BI Workspace Gateways Apps, User Info & Group Power BI Power BI Gateway Activity Inventory & & Data Capacities Service Prin Membership Licenses Admins Logs Events Security Sources etc. PowerShell Scripts Optional: Accessed Data Lake, Power BI Analytical by users NoSQL, or Auditing Datasets (RLS) File System Database & Reports Original raw data Historical transactions & Prepared data for adoption, JSON files point-in-time snapshots security & auditing Accessed only by →More info: What You Can Learn from the Power BI Activity Log and auditors & REST APIs at CoatesDS.com/presentations administrators
Things Begin retrieving all of the raw data from you can the Power BI Activity Log if you haven’t do right already. now Augment it with data from the various Power BI APIs.
Things Start learning from the Activity Log data you can how Power BI is *really* being used by do right the user population, and take action now accordingly.
Tip 9: Foster your internal community support, training & resources
Internal Power BI Community One of the biggest success factors is nurturing the internal Power BI community with: Support & Mentoring Knowledge Sharing Resources Training Documentation Communication
One of the best ways to decrease risk is to increase knowledge
Things Set up an internal Power BI “hub” for you can documentation, links, resources, training do right info & announcements. now Choose a convenient location where users frequently work.
Things Update the tenant settings so that the you can help menus in the Power BI Service direct do right users to your Power BI “hub.” now
Tip 10: Actively cultivate your data culture
“ A data culture promotes and encourages data-driven decisions by more stakeholders in more parts of the organization ” Matthew Roche – Data Culture Series on SSBIPolar.com
Goals of a Data Culture Increased: 1. Active and consistent reliance on data for decision-making 2. Effective use of data by more stakeholders Reduced: 1. Reliance on tribal knowledge 2. Hunches & gut decisions from the HIPPO (Highest Paid Person’s Opinion)
Do You Need a Center of Excellence? Who will nurture the internal Power BI community and advance the data culture on an ongoing basis? Who will mentor & build your network of Power BI champions & experts? Who will define & manage the governance model? Who will have a cross-departmental view into what’s happening across the organization?
Things Decide who will nurture the internal you can Power BI community. do right now Make sure they have time, focus, resources, funding & motivation.
Things Create & actively promote guidelines & you can policies which encourage a balance of do right control & empowerment. now
Things Define specific, measurable, goals for the you can Power BI community to monitor progress do right & retain support for it. now
Wrap-Up, Q&A, Links to More Info
Top Essential Actions for Governing Power BI 1. Deeply understand the centralized & decentralized BI approaches in use 2. Focus on data trustworthiness 3. Define your guiding principles 4. Implement short, clear, data policies for the most important items 5. Set accountability expectations with clear roles & responsibilities 6. Invest in a data-centric audit & protection strategy 7. Be transparent about system administration decisions 8. Analyze & act on data for adoption, auditing & training 9. Foster your internal community, support, training & resources 10. Actively cultivate your data culture
Top Essential Actions for Governing Power BI Two paths: Building out the governance groundwork Handling the most pressing analytical needs Focus on continual, iterative, progress on both paths.
Q&A
More Resources from Melissa Slides: Diagrams: CoatesDS.com/Presentations CoatesDS.com/Diagrams YouTube: Blog: YouTube.com/CoatesDataStrategies CoatesDS.com/Blog-Posts Power BI Governance Training: Twitter: CoatesDS.com/Workshop @SQLChick | @CoatesDS
Additional Resources Power BI Whitepapers https://docs.microsoft.com/en-us/power-bi/whitepapers Matthew Roche’s Data Culture Series https://ssbipolar.com/building-a-data-culture/ Power BI Adoption Framework https://github.com/pbiaf/powerbiadoption Power Platform Adoption Framework https://github.com/PowerPlatformAF/PowerPlatformAF
You can also read