Dashboard Design
Part of Selecting BI and dashboard software
Comparing self-service BI with centrally managed reporting
Compare who builds reports, owns shared measures and approves distribution when selecting a BI reporting model.
When selecting BI software, compare who creates reports and who controls shared measures. Centrally managed reporting assigns production and release to a designated team.
Self-service BI lets authorised business users investigate or create views. A blended arrangement can keep approved measures under one owner while report authors answer local questions.
Separate the activities
Decide who may view a report, build a new view from an approved dataset, change a shared calculation and publish to a wider audience. These are different tasks, even when one person performs several of them.
| Selection question | Central reporting emphasis | Self-service emphasis |
|---|---|---|
| Who builds a new view? | A reporting team handles requests. | An authorised business creator builds from available data. |
| Who changes a common measure? | A designated model or metric owner. | An accountable owner is still needed when reports share it. |
| How does a result reach wider readers? | A defined release process. | A creator may draft locally before review. |
| What must the tool support? | Controlled publication and maintenance. | Finding suitable data, exploration and clear usage limits. |
These are operating choices. Product labels alone do not show whether the proposed staff and permissions can carry them out.
Centralised Reporting vs Self-Service BI: Key Operational Differences
- Who builds a new view?
- A reporting team handles requests.
- Who builds a new view?
- An authorised business creator builds from available data.
- Who changes a common measure?
- A designated model or metric owner.
- Who changes a common measure?
- An accountable owner is still needed when reports share it.
- How does a result reach wider readers?
- A defined release process.
- How does a result reach wider readers?
- A creator may draft locally before review.
- What must the tool support?
- Controlled publication and maintenance.
- What must the tool support?
- Finding suitable data, exploration and clear usage limits.
Find the constraint that matters
Central delivery may suit a recurring, consequential figure used across departments, or a team with few model maintainers. Check whether the reporting team can handle changes at the required pace; do not assume a delay without observing the request process.
Self-service may suit trained users who need frequent investigations within a known dataset. Its risk is inconsistent choices about periods, filters or measures.
Documented definitions and a maintained dataset help, but creators still need to recognise questions outside that dataset's scope.
Microsoft describes a managed self-service Power BI pattern: a central team maintains semantic models and departmental creators build reports. It is a product-specific example, not a prescribed structure.
Power BI sharing also depends on licences and capacity.
In Tableau Cloud, licences and site roles limit capabilities, and content permissions govern particular actions. Looker lets users query analyst-defined LookML models in Explore when they have the relevant permissions.
Check those controls against your proposed division of work.
Licensing and Access Controls in BI Platforms (Australia)
- Power BI Licence Requirement
- Pro or Premium Per User licence for publishing and sharing content
- Tableau Cloud Site Role
- Determines user permissions; e.g., 'Site Administrator' vs 'Explorer'
- Looker User Permissions
- Requires specific roles to query LookML models in Explore
Use a role-based selection trial
Give a central author and a proposed business creator the same approved measure and fixed-period task. Ask each to produce a breakdown. Have a viewer find the result and explain its cut-off. Then request a change to the shared measure: identify who can propose, approve, implement and communicate it.
Record assistance needed, effective permissions and whether a draft can stay limited to its intended audience. Include a case in which the business creator must be unable to edit the shared measure. These are proposed acceptance checks, not results from a completed trial.
Choose the arrangement that can answer routine questions while keeping shared figures explainable. Put its required author, model-owner and viewer capabilities into the software brief. The wider program for training and governing self-service creators is a separate implementation decision.
Role-Based Selection Trial: Key Acceptance Checks
- Step 1Give central author and business creator same approved measure and fixed-period task
- Step 2Ask each to produce a breakdown; have viewer find result and explain cut-off date
- Step 3Request change to shared measure: identify who can propose, approve, implement and communicate
- Step 4Record assistance needed, effective permissions, and whether draft stays limited to intended audience
- Step 5Include case where business creator cannot edit shared measure; test access control
