Map the architecture to the business workload
We review source systems, connection paths, data preparation, published data sources and the workbooks that depend on them. The scope captures audience, concurrency, data volumes, refresh windows and decision deadlines. This establishes which workloads matter and where a delay or failure has a practical business consequence.
- Source connectivity and data-model dependencies
- Shared metric definitions and reusable analytical models
- Identity, access and content-publishing responsibilities
- Refresh workload, user activity and operating capacity
Investigate where the time is spent
A slow workbook may involve source queries, calculations, the amount of displayed detail or rendering. We use available performance recordings and workload evidence to isolate relevant components, then test changes against a repeatable scenario. Refresh problems are reviewed separately, including schedules, source availability and failed-job history. Each proposed change is assessed in the actual environment.
Define a governed metric and access layer
Reusable data models and documented measures give teams a shared foundation for analysis. We review where calculations belong, how definitions are approved and how access follows business responsibilities. Governance covers data owners, content promotion, freshness checks and retirement of duplicated or unused workbooks. It also identifies who maintains each dependency when a source or definition changes.
Assess Cloud or Server fit and capacity
Tableau Cloud and Tableau Server have different operating responsibilities. We assess the fit against your source networks, identity requirements, workload and support model. For Server, the review can include deployment topology and resource capacity; for Cloud, it can include connectivity and refresh design. Product availability, licensing and any Tableau Next or Tableau+ requirements are confirmed during scoping.
Deliver a sequenced architecture plan
Deliverables can include a current-state architecture map, a measured performance baseline, a dependency and ownership register, and a prioritized remediation plan. The plan separates immediate workbook or refresh changes from structural work and identifies acceptance checks. Changes are retested under agreed conditions, with operating notes for the team responsible for maintaining the platform.
Common questions
Will adding server capacity fix slow dashboards?
That depends on the measured bottleneck. Source queries, workbook design and workload contention can require different changes. The review establishes where time is spent before recommending additional capacity.
How does this differ from business performance diagnostics?
Architecture and performance work examines the analytics platform: speed, freshness, models, access and operational reliability. Business diagnostics investigates the drivers of commercial or operational results. The two can be scoped together when the business question and platform constraints are connected.