Most Power BI projects fail not because of the technology, but because of scope creep and stakeholder misalignment. After deploying dashboards for logistics companies, healthcare clinics, law firms, and retailers across Canada, we've distilled the process down to a repeatable two-week sprint.
Week 1: Foundation
Day 1–2: KPI Workshop
Before touching Power BI Desktop, sit in a room (or a Teams call) with three to five people: the executive sponsor, one or two end users, and your data analyst. Answer three questions:
- What decisions should this dashboard enable?
- What does "good" look like for each metric?
- What data sources contain the relevant data?
Time-box this to two hours. You should leave with a list of 8–12 KPIs, ranked by priority.
Day 3–4: Data Audit
Connect to every data source identified in the workshop. Common sources for Canadian SMBs:
- ERP/accounting: QuickBooks, Sage, NetSuite
- CRM: Salesforce, HubSpot, Dynamics 365
- Operations: industry-specific platforms exported to Excel or SQL
- HR: BambooHR, Ceridian Dayforce
Document data quality issues now — missing values, inconsistent date formats, duplicate records. These will eat your timeline if discovered in week two.
Day 5: Data Model
Build the star schema in Power BI Desktop. Fact tables connected to dimension tables via clean relationships. Avoid the temptation to build one giant flat table — it performs poorly and makes DAX measures difficult to write.
Week 2: Build and Ship
Day 6–8: Measure Library
Write your DAX measures before building a single visual. This forces you to define calculations precisely and makes the report layer straightforward.
Revenue YTD = CALCULATE([Total Revenue], DATESYTD('Date'[Date]))
Revenue vs Budget = DIVIDE([Revenue YTD], [Budget YTD], 0)
Customer Retention Rate =
DIVIDE(
CALCULATE(DISTINCTCOUNT(Orders[CustomerID]), Orders[IsRepeatCustomer] = TRUE()),
CALCULATE(DISTINCTCOUNT(Orders[CustomerID]), PREVIOUSMONTH('Date'[Date]))
)
Day 9–10: Report Layer
Build report pages in this order: executive summary page first (single-page overview with 4–6 KPIs), then drill-through detail pages. Use bookmarks for state management, not separate report pages.
Design principles that survive stakeholder review:
- Maximize data-ink ratio. Remove gridlines, borders, and chart backgrounds unless they carry meaning.
- Use a consistent colour palette. Your company's primary colour for the primary metric; grey for context; red/amber/green only for status.
- Every visual needs a title that states the insight, not the metric name. "Revenue is 12% below forecast" beats "Monthly Revenue."
Day 11–12: UAT and Iteration
Share the report via Power BI Service and schedule a one-hour walkthrough with stakeholders. Capture feedback in writing. You'll typically get one round of significant changes and two rounds of minor tweaks.
Day 13–14: Production Deployment
Configure scheduled refresh (direct query or import mode depending on data volume and latency requirements), set up row-level security if multiple business units will share the workspace, and publish to the production workspace.
Document the data sources, refresh schedule, and KPI definitions in a one-page data dictionary. This prevents the "what does this number mean?" question six months later.
What This Costs
A two-week sprint engagement with Cleva IT Analytics typically runs $8,000–$15,000 CAD depending on the number of data sources and complexity of the model. Organizations with a Power BI Pro or Premium Per User licence can publish and share immediately — if you don't have licences, we'll help you size the right Microsoft 365 plan.
The alternative — a months-long BI project that delivers a dashboard nobody uses — costs far more.