Originally published 13 September 2023. Fully updated for 2026.
Most businesses already have the data they need to run better. It's just scattered across a finance system, a CRM, a few spreadsheets on SharePoint and someone's inbox. A good Power BI dashboard pulls that together into one view that updates itself, so the Monday meeting starts with the numbers instead of an argument about which spreadsheet is right.
This guide walks through how to build one properly. The core steps haven't changed much since we first wrote it, but the platform around them has: Power BI now sits inside Microsoft Fabric, "datasets" are now called semantic models, and Copilot can draft report pages for you. We've updated everything accordingly.
What is Power BI in 2026?
Power BI is Microsoft's business intelligence platform. It has three parts you'll use:
- Power BI Desktop: a free Windows application where you connect to data, model it and design reports.
- The Power BI service: the cloud side (app.powerbi.com), where reports are published, refreshed, secured and shared. It's now one of the workloads within Microsoft Fabric, Microsoft's wider analytics platform.
- Power BI Mobile: apps for iOS and Android so people can check dashboards on the go.
Reports vs dashboards: a quick clarification
In everyday language, "dashboard" means any screen full of charts. In Power BI it has a specific meaning. A report is the multi-page, interactive thing you build in Desktop. A dashboard is a single-page canvas in the service made of tiles pinned from one or more reports. Most of the work in this guide is building the report; the dashboard is the executive summary you assemble at the end.
Before you start: licensing in brief
Building in Power BI Desktop is free. Sharing is where licensing comes in. Here's the simplified picture:
| Licence | What it's for |
|---|---|
| Free | Build and view your own content. You can't share with or collaborate with others. |
| Power BI Pro | Publish, share and collaborate in workspaces. Both the creator and the viewers generally need Pro. Included in Microsoft 365 E5. |
| Premium Per User (PPU) | Pro features plus larger models, more frequent refresh and advanced features, licensed per person. |
| Fabric capacity (F SKUs) | Capacity bought for the organisation. Unlocks Fabric workloads and Copilot (on paid capacities). On larger capacities, viewers with a free licence can consume shared content. |
For most small and medium businesses, Power BI Pro for the people who build and view reports is the sensible starting point. Microsoft adjusts these tiers from time to time, so confirm the current details on Microsoft Learn before you buy.
The step-by-step build
Step 1: Decide what the dashboard is for
The step everyone skips. Before opening Desktop, write down who the dashboard is for and the five to ten questions it must answer. "How are sales tracking against target this month?" is a good question. "Show me everything" is not. Agree on definitions too: does "revenue" mean invoiced, paid or booked? Settling this now saves weeks of rework.
Step 2: Connect to your data sources
In Power BI Desktop, choose Get data. Power BI has connectors for Excel, SQL Server, SharePoint lists and files, Dataverse, Azure services, Salesforce, Google Analytics and many more. Some practical tips:
- Connect to Excel files stored in SharePoint or OneDrive, not on someone's desktop, so the service can refresh them.
- For on-premises databases, you'll need the on-premises data gateway installed on a server that stays on.
- Where a business system has no native connector, its API or a scheduled export into SharePoint is usually the pragmatic route.
Step 3: Clean and shape the data in Power Query
Power Query is where raw data becomes usable. Remove columns you don't need, fix data types, split or merge fields, filter out test records and standardise names (so "Vic", "VIC" and "Victoria" become one value). Every step is recorded and replayed on each refresh, so you do the cleaning once.
Step 4: Build the data model
This is what separates a dashboard that scales from one that breaks. Aim for a star schema: fact tables (sales transactions, tickets, timesheets) in the middle, linked to dimension tables (customers, products, dates, staff). Always add a proper date table and mark it as such, so time comparisons like month-to-date and year-on-year work correctly.
Step 5: Create measures with DAX
Measures are the calculations behind your KPIs, written in DAX (Data Analysis Expressions). Start simple: total sales, gross margin percentage, average days to pay. Define each KPI once as a measure and reuse it everywhere, rather than recalculating it in individual visuals.
Step 6: Design the visuals
Now the fun part. Some rules that consistently make reports easier to read:
- Put the headline KPIs as cards across the top, trends in the middle, detail at the bottom.
- Use line charts for trends over time, bar charts for comparisons and tables for detail. Be sparing with pie charts and gauges.
- Limit each page to one audience and one purpose.
- Apply a report theme with your brand colours so every page looks consistent.
Step 7: Add interactivity
Power BI visuals cross-filter each other by default: click a region on a map and every other visual filters to that region. Add slicers for date ranges, teams or product lines, set up drill-through pages so people can click from a summary to the detail behind it, and use bookmarks to save useful views.
Step 8: Publish, secure and schedule refresh
Publish from Desktop to a workspace in the Power BI service. Then:
- Configure scheduled refresh for the semantic model (up to eight times a day on Pro, more on PPU and capacity).
- Set up row-level security if different people should see different data, for example each branch manager seeing only their branch.
- Control access through workspace roles and apply sensitivity labels if your organisation uses them.
Power BI's "Publish to web" option creates a public link that anyone on the internet can view, with no sign-in. It's meant for public data only. For internal dashboards, share through workspaces, apps, Teams or SharePoint instead, and consider having your administrator restrict the feature in the tenant settings.
Step 9: Build the dashboard and share it
In the service, pin the most important visuals from your report onto a dashboard. You can set data alerts on card and KPI tiles so people get notified when a number crosses a threshold. Then share the content where people already work: package it as a Power BI app for a clean audience-specific experience, add it as a tab in a Microsoft Teams channel, or embed it in a SharePoint page on your intranet.
Where Copilot fits
Copilot in Power BI can suggest and draft report pages, summarise what a report is showing in plain language, help write DAX and answer questions about your data. It requires a paid Fabric capacity (or eligible Premium capacity) and has to be enabled by your administrator, and it's worth checking where Copilot processes data if residency matters to you (see Microsoft's Copilot for Power BI overview).
The catch is that Copilot is only as good as the model underneath it. Clear table and column names, a proper date table, well-defined measures and descriptions on your fields make a big difference to the quality of its answers. Steps 3 to 5 matter more with Copilot, not less.
Common mistakes to avoid
- Refreshing from a laptop: reports sourced from files on someone's local drive stop refreshing when that laptop is off.
- One giant flat table: it works for a demo and falls over as data grows. Model it properly.
- No owner: every dashboard needs a named owner who fixes it when a source system changes.
- Sharing from "My workspace": use shared workspaces so content doesn't disappear when someone leaves.
How synfo can help
Getting a first dashboard live is very achievable in-house. Where businesses tend to want help is the plumbing: connecting line-of-business systems, setting up gateways, licensing, security and governance so the reporting keeps working as you grow. Our app development and integration team connects systems and builds on the Microsoft Power Platform, and our managed IT plans look after the Microsoft 365 environment it all runs on. If you're still deciding on your productivity platform, our Microsoft 365 vs Google Workspace comparison is a good place to start.