How to plan a Power BI dashboard project
A good dashboard starts with the decisions it supports, not with the charts. This guide takes you from the first question to a report people open every week: what to agree, what to prepare, how to design, and what to check before you share it.
The short version
Six things that make a dashboard work
These six points decide whether a dashboard gets used. The sections below explain each one.
- Start from a decisionName who uses the dashboard and which decision or weekly meeting it supports.
- Agree each KPI in writingOne definition per measure, one owner, one source. Arguments about the number kill trust.
- Fix the data before the visualsClean, joined and refreshed data matters more than any chart type.
- Keep the design simpleA few clear visuals per page, consistent colours and the most important number first.
- Decide who sees whatRow-level security and sharing rules are designed from the start, not added at the end.
- Plan for adoptionTrain users, collect feedback and review the report after go-live.
The planning order
Plan in this order
Most dashboard rework comes from skipping the first two steps.
- Start with the question the page answers
- List where the data lives and who owns it
- Sketch the pages with the people who will use them
Quick enquiry
Planning a Power BI dashboard? Talk to us.
Questions to answer before you open Power BI
Write the answers down. They become the brief for whoever builds the report, whether that is your team or ours.
Who is it for?
Executives, department heads, field teams or customers. Each group needs a different level of detail.
What decision does it support?
For example: which projects are over budget, which branches miss target, which invoices are overdue.
Which KPIs matter?
Pick the few measures that drive the decision. Write the formula and the data source for each.
Where does the data live?
Excel files, SharePoint lists, an ERP, a database, a CRM. List every source and who owns it.
How fresh must it be?
Daily, hourly or on demand. Refresh needs affect cost and design.
Who may see which numbers?
Decide early if regions, departments or customers must see only their own data.
How a Power BI project runs, step by step
The names differ between teams. The order should look like this.
- 1
Define the questions and KPIs
Agree the decisions, the measures and their definitions with the people who will use them.
- 2
Connect and clean the data
Bring the sources together, fix duplicates and gaps, and agree what counts as the source of truth.
- 3
Build the data model
Shape tables and relationships so the report is fast and the measures give the same answer everywhere.
- 4
Design the pages
Sketch the layout first. Put the headline numbers at the top, then trends and detail below.
- 5
Test with real users
Check the numbers against figures you already trust. Ask users to find an answer without help.
- 6
Publish, secure and train
Set up the workspace, refresh schedule and access. Show users how to read and filter the report.
Build it yourself, use a template, or hire a partner?
All three are valid. The right choice depends on your data and your team.
| Route | Best when | Watch out for |
|---|---|---|
| Build it in-house | You have someone who knows Power BI and the data is already tidy. | Report sprawl and measures that differ between reports. |
| Start from a template design | Your question is common, such as sales, finance or HR, and you want a proven layout. | Forcing your data into a layout that does not fit. |
| Work with a partner | Several systems must be joined, the data needs cleaning or you need a model that others can extend. | A build with no handover. Ask for the files and the measure notes. |
Common mistakes in Power BI projects
Too many visuals on one page
Crowded pages hide the message. Split detail into drill-through pages.
Measures defined in several places
Put each calculation in the model once so every page agrees.
No owner for the numbers
When two reports disagree and no one owns the answer, trust drops fast.
Ignoring licences
Sharing reports in the Power BI service usually needs the right licences for viewers. Check this before launch.
Skipping the pilot
Test with a small group first. Fixing issues before everyone sees the report is cheap.
Treating go-live as the end
Reports need a review after a few weeks and an owner who keeps them up to date.
Power BI planning questions
How do I plan a Power BI dashboard?
Start with the decision the dashboard supports and the people who will use it. Agree each KPI and its source in writing, fix the data, build a clean model, design a simple layout, test with real users and plan access and training before go-live.
Which KPIs should a dashboard show?
Only the few that drive the decision it supports. Write the formula, the source and an owner for each. Extra measures can sit on drill-through pages.
Do I need a data warehouse for Power BI?
Not always. Small and medium needs can be met by connecting Power BI to your existing sources. A warehouse or a shared data model helps when many reports use the same data or when the sources are slow or messy.
Do I need Power BI Pro licences?
Power BI Desktop is free to build in. Sharing reports with other people in the Power BI service usually needs Pro licences for them, or a capacity that covers viewers. Check current Microsoft licensing and ask your partner to list what a design needs.
How do I limit who sees which data?
Use row-level security so that each person or group sees only their own region, department or customer. Design it with the model and test it with real accounts.
How do I get people to use the dashboard?
Build it with them, keep it simple, show it in the meeting where decisions are made, and train users to filter and read it. Review usage after a few weeks.
Can AscenWork build the dashboard for us?
Yes. We build Power BI dashboards and data models, and hand over the files and measure notes. The first call is free.
Talk to an expert