A Power BI implementation runs in six stages: agreeing what the business needs to know, connecting and cleaning the data, building the data model, building the reports, setting up security and sharing, and handing over with training and support. For a business with two or three standard systems the whole thing takes four to eight weeks. This guide explains what each stage produces, where projects stall, and the two settings an Australian business should get right at the start: where the data lives and who can see what.<\/p>\n
Before anyone opens Power BI, the people who will read the reports say what they need to know each week and each month, in their own words: “are we on target this month”, “which jobs are over the hours quoted”, “who is overdue”. Five to ten questions is normal. Each one is matched to a number and a definition, written down, so that “revenue” means the same thing to sales and to finance. Microsoft’s own implementation planning guidance<\/a> starts the same way, with strategy before tenant setup or tools. Our guide to writing a measurement plan<\/a> covers this stage.<\/p>\n Power BI connects to Xero, MYOB, Dynamics, SQL databases, SharePoint, Excel, Google Analytics and hundreds of other sources. For cloud systems the connection is direct. For anything on a server in your office, a small piece of software called a gateway is installed so the Power BI service can reach it safely. Then the data is looked at properly, and this is where most of the surprises are: customers entered three ways, invoices coded to the wrong account, dates in text fields. The cleaning is limited to the fields the questions from stage one depend on, and anything structural is fixed at the source rather than patched in the report.<\/p>\n The model is the part nobody sees and everything depends on: the tables, how they relate, the calculations (revenue, margin, hours over quote), and the date logic for comparing this month with last. A good model means every report built on it agrees with every other, and a new question is a ten-minute change rather than a new project. A poor model means every report has its own version of the truth. This stage is where an experienced implementer earns their fee.<\/p>\n One page per question, the most important number top left, five or six visuals rather than twenty, consistent colours, few filters. The first version reproduces the report people already trust, so they can check it against what they know. Then the questions nobody could answer before are added, one at a time. Reports are laid out for phones as well as desktops, because owners read them from the car.<\/p>\n Two settings matter for an Australian business. The first is where the data is stored: a Power BI tenant has a home region, and Microsoft’s documentation on tenant location<\/a> explains how to check it. For most Australian organisations it should be an Australian region, and that is decided when the Microsoft 365 tenant is set up, so check before building anything. The second is row-level security: who sees which rows. A state manager opens the same report as the national manager and sees only their state. Set up once in the model, it applies to every report. Then sharing: workspaces for each audience, reports published to them, and a scheduled refresh so the numbers update overnight without anyone doing anything.<\/p>\n Documentation of the model and the definitions, a walkthrough for the people who will read the reports, and training by task for whoever will own the reports internally. Then support: watching the refresh, fixing it when a source system changes, and adding the next question. Some businesses take this in house once someone is trained; others keep a monthly arrangement. The failure is having neither, because a report nobody maintains stops being trusted within a few months.<\/p>\n Definitions not agreed, so the first meeting about the report is an argument. Dirty source data discovered late. Access to systems not granted for weeks. The report trying to show everything. And nobody owning it after launch. None of these is a Power BI problem, and all of them are avoided by doing stage one properly and keeping the first version small.<\/p>\n Want this as a one-page checklist?<\/strong> Six stages and ten steps on one A4 page, with tick boxes and the places projects usually stall.<\/p>\n Download the checklist (PDF)<\/a><\/p>\n<\/div>\n This is a hypothetical, not a client. Take a Gold Coast labour hire company with 80 staff, running a rostering system, Xero, and a spreadsheet of client rates. Stage one produces seven questions, of which “which clients are we losing money on” is the one nobody can currently answer. Stage two connects the three sources and finds that client names differ between the roster and Xero, which is fixed at the source. Stage three builds a model with a single definition of margin. Stage four reproduces the monthly report in week three, then adds margin by client in week four. Stage five sets row-level security so each branch manager sees their branch, with the tenant confirmed as Australian. Stage six trains the finance manager to own it. Six weeks. The margin-by-client page changes which contracts the business renews.<\/p>\n High10 implements Power BI<\/a> for Australian businesses in these six stages, starting with the questions and the definitions, reproducing the report the business already trusts, and handing over to someone in your team or supporting it monthly. Where a website or Google Analytics feeds the reports, we handle the analytics implementation<\/a> too. Every project is quoted in writing after a short scoping call.<\/p>\n Four to eight weeks for a business with two or three standard systems and a first version that reproduces an existing monthly report. Larger implementations with many sources, complex security or a migration from another tool run three to six months. Most of the calendar time is agreeing definitions and getting access, not building.<\/p><\/div><\/details> It can be. A Power BI tenant has a home region set when the Microsoft 365 tenant is created, and for Australian organisations it should be an Australian region. Check it before the implementation starts, because changing it later is a significant exercise.<\/p><\/div><\/details> Not for a first implementation with a few standard systems; Power BI’s own model handles it. A warehouse becomes worth considering when there are many sources, large volumes, or several tools reading the same data. Start without one and let the need show itself.<\/p><\/div><\/details> It depends on the number of sources, how clean the data is, how many questions the first version answers, and the security required. We quote each project in writing after a scoping call rather than publishing a rate. Our guide to Power BI consultant costs<\/a> gives the Australian market ranges.<\/p><\/div><\/details><\/div>\n Want this as a one-page checklist?<\/strong> Six stages and ten steps on one A4 page, with tick boxes and the places projects usually stall.<\/p>\n Download the checklist (PDF)<\/a><\/p>\n<\/div>\n A Power BI implementation runs in six stages: agreeing what the business needs to know, connecting and cleaning the data, building the data model, building the reports, setting up security and sharing, and handing over with training and support. For a business with two or three standard systems the whole thing takes four to eight […]<\/p>\n","protected":false},"author":2,"featured_media":192,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[],"class_list":["post-183","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-data-analytics"],"yoast_head":"\nStage 2: connect and clean the data<\/h2>\n
Stage 3: build the data model<\/h2>\n
Stage 4: build the reports<\/h2>\n
Stage 5: security and sharing<\/h2>\n
Stage 6: handover, training and support<\/h2>\n
Where implementations stall<\/h2>\n
A worked example<\/h2>\n
How High10 helps<\/h2>\n
Frequently asked questions<\/h2>
How long does a Power BI implementation take?<\/h3><\/summary>
Is Power BI data stored in Australia?<\/h3><\/summary>
Do I need a data warehouse for Power BI?<\/h3><\/summary>
What does a Power BI implementation cost?<\/h3><\/summary>
Related reading<\/h2>\n
\n