Power BI reporting built on a model your business agrees with
The visuals are the easy part. What makes Power BI worth the licence is a semantic model where every measure has one definition, so two departments stop arriving at a meeting with different revenue figures.
What is power BI?
Power BI is Microsoft's business intelligence platform for building governed reports over a shared semantic model, with row-level security, scheduled refresh and integration across the Microsoft stack. Consulting work covers modelling the data, connecting sources, securing access and training report authors. It suits Australian organisations already using Microsoft 365, Dynamics or Azure for their core systems.
Get a fixed written quote- Typical timeline
- 4 to 10 weeks
- What drives cost
- The number of source systems, how much cleaning the data needs before it can be modelled.
- Best for
- Organisations on Microsoft with reporting that must reconcile to the ledger
- You own
- The workspaces, the datasets, the gateway and the source code for pipelines
- Built with
- Semantic models, DAX, Dataverse and SQL sources, row-level security
Your handover
Why the semantic model matters more than the visuals
A semantic model is the layer that turns tables from your source systems into business concepts: a customer, an order, a site, a financial period, and the measures calculated from them. It holds the relationships, the date logic and the definitions. Build it well and every report drawn from it agrees, because gross margin is calculated in exactly one place and every author uses that calculation whether they understand it or not.
- 01Documented semantic model with a star schema
- 02Core measures written once and certified
- 03Refresh pipelines with monitoring and alerts
- 04Row-level security roles tested by impersonation
- 05Reports by subject area, built with their users
- 06Financial year and period logic for Australian reporting
- Workspace structure and publishing governance
- Reconciliation notes against source systems
- Author training and a documented support path
Build it badly and you get the situation that drives organisations to call us
Build it badly and you get the situation that drives organisations to call us. Six analysts have each connected directly to the warehouse, each written their own version of active customer, and the executive meeting spends its first twenty minutes reconciling numbers instead of deciding anything. The fix is not more dashboards. It is one governed model, published as a shared dataset, with reports built on top of it and direct source connections discouraged by policy rather than by hope.
- A star schema rather than reports querying raw operational tables
- Measures written once in the shared model and reused everywhere
- A proper date table supporting Australian financial year reporting
- Documented definitions visible to report authors inside the tool
- Certified datasets so authors know which model is authoritative
- Naming conventions that survive the analyst who wrote them leaving
Ask the question early, because the home region of a tenant is not trivial to change later.
Row-level security and who sees which numbers
Most organisations of any size have data that not everyone should see. A regional manager should see their own region. A franchisee should see their own store and a benchmark, not their neighbour's takings. Payroll detail belongs to a small group. Power BI handles this with row-level security, where the model filters itself according to who is signed in, so one report serves everyone and each person sees only their slice.
Getting it right is a design task, not a checkbox
Getting it right is a design task, not a checkbox. Security roles have to be modelled against your actual organisational structure, which is rarely as tidy as the org chart suggests, and then tested by impersonating each role rather than assumed to work. We also plan for the awkward cases: staff who cover two regions, acting managers, and people who move teams. This capability is a genuine reason to choose Power BI over lighter reporting tools, and it is the point where Looker Studio stops being sufficient for many organisations.
How the engagement runs
How a Power BI implementation runs
We deliver in slices. One subject area, modelled properly, reconciled and in the hands of users, beats a comprehensive model that arrives in month six and disagrees with the finance system on its first day.
- 01DiscoveryThe decisions the reporting supports, the source systems and who is allowed to see what
- 02Data auditQuality, completeness and the joins that will be difficult, identified before commitments are made
- 03Model designStar schema, grain, relationships, date logic and the first set of core measures
- 04Pipeline buildRefresh from each source, with a gateway or cloud connection, monitoring and failure alerts
- 05ReportsOne subject area at a time, reviewed with the people who will use it weekly
- 06SecurityRow-level roles modelled, applied and tested by impersonation before release
- 07RolloutWorkspace structure, publishing rules, author training and a support path when a refresh fails
Two decisions on your side that keep the project moving
Reconciliation is the stage clients underestimate. Every measure gets checked against the source system for a defined period, and where a variance exists we document why rather than adjusting the number until it matches. Explaining a known difference is credible. Silently forcing agreement is how a model loses trust permanently the first time somebody checks.
Where the data comes from and how it stays current
Reporting projects rarely fail on visualisation. They fail on the plumbing: a source system with no clean export, a nightly refresh that times out, a gateway installed on someone's desktop that gets shut down over Christmas. We treat data movement as engineering with monitoring and alerting attached, not as a scheduled task nobody watches.
The rest of the answer
Common sources for Australian organisations are Dynamics or Dataverse, an SQL database behind a line of business system, accounting platforms, and increasingly a custom application built for the business. Where an ERP or operational system holds the truth, we connect to it directly or stage it into a warehouse depending on load and refresh needs. If that source is a system we or another partner built, the reporting requirement should shape its API design rather than being solved with screen scraping later. Manufacturers and distributors typically hit this first, which is why reporting often surfaces during an ERP project.
Licensing, capacity and where your data sits
Power BI licensing is a Microsoft matter and you buy it directly, but it shapes the architecture so we plan around it from the start. The main decisions are whether users are licensed individually or you provision capacity, and whether you adopt the wider Fabric platform for storage and pipelines or keep those elsewhere. Capacity based licensing changes what you can do with large models and with sharing to people who do not have their own licence, and choosing wrongly is expensive to unwind.
Ask the question early, because the home region of a tenant is not trivial to change later
For data residency, Australian organisations can provision their tenant and capacity in Australian regions, which matters for government, health and financial services buyers whose procurement rules or internal policies require it. Ask the question early, because the home region of a tenant is not trivial to change later. Where personal information is involved, we also apply the Australian Privacy Principles to the model itself: collect the fields the reporting needs, aggregate where identity adds nothing, and restrict the rest through security roles instead of trusting that nobody will look.
When Power BI is heavier than you need
If your reporting need is marketing performance across a few advertising platforms and a website, Power BI is more platform than the problem requires. You will pay licences, build pipelines and maintain a model to answer questions a free tool answers natively. We say this often enough that it is worth putting on the page.
Moving up is straightforward
Power BI earns its place when at least one of these is true: numbers must reconcile to your accounting system, different people must see different rows of the same report, several source systems have to be joined into one version of the truth, or you are already invested in Microsoft and the governance and identity integration come for free. If none of those apply, start with Looker Studio and move later if you outgrow it. Moving up is straightforward. Paying for governance you never use is just a subscription.
How we scope it
Four ways to scope your Power BI project
We do not publish package prices, because the same brief can be a short build or a long one. These are the shapes the work usually takes. Tell us which one sounds like you and you will get a fixed written quote that spells out exactly what it covers.
Power BI Setup
Tracking that is correct, so the rest is worth reading
Fixed written quote, agreed before work starts
- Documented semantic model with a star schema
- Core measures written once and certified
- Refresh pipelines with monitoring and alerts
Measurement build
The measurement your decisions actually depend on
Fixed written quote, agreed before work starts
- Everything in Power BI Setup
- Row-level security roles tested by impersonation
- Reports by subject area, built with their users
- Financial year and period logic for Australian reporting
Power BI Full stack
Warehouse, pipelines and reporting across the business
Fixed written quote, agreed before work starts
- Everything in Measurement build
- Workspace structure and publishing governance
- Reconciliation notes against source systems
- Author training and a documented support path
Power BI Ongoing
Keeping it accurate as the site and the tools change
Rolling monthly, quoted in writing
- Tracking checked after every site or tool change
- A named person who knows the account
- Reports maintained as the questions change
- Rolling, cancel with 30 days notice
These are shapes, not menus. Most quotes end up somewhere between two of them, and we will say so when the honest answer is the smallest one. Describe the problem and we will tell you which it is.
Questions buyers usually ask
Frequently asked questions
Ownership and handover
Do we own the model and the reports?
Yes. Everything is built inside your Microsoft tenant, in workspaces your organisation controls, and any pipeline code is delivered in a repository you own. We hold access you can revoke. The model is documented so an internal analyst or another consultancy can extend it, and we would rather hand over something maintainable than remain necessary by default.
Can our team build their own reports afterwards?
That is the aim of a certified shared dataset. Analysts build reports on top of the governed model without needing to understand the underlying joins, and because the measures are defined once, their reports agree with everyone else's. We train report authors as part of rollout and set publishing rules so a personal experiment does not end up circulating as an official figure.
Detail and edge cases
How long does a Power BI implementation take?
Four to ten weeks for a first subject area including the model, pipelines, security and reports. Broader programs run longer and we deliver them in slices so value arrives early. The variable is data quality at the source. Clean, well structured systems move quickly. A decade of inconsistent data entry in an operational system can add weeks that no amount of modelling skill removes.
What drives the cost of a Power BI project?
The number of source systems, how much cleaning the data needs before it can be modelled, the complexity of your security requirements and how many report audiences you serve. Microsoft licences are separate and paid by you directly to Microsoft. We scope the work in discovery and send a fixed written quote, then flag in writing if anything we find during the data audit changes the picture.
Should we adopt Microsoft Fabric as well?
Only if you have a reason. Fabric brings storage, pipelines and reporting into one platform, which suits organisations consolidating a scattered data estate. If your sources are few and a straightforward model over a database serves you, adding Fabric adds cost and concepts your team must learn. We assess it against your actual data volume and roadmap rather than defaulting to the newest option.
Can Power BI report on data from a system you did not build?
Usually yes, provided the system offers a database connection, an API or a reliable scheduled export. The awkward cases are older on premises systems with no documented interface, where we may need a gateway and a staging database. We check connectivity during discovery, before scoping the reporting, because that answer changes the shape of the whole project.
Related services
Get a scoped Power BI engagement
Tell us your source systems, who needs to see what, and the report that currently starts an argument. One business day for a response, and the number arrives in writing.