Power BI: Where It Fits and What It Really Costs

When a business asks us which BI tool to use, Power BI is usually the first one on the table, and often the right one. That is not because it is the best at everything. It is because for most businesses already running Microsoft 365, it is the tool that fits into what people already use, at a price that starts small. Fit is the whole point of this series: each of the major BI tools was built with a different purpose in mind, and the right choice depends on yours.

Power BI was built to put analysis into the hands of people who live in Excel and Teams. It is at its best inside a Microsoft business with a modest number of report builders. It strains when the audience grows large, when the data model gets complicated without anyone owning it, or when nobody in the business is willing to learn its formula language.

Where Power BI fits best

• You already run on Microsoft 365. Sign-in uses the same accounts, reports can be pinned in Teams, data can be pulled into Excel, and security groups you already manage can control access. That integration is worth more than any single feature. If you are getting value from Microsoft 365 beyond email, Power BI is the natural next step.
• Your analysts are Excel people. Power Query, the tool that cleans and shapes data in Power BI, is the same one built into Excel. People who already build serious spreadsheets find the first steps familiar.
• A handful of people build, and a defined group reads. Ten, twenty, fifty users is the comfortable range for per-user licensing.
• Your data sits in common systems. SQL databases, Excel and SharePoint, accounting packages, CRMs and most cloud services have connectors. On-premise data reaches the cloud service through a gateway installed on a machine in your network.
• Data residency matters. Microsoft runs cloud regions in South Africa, and a tenant's home region can sit here, which helps the POPIA conversation.

How the licensing actually works

This is where most of the surprises come from, so it is worth being precise. Prices below are Microsoft's published US-dollar list prices on annual commitment at the time of writing; check the current price in rand with your licensing partner.

• Power BI Desktop is free. It is where reports are built, and it runs on Windows only, which matters if your analyst uses a Mac.
• Power BI Pro, about US$14 per user per month, lets people publish and share reports and view reports shared with them. The rule that catches people out: both the publisher and every viewer need Pro. Some Microsoft 365 enterprise plans include it.
• Premium Per User (PPU), about US$24 per user per month, adds larger datasets, more frequent refreshes and advanced features. Again, everyone viewing PPU content needs PPU.
• Fabric capacity is bought as a block of computing power rather than per person. At F64 and above, people with free licences can view reports, which is what makes it economical for a large reading audience. An F64 costs in the region of US$5,000 a month, so it pays off only when the number of viewers is in the hundreds.

The practical result: Power BI is inexpensive for a small team and gets steadily more expensive per head as you share more widely, until a large audience makes capacity the cheaper route. Plan the licensing around how many people read reports, not just how many build them.

Where it strains

DAX is a real skill

Simple totals are easy. Year-on-year comparisons, rolling averages, measures that behave correctly when filtered several ways at once: those need DAX, Power BI's formula language, and DAX has a genuine learning curve. A business that expects anyone to pick it up in a week ends up with reports that give subtly wrong numbers. Budget for training, or for someone who already knows it.

Sprawl

Because it is easy to start, Power BI grows organically: personal workspaces, copies of the same dataset with slightly different logic, five versions of "sales". Six months later two managers bring different revenue figures to the same meeting. The fix is governance from the start: shared, certified datasets for the core numbers, a naming convention, and an owner for each workspace. It is the same clean data discipline that every BI tool needs, just more urgent when building is this easy.

Exploration across messy, many-source data

Power BI works best on a well-designed model. Where the data comes from many loosely related systems and users want to roam freely through it, a tool built around associative exploration, such as Qlik, can be the better fit.

The audience outgrows the licence

As above: past a certain number of viewers, per-user cost adds up quickly, and the jump to capacity is large. Some businesses sit uncomfortably in between.

Getting the most out of it

1. Start with decisions, not visuals. The same principle as in our first look at Power BI: decide what the reports must help people decide.
2. Build one shared model for the core numbers. Sales, margin, stock, debtors. Certify it, and have every report use it.
3. Train the builders properly in Power Query and DAX, and give readers a short session on filtering and drilling.
4. Plan refresh and the gateway. Decide how fresh each dataset needs to be, and make sure the gateway machine is maintained and backed up.
5. Grow it a question at a time. As we wrote in why you have only one report, the business must keep pointing the specialists at the next question.

The short answer

If you run on Microsoft 365, have a modest number of people who need reports, and are willing to invest in a little DAX skill and some governance, Power BI is very likely the right tool, and the cheapest good one to start with. If your data is sprawling and exploration matters more than polished reports, or your reading audience is large and outside Microsoft, read on through the rest of this series before you decide.