Back to Blog

The Power BI Store Sales Sample - A Practical Way to Learn Retail Reporting

November 19, 2026•7 min read•Michael Ridland

When someone in a retail business tells me they want to "get into Power BI", I usually suggest they start with somebody else's data. Not because their own data isn't interesting, but because their own data comes with baggage. The point-of-sale export has a weird date format. Store codes changed in 2023. Half the product hierarchy lives in a spreadsheet on the merchandise manager's desktop. You spend the first three weeks cleaning data and never get to the part where you learn how reporting works.

That's what the Store Sales sample is for. It's one of the built-in Power BI samples Microsoft ships, and it's modelled on a fictional retailer with a set of stores, product categories and sales over time. Clean data, a ready-made semantic model, a report to pull apart. For anyone in retail who wants to understand what Power BI can do with store performance data, it is the quickest path I know.

What's in the sample

The Store Sales sample represents a fictional company selling clothing and related goods through a chain of stores. The data covers sales across stores, districts and product categories, with comparisons between this year and last year. There's a split between new stores and existing stores, which matters because new stores distort like-for-like comparisons if you lump them in with the rest.

The report that comes with it looks at things like:

  • Total sales this year versus last year
  • Sales by district and store
  • Performance of new stores versus established ones
  • Sales per square foot, which is how retailers compare stores of different sizes
  • Category breakdowns and trends over time

If you've worked in retail, you'll recognise these immediately. They're the questions every area manager gets asked on a Monday morning. That familiarity is the sample's biggest strength for retail people. You're not learning Power BI and a foreign business domain at the same time.

(One small note for Australian readers: the sample uses US conventions, including square feet and US-style store data. It doesn't matter for learning purposes, but don't be surprised.)

How to get it

There are two common ways in. You can install it as a built-in sample straight into the Power BI service, which gives you the semantic model, report and dashboard in a workspace in a few clicks. Or you can download the .pbix file and open it in Power BI Desktop.

My recommendation: get the .pbix and open it in Desktop. The service version is fine for clicking around, but the learning happens when you open the model view, look at the relationships, read the DAX measures and start breaking things. You can't do much of that in the browser.

A practical warning: if you're installing the sample into the service at work, put it in your personal "My workspace" or a sandbox workspace. Don't drop it into the shared Finance workspace. I've seen "Store Sales" sitting in a production workspace at a client two years after a training session, confusing new starters who thought it was real data.

How we use it in training

We use samples like this in our Power BI training sessions, and the sequence that works best is roughly this.

Step 1: Read the report like a user. Spend ten minutes just using it. Click the bars, watch cross-filtering, hover for tooltips, use the slicers. Most people underestimate how much interactivity a good report gives you for free, and seeing it first makes the building part make sense.

Step 2: Open the model view. Look at the tables and relationships. You'll see the classic star schema shape: a sales fact table in the middle with dimension tables around it for store, item, time and district. This is the single most important concept in Power BI, and seeing a working example is far better than reading about it. When people's own models go wrong, it's nearly always because they skipped this shape.

Step 3: Read the measures. Find the measures for this year's sales, last year's sales and the variance. Look at how they're written. Time intelligence measures in particular are worth studying, because "same period last year" is a question every retailer asks and the DAX for it isn't obvious the first time.

Step 4: Change something. Add a new visual. Write a new measure, say average sale per store or the percentage of sales from new stores. Build a page that answers a question the original report doesn't. This is where it sticks.

Step 5: Ask what's missing. This is my favourite part with retail groups. Once people understand the sample, they quickly start saying "but we'd also need margin", or "we track foot traffic", or "where's the online channel?" That list of gaps is basically the requirements document for their own first report. We write it down on the whiteboard and keep it.

What the sample teaches well, and what it doesn't

Things it's genuinely good for:

  • Star schema modelling, with a clean real-world-ish example
  • Basic time intelligence (this year versus last year)
  • Retail-specific KPIs like sales per square foot
  • Seeing how a finished, reasonably designed report hangs together

Where it falls short of real life:

The data is too clean. Real retail data has returns, voids, stock adjustments, gift card redemptions and the occasional negative sale that nobody can explain. The sample has none of that mess. People who learn on the sample sometimes get a shock when their own POS data won't behave.

No margin or cost data to speak of. Revenue is half the story. Most Australian retailers I talk to care more about gross margin and markdowns than top-line sales, especially when costs have been climbing. The sample won't teach you how to model cost of goods or promotional markdowns.

No inventory. Store performance without stock levels is incomplete. Sell-through rates, stock cover and out-of-stocks are where a lot of the money is in retail analytics, and they need a different kind of fact table (snapshot rather than transactional) that the sample doesn't cover.

It's a little dated in design. The report works, but it doesn't reflect the newer formatting options or how most people design Power BI reports today. Treat the visuals as a starting point rather than a template.

None of that makes it less useful. It just means the sample is step one, not the destination.

From sample to your own store reporting

The jump from the sample to real store reporting is where most retail Power BI projects either take off or stall. A few things we've learned helping Australian retailers make it:

Get the store dimension right early. Store openings, closures, refits, relocations and changes in floor area all affect like-for-like comparisons. The sample's new-versus-existing split is a simplified version of this. In reality you want a store table with open dates, close dates and size history, and a clear rule for when a store becomes "comparable". Agree that rule with the business before building anything. It saves weeks of argument later.

Use a proper retail calendar. Many retailers report on a 4-4-5 or 52/53-week calendar rather than calendar months. Standard DAX time intelligence assumes a normal calendar, so you'll need a custom date table and measures written against it. Getting this wrong means every "versus last year" number is subtly off.

Bring margin in from day one. Even if the first release is sales-only, design the model so cost and margin can be added without a rebuild.

Think about where AI actually helps. There's a lot of interest in using AI for retail reporting: automated commentary on weekly performance, natural language questions over sales data, anomaly detection for stores that suddenly drop. Some of this works well now. Copilot can produce a decent narrative summary over a well-built model. But it only works if the model underneath is clean and well-named, which brings us straight back to the fundamentals the Store Sales sample teaches.

Is it worth your time?

Yes, if you're new to Power BI and you work anywhere near retail. An afternoon with the Store Sales sample will teach you more about how Power BI is meant to work than a week of fighting your own exports. If you're already experienced, it's less useful, though it's still a handy demo dataset when you want to try a new feature without risking real data.

If you're a retailer ready to move past the sample and build store reporting on your own data, our Power BI consultants have done this for Australian retail businesses of various sizes. You can also read more about our work in AI for retail, or if your team needs to get comfortable with the tools first, take a look at our business AI training.

Reference: Store Sales sample for Power BI - Microsoft Learn