The Power BI Sales and Marketing Sample - What You Can Actually Learn From It

Microsoft ships a handful of sample datasets with Power BI, and the Sales and Marketing sample is probably the one I've seen referenced most often in client training material. It's the "everyone has used this at some point" dataset. New analysts open it on their first day. Trainers use it for workshops. People copy patterns from it into production reports without quite realising that's what they're doing.

I've been on the receiving end of that last one a few times. A client sends me a report to review, I open it up, and I can tell within thirty seconds that whoever built it learned Power BI on the Sales and Marketing sample, because the visual layout, the colour choices, and a few measure patterns are dead giveaways. That isn't always a bad thing. The sample teaches some good habits. It also teaches a few things that don't translate well to real work. So this post is what to take from it, what to leave behind, and how to use the sample as a learning tool without building the wrong muscle memory.

What's actually in the sample

The Sales and Marketing sample tells the story of a fictional manufacturing company called VanArsdel that sells bicycles and accessories. The dataset contains sales data, market share data, sentiment data, and some category-level competitive data. The report has multiple pages that walk through how the business is performing across regions, segments, manufacturers (competitors), and time.

You can download it as a .pbix file directly, or you can load the sample app from the Power BI service. I usually grab the .pbix because that's where you can actually inspect the model, the measures, and the visual configuration. The service app is fine for a quick look but you can't see how anything is built.

The model is small (a few fact tables, a few dimensions), and the DAX is mostly straightforward. That's deliberate. Microsoft built this sample to teach concepts, not to model real complexity. Which is also where the limits come in. If you treat the Sales and Marketing sample as an example of how to build a real production model, you're going to miss most of what makes a real production model hard.

What the sample does well

A few things in this sample are genuinely worth copying.

The page narrative. Each page tells a self-contained story. There's a clear headline metric at the top, a few supporting visuals underneath, and a way to break the metric down by relevant dimensions. The pages don't try to do everything at once. If you opened most enterprise Power BI reports you'd see the opposite pattern, where every page is a wall of small visuals competing for attention. The sample's restraint is the better template.

The use of small multiples and category context. Several pages use small multiples or repeated patterns to show how a metric varies across a dimension. That's a more honest visualisation choice than a single chart with a busy legend. You can see what's actually happening to each category, not just the aggregate.

The way KPIs are introduced with a target. A KPI without a target is just a number. The sample is good about pairing each headline number with what it should be, so the reader has context. We push this on every client report we build. A revenue number is meaningless without "compared to what". The Sales and Marketing sample bakes this into its visual choices and it's worth copying.

Tooltips that earn their keep. Custom tooltips are used on several visuals to give extra context when you hover. They aren't doing anything fancy, but the principle (use the tooltip for context, don't make the user click into a detail page for a small extra fact) is sound.

This is the bit I want most junior analysts on our training programs to take from the sample. We cover it in our Power BI consulting work and in some of our broader AI and data training. The Sales and Marketing sample's structural choices teach you more about good reporting than its data choices do.

What I'd actively avoid copying

Now the parts that get repeated too often, that aren't really good models for production work.

The colour palette. The sample uses a punchy palette that works fine for a tutorial dataset but is way too saturated for most corporate reports. If you copy these colours into a real-world report, the report will look like a sample. Your business has a brand palette. Use it. Spend ten minutes setting up a custom theme JSON file and apply it across every report in the workspace. We do this on every engagement and it changes how the reports feel before any data is even loaded.

The map visual choices. The sample uses a few map visuals that look slick but are awful for actual analysis. A bubble map with overlapping bubbles doesn't communicate values clearly. A choropleth map without proper colour breaks misleads more than it informs. Use maps sparingly and only when geography is the question the user actually has. If the user wants to know "which region is performing best", a sorted bar chart beats a map every time.

The cluttered single-page summary. One of the pages in the sample is a "everything at once" summary page that's the exact pattern I just warned about. It exists in the sample because Microsoft wants to show off many visual types in a small space. Don't take it as a layout suggestion. It's a feature catalogue, not a design pattern.

The simplified data model. The model in the sample is essentially a flat star schema with three or four tables. Real businesses have more complex relationships. Multiple fact tables with different grains. Slowly changing dimensions. Many-to-many relationships. The sample doesn't teach you how to handle any of that. If you only ever practice on the Sales and Marketing model, you'll be unprepared for the first real client model you see. The most useful follow-up to the sample is to find a public dataset with real complexity and work through modelling it from scratch. Or even better, sit next to someone who has done it for a few years.

How to actually use the sample productively

If you're learning Power BI, here's the path I'd suggest. Not just "open the sample and click around", which is what most tutorials suggest, but a more structured way to extract value.

First, open the .pbix and look at the model view. Note the relationships, the cardinality, the cross-filter direction. Try to predict what each measure does before you click on it. This forces you to actually think about the model rather than passively absorbing it.

Second, open the View > Performance analyzer and watch what happens when you click around. Performance analyzer shows you how long each visual takes to render and how that breaks down between query time and visual time. Most beginners never open this tool. It's how you learn that a chart with a few hundred categories will be slower than one with a few dozen, and that DirectQuery and Import have very different performance profiles.

Third, rebuild a page from scratch. Don't just admire the existing layout. Pick the "Yearly Sales" page, hide it, and try to build it again using the same data. You'll discover things you never noticed by reading. This is how you learn what each visual setting actually does.

Fourth, try to break something. Add a measure that does the wrong calculation on purpose. Watch it propagate through the report. Then fix it. The thing that makes you good at Power BI isn't knowing how to use the features. It's knowing how to diagnose when something goes wrong, and you only get there by deliberately breaking things in a safe environment.

This is roughly the structure we use when we run Power BI training sessions for client analyst teams. Open the sample. Make people predict, rebuild, and break. Then move them onto a real dataset where the model isn't so neat.

Where the sample isn't going to help you

The Sales and Marketing sample isn't a guide to building a data model. It isn't a guide to writing complex DAX. It isn't a guide to row-level security, incremental refresh, deployment pipelines, or any of the operational concerns that actually take up your time once you're past the basics. Those are the things that make Power BI a craft. The sample is a starting point and you should treat it that way.

For Australian businesses doing real work in Power BI, the next steps after the sample are usually about three things: understanding your real data model (which is rarely as clean as the sample), building patterns for ongoing maintenance (because reports rot if nobody's tending them), and integrating Power BI with the rest of the data platform (especially if you're moving into Fabric or building AI-driven insights on top). We cover all of that as part of our broader Microsoft data and AI consulting. The sample is where you start. It's not where you end up.

A small recommendation

If you do download the Sales and Marketing sample, also download two or three of the other built-in samples. The Retail Analysis sample and the Customer Profitability sample teach different things. The Retail one is closer to a realistic operational report. The Customer Profitability one has a more interesting data model. Looking at all three together is more educational than studying any one of them in depth, because you can see how the same kinds of problems get solved with slightly different patterns.

For more on what's in this particular sample and how to load it, the Microsoft documentation page has the download links and a walkthrough of each page.

If you're trying to lift your team's Power BI capability past the "we can open the sample" stage and into "we can build and maintain production reports", that's the kind of thing we help Australian companies with on a regular basis. Have a look at our Power BI services and get in touch if you want to talk through what good looks like for your team.