Standard BI Was Not Going to Cut It
Our client is a luxury goods retailer with stores around the world, a strong online presence, wholesale partners, and a brand built on craftsmanship and detail. They came to us with what most growing retailers have: data across many systems, a mix of dashboards built over time, and a sense that what they had was not really doing the brand justice.
For a luxury business, this matters more than it does anywhere else. When everything about your brand is designed, from the product to the packaging to the in-store experience, having a business intelligence dashboard that looks generic creates a small but persistent friction. The people running the business open something every morning that does not feel like the business they are running.
For a luxury brand, generic dashboards are not a neutral choice. They quietly undermine the experience of running a business that takes design seriously.
The brief, more or less, was to build something that felt like the brand. Polished. Considered. Interactive in ways that supported how people actually wanted to explore their data. And underneath all that, technically modern and built to last.
A Designed Experience, Not a Set of Dashboards
The first decision we made was to treat the BI environment as a designed product. That meant thinking about layout, typography, navigation, and interaction patterns as deliberate choices, not afterthoughts. It meant having opinions about what should be on the front page, what should sit behind a drillthrough, what should appear in a tooltip, and what should never be shown at all.
It also meant accepting that the standard Power BI building blocks would not be enough on their own. Out of the box, Power BI gives you good components. But to build something that genuinely feels designed, you sometimes have to go further. We did, in several places.
The Portal: One Designed Front Door
Power BI by default presents users with a list of reports. Functional, but anonymous. We replaced that with a custom-built Portal: a single landing page that renders a designed, branded navigation grid showing every report grouped by category — Sales, Customers, eCommerce, Finance, Operations.
The Portal is built entirely from a single DAX-generated HTML measure rendered through an HTML content visual. That sounds technical, but the experience is simple. Users open Power BI, see a designed front door that looks like the brand, click the report they want, and go straight there. No hunting through workspace folders. No mental mapping between report names and what they actually contain.
The result is that opening BI feels like opening a properly built product, not a directory of files.o be one of the most important design decisions in the whole build.
Custom Visuals Where Standard Ones Did Not Land
The other place we went beyond standard Power BI is in the Daily Sales Orders report. This is the most-used report in the business: a snapshot of how every channel is performing across Daily, Week-to-Date, Month-to-Date, and Year-to-Date, with comparisons against last year and against target. It feeds an automated email sent to the leadership team every morning.
The default Power BI matrix could do this, but not gracefully. Eleven channel groups, four time periods, two comparisons each — the standard visual would have produced a cluttered, hard-to-scan table. So we built a custom HTML-rendered Sales Matrix, composed of eleven HTML section measures concatenated together, with a designed CSS layout that gives every channel group its own clean row, formats the comparisons in the right way at the right size, and reads at a glance.
That same approach — DAX-rendered HTML for the visuals where it mattered — gives a level of design control that standard Power BI components simply cannot match. The trick is knowing where to use it. Too much, and the report becomes hard to maintain. Used selectively, it elevates the few places that need to be perfect.
Interaction Designed Around the Question, Not the Visual
The other piece of polish that often gets overlooked is interaction. How a user moves through the data, what they see when they hover, where they can drill, what they can compare. Done well, this is invisible: it just works the way the user expects.
Across the build, we used the full range of Power BI interaction patterns where each one earned its place:
- Tooltip pages. When a user hovers over a key KPI, a custom tooltip page appears showing the breakdown behind the number, formatted as carefully as the main report.
- Drillthrough pages. Clicking a channel takes the user to a dedicated drill page focused on that channel, with its own KPIs, charts, and breakdowns. The hidden drillthrough pages do not clutter the main navigation but they are always one click away.
- Action buttons. Designed buttons sit on the main pages, taking users to the right drill page in a way that feels intentional rather than incidental.
- Dynamic titles. Page titles update based on the slicers the user has applied, so they always know exactly what they are looking at.
None of these are exotic features. They are all in standard Power BI. The difference is that they have been used together, deliberately, to create an environment that feels coherent.
Sophistication Behind the Polish
The polish would not matter if the analysis underneath was thin. It is not. The build covers some of the most sophisticated analytical thinking we have done in business intelligence in retail industry projects.
- The Executive report covers every channel — Retail standalone stores, Concessions, Direct, China, Corporate, Franchise, Wholesale, third-party online — with revenue, gross margin, ATV, UPT, personalisation rate, upsell rate, footfall data from in-store sensors, and web traffic, all in one designed environment with drillthroughs into each channel.
- The Customer Lifecycle report breaks the entire customer base into nine distinct segments: Active, First-Time Buyers, New Repeat Buyers, Reactivated, At-Risk, Defecting, Defected, plus rolling one-year and two-year repeat purchase cohorts. Each segment has its own page, its own definition, its own trend analysis, and its own LTV view. The business can see exactly how customers are moving through the lifecycle and where the value is being created or lost.
- The Returns report splits every return into one of four root-cause categories: Production defects, Operations and warehouse errors, Subjective customer preferences, and Courier issues. Each has its own analysis, with year-on-year comparisons, preventable-versus-non-preventable splits, and a change log that tracks the impact of operational changes on return rates. For a luxury business where returns matter to both margin and customer experience, this kind of root-cause visibility is genuinely valuable.
The Foundation Underneath
None of this would be sustainable without a solid platform underneath. The data layer is built on Microsoft Fabric using a Medallion architecture: Bronze for raw ingestion, Silver for cleaned and harmonised data, Gold for curated reporting-ready views. Sources span Microsoft Dynamics 365 Business Central, NAV, Google Analytics 4, RetailNext for in-store footfall, and TripleWhale for marketing attribution.
The whole thing is orchestrated to refresh automatically every morning, with the Daily Sales Orders email landing in leadership inboxes shortly after. The Power BI semantic model is the single source of truth across every report, so a definition changed once is reflected everywhere — no drift between dashboards, no contradictory numbers between teams.
The Broader Lesson
For some businesses, BI is functional and that is fine. For others, particularly those whose brand depends on design and considered detail, business intelligence for luxury retail needs to be more than that. It needs to feel like part of the business it serves.
Getting there is not about using a different platform. It is about treating the build as a designed product, going beyond defaults where it matters, using the full interaction toolkit deliberately, and grounding the polish in genuine analytical depth.
The result is BI that the business actually wants to use. Reports that open every morning across the leadership team. Conversations that move from “can we trust these numbers” to “what should we do about what we are seeing.” A daily rhythm where the data is not friction but rather a tool that helps the business operate the way it wants to. And the principles transfer cleanly to any organisation thinking about how to take AI from interesting demos to genuine business capability.
