Microsoft Fabric and Azure data platforms in La Louvière: architecture, capacity and cost
How to build a Microsoft Fabric or Azure data platform for a La Louvière organisation: medallion layout for metals and materials and logistics and freight, honest capacity sizing, Direct Lake trade-offs and the Belgium residency position.
Local view for La Louvière, Belgium
A Fabric capacity bought before anyone has profiled the workloads is an expense in search of a benefit. That is the position several La Louvière organisations find themselves in when they call us.
This page is about building a data platform for an organisation in La Louvière, Microsoft Fabric, Azure Synapse, Data Factory and the lakehouse patterns underneath them, with the local economy taken seriously. La Louvière's base in metals and materials, logistics and freight, and industrial machinery and engineering determines the data volumes, the latency requirements and the retention obligations, and those three things decide the architecture. The client population here is shaped by steel and metals processing and a major canal and intermodal logistics position.
Does a La Louvière organisation actually need Fabric?
Look for a cluster rather than a single trigger: import refreshes that are getting uncomfortable, demand for SQL and notebooks against one copy, self-service across more than one team, retention that rules out a desktop model, or an existing Premium commitment. One analyst maintaining three import models that refresh in twenty minutes does not need a capacity, and buying one will not improve anything.
| Situation | Recommendation | Why |
|---|---|---|
| Under about 10 GB of modelled data, one or two authors, overnight refresh is fine | Stay on Power BI Pro with import models | A capacity adds cost and operational surface with no benefit you can measure |
| Refresh windows no longer fit overnight, incremental refresh already in use | Lakehouse or warehouse serving layer | The constraint is transformation throughput, not the reporting tool |
| Analysts need SQL, data scientists need notebooks, both on the same data | Fabric, serving layer first | One copy of the data with multiple engines is the thing Fabric genuinely does well |
| Retention obligations measured in years at event grain | Lakehouse with lifecycle policies | Delta on cheap storage is far less expensive than keeping history in a semantic model |
| Existing Synapse or Databricks platform that works and costs are stable | Keep it; surface through Fabric or Power BI | Re-platforming a working estate to tidy an architecture diagram is not a business case |
What the data actually looks like in La Louvière
Metals and materials
Heat and coil genealogy, energy-intensive process data, and quality certification per batch, with energy cost a dominant and volatile input. Source systems typically include process control systems, ERP, quality certification systems, energy monitoring and logistics systems.
For a platform design, the consequence is specific. Energy and production data are reported on different period bases, so the cost per tonne figure moves for reasons nobody can explain. So the architecture has to make the raw grain replayable rather than aggregating at ingestion, and the first serving model should be a genealogy and yield model with energy allocation, which supports both operational improvement and emissions reporting from one source.
Construction and real estate
Project cost and progress data with subcontractor valuations, variations and retentions, plus asset and tenancy data on the property side. Project reporting runs monthly on manually consolidated spreadsheets, so a problem contract is identified two months after it turned.
Systems to plan for: project costing systems, ERP, BIM and design systems, site progress apps and property management systems. Subcontractor payment and tax reporting obligations apply, and health and safety records carry their own retention duties.
Logistics and freight
Event data from scanners, telematics, EDI and transport management systems, each with its own timestamp semantics and time zone handling. Subcontracted volume arrives as spreadsheets, so a third of the network is effectively unmeasured exactly where control is weakest. Drivers' hours records, customs declarations and posting documentation all carry retention obligations and are frequently requested in audit.
A medallion layout for a metals and materials business
The medallion pattern is not complicated, and it is routinely implemented badly: a bronze layer that has already lost fidelity, a silver layer with no testable rules, and a gold layer that is really just the silver layer with different names. Here is how we would lay it out for a La Louvière client in metals and materials.
| Layer | What lands here | Ownership | Rules |
|---|---|---|---|
| Bronze | Raw extracts from process control systems, ERP and quality certification systems, at source grain, with ingestion metadata | Platform team | Append-only, replayable for any date, no business logic, retention set against your obligations |
| Silver | Conformed, deduplicated, type-safe entities with the agreed business keys | Platform team with domain sign-off | Every transformation testable; data quality assertions block promotion on failure |
| Gold | Dimensional models for reporting: genealogy and yield model with energy allocation | Domain team | One definition per measure, documented; changes go through review |
| Semantic | Direct Lake or import model serving reports and Excel | Analytics engineering | No business logic that is not also in gold; measure dictionary maintained |
Capacity sizing and what it really costs
Fabric capacity is billed by capacity unit. An F2 is roughly US$0.36 per hour pay-as-you-go, about US$260 a month, with around 40 per cent off for a one-year reservation. An F64, the smallest SKU at which report consumers no longer need an individual Pro licence, is roughly US$8,400 a month pay-as-you-go or around US$5,000 reserved.
The arithmetic that matters: F64 only becomes cheaper than per-user licensing somewhere in the region of 350 to 600 readers, depending on the discount you actually pay and whether you reserve capacity. Below that, Pro or PPU per user is normally the right answer, and buying capacity for prestige is a straightforward waste of money.
| Option | Who it suits | Approximate list price | The catch |
|---|---|---|---|
| Power BI Free | One person exploring data with no sharing | No charge | Cannot share into a workspace, so it is not a business option |
| Power BI Pro | Small reader populations, typically under about 200 people | Around US$14 per user per month at list | Every reader needs a licence, so cost scales linearly with the audience |
| Premium Per User | Teams needing larger models, higher refresh rates or paginated reports | Around US$24 per user per month at list | Still per-user, so a large reader base gets expensive quickly |
| Fabric F2, F32 capacity | Data engineering workloads with a modest reader population | From roughly US$260 a month for an F2, pay-as-you-go | Readers still need Pro licences below F64, and small SKUs throttle under load |
| Fabric F64 capacity | Large reader populations and in-platform data engineering | Roughly US$8,400 a month pay-as-you-go, around US$5,000 reserved | Only pays for itself above roughly 350 to 600 readers, depending on your discount |
We size capacity by measuring rather than estimating: each workload profiled for compute shape, background refresh against interactive query, concurrency and peak hours, then validated on a short-lived trial capacity with your real data. The output is a sizing model that states expected utilisation at launch, at double the readers, and the specific point at which you would need the next SKU. Always price from the current Microsoft rate card for your region and agreement. Any figure in an article older than a quarter should be treated as an order of magnitude, not a quote.
Direct Lake, import or DirectQuery
The appeal of Direct Lake is that nothing is copied and nothing is refreshed. The risk is the fallback: one unsupported construct and the model quietly behaves like DirectQuery instead. We test it against an import model on your own data and choose on the measured result. In practice Direct Lake tends to win on large, frequently landing tables, which metals and materials workloads often are, while import still wins on smaller models with heavy DAX.
Migrating from what you have now
Few La Louvière organisations start from nothing. The realistic options are a serving-layer-first adoption, where Fabric takes the gold layer and reporting while existing engineering stays in place, or a domain-by-domain migration with the old platform retained until each domain is validated. What both have in common is that nothing is switched over in a single weekend. Single-weekend cutovers are what produce the rollbacks. The reference pipeline we build first is usually the system that hurts most, which here is normally process control systems.
Residency, latency and the Belgium question
Most Belgian deployments land in West Europe, which is physically in the Netherlands, or in France Central. Microsoft has announced a Belgian datacentre region, so check current availability before committing to an architecture that assumes in-country residency, and in the meantime document the EU-but-not-Belgian position for your data protection officer.
For La Louvière specifically, West Europe (Netherlands) is the normal primary with France Central (Paris) as the pair. Latency from La Louvière to that region is not usually a constraint for analytical workloads; the constraints that bite are feature availability per region and the processor chain your data protection officer will ask about.
- Belgium's federal structure means some reporting obligations sit with the regions (Flanders, Wallonia and Brussels) rather than nationally, which multiplies the datasets a single organisation must produce.
- Language legislation affects employee-facing systems: reporting used in employment contexts may need to be available in the language of the region where staff are employed.
- As the host of the EU institutions, Brussels has an unusual concentration of organisations working to EU-level reporting rules ahead of national implementation.
Where to go next
If you are weighing a platform decision for a La Louvière organisation, a workload assessment will answer it with evidence in two to three weeks, and the written recommendation is yours whatever you decide. Book a technical review.
We work across Belgium and deliver into neighbouring markets regularly. Nearby pages worth reading if your group has sites elsewhere: Genk, Seraing, Liège, Charleroi, Hasselt and Kortrijk.
Frequently asked questions
Is Microsoft Fabric worth it for a mid-sized La Louvière business?
Sometimes. The case for Fabric gets strong when more than one of these holds: import models under strain, the same data needed through both SQL and notebooks, governed self-service for several teams, retention that a desktop model cannot carry, or a Premium capacity already on the bill. If one analyst maintains three models that refresh in twenty minutes, a capacity is a cost without a return, and we will put that in writing.
How do we size a Fabric capacity without guessing?
By measuring. Each workload is profiled for compute shape and concurrency, mapped to capacity units, then validated on a trial capacity using your real data. You get a sizing model showing utilisation at launch, at double the users, and the point at which the next SKU becomes necessary.
Can Fabric coexist with our existing Synapse or Databricks platform?
Yes, and for most organisations that is the pragmatic path. Shortcuts and mirroring let Fabric read data in place, so you can serve analytics from Fabric without re-platforming the engineering layer. Start with the serving layer. Whether anything else moves is a question to answer with measurements from your own estate rather than in advance.
Where will our data physically sit?
Most Belgian deployments land in West Europe, which is physically in the Netherlands, or in France Central. Microsoft has announced a Belgian datacentre region, so check current availability before committing to an architecture that assumes in-country residency, and in the meantime document the EU-but-not-Belgian position for your data protection officer.
What does this cost for an organisation in La Louvière?
A workload assessment is a fixed fee in the low five figures. A landing zone plus a reference pipeline is typically a six to ten week engagement; full programmes run in quarterly increments and are quoted per phase. Platform licensing is separate and should be priced from the current Microsoft rate card for Belgium.
Who operates the platform afterwards?
Your team, with explicit layer ownership documented, or us under a managed service if you would rather not carry the on-call. Either way the runbooks, architecture decisions and monitoring configuration are yours from day one. La Louvière draws on the Mons, Charleroi and Brussels catchments, working in French.
Read next
Power BI consultancy in La Louvière: what it costs, what breaks and what to build first
A practical guide to Power BI consultancy in La Louvière: the definition traps in metals and materials and logistics and freight, honest licensing arithmetic, hire-versus-engage costs, and the compliance position under the EU GDPR and the Belgian Data Protection Act of 2018.
Data governance and GDPR for La Louvière organisations: controls that produce evidence
Practical data governance for analytics estates in La Louvière: what the Data Protection Authority (APD / GBA) expects under the EU GDPR and the Belgian Data Protection Act of 2018, the sector rules that bite locally, Microsoft Purview in practice, and a 90-day plan.
Microsoft Fabric: the build-or-wait decision, with the cases where the answer is wait
A decision framework for Microsoft Fabric adoption: the five conditions that justify a capacity, medallion architecture that works, capacity sizing from measured consumption, and the situations where staying on Power BI Pro is correct.