Every month, an SMB manager shows me their shared Excel sheet, saved as v_final_final_OK_v3.xlsx, and asks: "Power BI, Tableau, or do you build me something?"
This time, it's a good question. The wrong answer would be "it depends". The real answer is a 3-year total-cost calculation — and in 80% of SMB cases I see, custom is no longer the luxury choice we imagine.
Here are the figures, the cases, and the grid I apply when a client hesitates.
The 3 options in one sentence
| Option | Positioning |
|---|---|
| Power BI | The default Microsoft SaaS dashboard. Simple, integrated with Excel, cheaper than Tableau. |
| Tableau | The Rolls Royce of data viz. Powerful on large volumes, expensive, requires a dedicated BI profile. |
| Custom dashboard | A web app built for your business. Owned, no subscription, sized to fit. |
These three options solve the same problem — centralising your data to steer — but with radically different economic logic.
3-year comparison table
Figures for an 80-person SMB, 20 dashboard users, 5 data sources.
| Item | Power BI | Tableau | Custom |
|---|---|---|---|
| Initial setup | €3,000 - €8,000 | €15,000 - €30,000 | €8,000 - €25,000 |
| User licences | €10/user/month | €75/user/month | €0 |
| Infrastructure (hosting) | included | included | €50-150/month |
| Maintenance | €200-500/month | €500-1,500/month | €100-300/month |
| BI team training | €2,000-5,000 | €15,000-50,000 | €500-1,500 |
| 3-year cost (20 users) | ~€20,000 | ~€95,000 | ~€17,000 |
Note: the "custom" cost assumes you only build what's used — 6 to 8 targeted dashboards, not a full BI platform. Power BI assumes you already have Microsoft 365 (true for 70% of SMBs in 2026).
When Power BI wins
Power BI is the rational default choice for an SMB that already has Microsoft 365 and standard BI needs.
Signals that push toward Power BI:
- Team already comfortable with Excel.
- Need for 5 to 10 classic dashboards (sales, finance, ops) without strong business specificities.
- Moderate data volumes (less than 10 GB per dataset).
- No dedicated internal data profile, but at least one advanced Excel analyst.
- Tight budget, need to start within 6 weeks.
What it does well: native Excel/SharePoint connectors, well-built standard visuals, powerful DAX for calculations, fast deployment.
What it does badly: deep customisation quickly becomes painful (you fight the tool), pricing climbs mechanically with the user count, and the day Microsoft changes its pricing — it has happened — you take the hit.
When Tableau wins
Tableau is overpowered on visualisation and large volumes — but its pricing positioning reserves it for mid-caps and SMBs that genuinely need it.
Signals pushing toward Tableau:
- Volumes of dozens of GB per dataset.
- Advanced visualisation requirement (mapping, geospatial, deep interactions).
- Dedicated data team of at least 1 FTE, trained on the tool.
- Annual BI budget over €50k.
For 95% of SMBs of 10 to 100 people, Tableau is a disproportionate choice. You'll pay €75/user/month to use 20% of the tool. I've never recommended Tableau to an SMB under 150 people.
When custom wins
Three typical cases where I recommend building rather than subscribing:
1. You have business logic Power BI can't handle cleanly. A margin calculation crossing 4 sources with an allocation rule specific to your sector. An operational dashboard that must trigger an action (send an email, create a task) — Power BI doesn't do that, a custom dashboard does. That's exactly where my business applications become more profitable than a SaaS.
2. You have more than 15 permanent users. At 20 users, Power BI costs €2,400/year in licences. At 50 users, €6,000/year. At 100 users, €12,000/year. Custom pays back in 2 to 3 years at 50 users, in less than 1 year at 100 users.
3. You want to stay the owner. Business data is your asset. Putting it at Microsoft with a tool you don't control is a choice — not necessarily the right one. A custom dashboard means code yours, data yours, no forced price hikes.
Simple rule: below 15 users and without specific business logic, Power BI wins. Above 30 users or with real business logic, custom wins almost always over 3 years.
Concrete case — the executive cockpit
An industrial client, 120 employees, contacts me. They have 3 active Power BI dashboards for 18 months, paid €10/month × 22 users = €2,640/year. The manager wants a mobile cockpit (5 indicators) to steer while travelling.
Power BI does it — but badly. The mobile app is slow, the 5 indicators have to be split across 3 views, and the "executive" dashboard requires a Premium licence at €20/user/month.
Custom alternative: €6,800 in development, €40/month hosting, a dedicated mobile-first webapp with 5 indicators updated in real time from the same sources as Power BI (CRM + accounting + production APIs).
Over 3 years: custom €8,240 vs Power BI Premium €13,200. And above all, the manager has exactly the UX they want, not the one Microsoft imposes.
What systematically goes wrong
1. The dashboard built once and never reviewed. 60% of Power BI dashboards I audit haven't been updated in over 6 months. Metrics have shifted, sources too, but no one touches the dashboard anymore. Not a tool problem — a governance problem, and it shows up the same way on custom builds.
2. BI "instead of the spreadsheet". A dashboard isn't a prettier Excel. If the 5 indicators you watch fit in an Excel sheet with a pivot table, stay on Excel. BI becomes profitable when you have several sources, several users, and need fresh data.
3. The "we get Tableau to look serious" trap. I've seen two 40-person SMBs pay €30-40k/year in Tableau licences "because it's the reference". They were using 15% of the tool. The alarm signal is when no team member can edit a dashboard alone — you're paying for a Rolls to do the groceries.
Decision grid — 3 questions
Question 1 — How many permanent users?
- Under 10: Power BI if Microsoft 365 already in place, otherwise light custom (€5-8k).
- 10 to 30: Power BI in most cases. Custom if specific business logic.
- Over 30: custom wins almost always over 3 years.
Question 2 — Data volume and complexity?
- Simple data (1-3 sources, < 5 GB): Power BI or custom.
- Complex data (5+ sources, business rules): custom.
- Massive volumes (> 50 GB, critical performance): Tableau or custom with dedicated data architecture.
Question 3 — Do you want ownership?
- Yes, strong priority: custom.
- Doesn't matter: Power BI if standard, custom if specific.
Where to start
If you have an Excel dashboard starting to crack, or you're already paying BI licences you doubt, 30 minutes are enough to look at your case. I look at your sources, your uses, your volume — and tell you frankly if you should pay for a SaaS or if a targeted custom solution will cost you less.
If you want to see my price ranges before talking, they're public here.
To go further
- Related service: Custom dashboards
- Related articles: SaaS vs custom — the real math · When to go custom: 5 signals
