Business intelligence
Some links on this page are affiliate links: we may earn a commission at no extra cost to you. Our approach
Business intelligence answers a concrete question: how do you know in real time whether the company is doing well, and where it is losing money? It turns scattered data into reliable dashboards that the right people can consult at the right time. ## BI starts with a definition, not a tool Before choosing Power BI or Looker Studio, you must agree on what you measure and how. Does revenue mean the amount invoiced or the amount received? Does margin include returns? These questions are settled in workshops, with the people who would use the reports. ## Levels of maturity Most SMEs go through three stages: manual tracking in Excel, then dashboards connected directly to the sources, then a data warehouse as sources and volumes grow. Each stage has its purpose, and jumping straight to a complex architecture costs a lot without bringing immediate value. ## From dashboard to recommendation A good dashboard also says what to do: alert on falling margin, fast-moving stock or drifting acquisition cost. With clean historical data, AI models can then go further and propose forecasts or segmentations. ## An example journey A distribution company builds a sales dashboard by shop and by product family, connected to its till software and a stock spreadsheet. The manager sees every Monday which products sell poorly and which shops are struggling, without asking the accountant for an export. The salesperson has a view by customer, fed by the CRM and invoices.When this guide is for you
- Every monthly report is rebuilt by hand by copying data into Excel
- You cannot tell which products, customers or branches are profitable without checking several files
- The sales, marketing and accounting figures do not match each other
- You want to track your indicators every week without waiting for an analyst
- You are preparing a management committee that asks for reliable figures comparable from month to month
Step by step
- Defining key indicators — Choose 5 to 10 indicators that really steer the business: margin per product, acquisition cost, revenue per branch, payment delay. Define for each one the formula, the source and the refresh frequency.
- Collecting and preparing data — Gather the sources (ERP, till, e-commerce, Google Ads, CRM) and clean them: duplicates, dates, currencies, categories. Many BI projects fail here, before the first chart.
- Modelling and warehouse — Organise data into fact and dimension tables, or rely on a managed warehouse for large volumes. Clear modelling makes dashboards fast and easy to maintain.
- Creating dashboards — Build views by use: management, sales, marketing, finance. Each dashboard answers a precise question, with simple filters and alerts on significant gaps.
- Deployment and adoption — Train users, define access rights and plan a monthly review of indicators. A BI that is not consulted every week is not worth its cost.
What it costs (in MAD)
| Service | Indicative price | Good to know |
|---|---|---|
| Scoping and indicator definition | 10,000 – 30,000 MAD (≈ 900 – 2,700 €) | Workshops with business teams and an indicator dictionary |
| Data preparation and connection | 15,000 – 60,000 MAD (≈ 1,350 – 5,400 €) | Depending on the number of sources and data quality |
| Dashboards (per theme) | 5,000 – 20,000 MAD (≈ 450 – 1,800 €) | Sales, marketing, finance, operations |
| Power BI Pro licence | 110 – 130 MAD / user / month (≈ 10 – 12 €) | Per user, billed in USD by Microsoft |
| Looker Studio | 0 – 8,000 MAD / month (≈ 0 – 720 €) | Free with Google connectors, paid for custom connectors or large volumes |
| Self-hosted Metabase or Superset | 0 – 1,500 MAD / month (≈ 0 – 135 €) | Hosting cost; commercial licence if needed |
| Maintenance and enhancements | 3,000 – 12,000 MAD / month (≈ 270 – 1,080 €) | Fixes, new indicators, support |
Indicative ranges for the Moroccan market in 2026, excluding tax. Prices vary with scope and provider.
Mistakes to avoid
- Building dashboards without first defining indicators and their formulas
- Ignoring source data quality: a wrong figure presented neatly is still wrong
- Multiplying charts and pages until nobody knows which version is right
- Delivering a BI without training or access: the tool is abandoned within weeks
- Remaining dependent on a single provider for the slightest report change
Frequently asked questions
Power BI or Looker Studio in 2026?
Power BI is more powerful for modelling and better integrated with the Microsoft ecosystem, but each user needs a licence. Looker Studio is free and perfect with GA4 and Google Ads, with limited performance on large volumes.
Do I need a data warehouse for BI?
Not always. For a few sources and modest volumes, a direct connection to the database or spreadsheets is enough. A warehouse becomes useful when you need to cross many sources, keep a history or handle large volumes.
How long before a first useful dashboard?
Allow 4 to 8 weeks for initial steering (one or two sources, 5 to 10 indicators). The timeline depends mainly on data availability and indicator clarity, not on the tool.
Is BI only for large companies?
No. An SME with a few data sources and well-defined indicators can set up simple steering on a moderate budget. It is large companies that have more sources and governance constraints.
Is an Excel dashboard not enough?
For a simple activity, yes. As soon as several people work on the same file, data comes from several sources or an automatic history is needed, a BI tool avoids errors and time wasted re-stitching figures.
Updated on
Related guides
Version française: /services/data-ia/business-intelligence/ · AR · NL