If you still run your sales, stock or margin reporting in a workbook that three people email back and forth, the real question is not whether Excel is a bad tool. It is this: at what point does the time you spend producing the report cost more than a tool that would produce it on its own?
This article is not selling you a dashboard. It helps you decide whether to leave the spreadsheet in 2026, with which tool, and when that change actually starts paying for itself — before you ask anyone for a quote.
You have three honest options: structure your Excel better, connect a standard BI tool (Power BI, Looker Studio) to your data, or have a custom dashboard built. For a Moroccan SME of 10 to 200 people, they should be examined in that order, not the reverse.
Why so many Moroccan SMEs stay on spreadsheets too long
Because the cost of Excel is invisible. It never appears on an invoice: it is paid in controller hours, in meetings spent arguing about the number instead of the decision, in month-end closings that slip by three days. The Microsoft 365 licence is already paid for, everyone knows the tool, nobody has to ask for budget. Standing still looks free.
Second reason: whoever built the file becomes irreplaceable. Nested formulas, hidden tabs and lookup chains form a system nobody else understands. The week that person is on leave — or leaves the company — reporting stops. That is an operational risk, not an IT detail.
Third and more legitimate reason: many managers have already paid for an ERP module nobody ever used. They are wary, and they are right to be. But often the real blocker is not the tool, it is the data. If your customers exist under three spellings and your product codes in two formats, no tool will save you. That is exactly what a data quality audit before a BI or AI project is designed to find.
Excel: how far it holds, and where it actually breaks
Let us be fair: Excel holds up very well in one specific case. A single data source, one person editing the file, a monthly rhythm, a few thousand rows, no need for fine history, no need to read it from the field. In that situation, replacing Excel is money with no return. Tidy the workbook, document the formulas, move on.
Where it breaks, the symptoms are always the same, and you will recognise them:
- The file exists in five versions, one of them named something like final_v3_fixed. Nobody knows which one is authoritative.
- Every month someone hand-pastes exports: the ERP, the till, the CMI statement, the sales team’s own sheet. One column shifted and the total is wrong, with no warning.
- A formula breaks silently. Nobody notices before the management meeting.
- The workbook is 40 MB and takes two minutes to open.
- There is no audit trail: you cannot tell who changed what, or rebuild last month’s picture.
- Your sales director cannot check anything from a phone between two appointments.
The real breaking point is not a row count. It is the moment several people have to produce or read the same number, regularly, from more than one source.
Power BI or Looker Studio: what these tools genuinely change
Four things change. Data loading becomes automatic, so nobody copies anything. Each metric is defined once, inside the model. History is kept. And access is managed by role — the rep sees their region, the CEO sees everything.
Power BI is the default choice if you already live in Microsoft 365 and your data sits in a SQL database, an on-premise ERP or well-structured files. It models seriously, handles row-level security, and reaches internal databases through a gateway. The trade-off is a per-viewer licence, roughly 130 to 160 MAD per person per month depending on your contract and the exchange rate — check Microsoft’s official price list when you decide, because it moves.
Looker Studio is free at its base tier and excellent for marketing and web data: Google Analytics 4, Google Ads, Google Sheets, a clean export. For multi-source commercial or financial reporting with real calculation rules and granular permissions, it hits its limits quickly.
What neither tool fixes: dirty data, no unique customer identifier, a source nobody trusts. A BI tool pointed at a mess produces a beautiful wrong chart, faster than before.
Custom BI: when your complexity justifies leaving standard tools
Custom BI is not the expensive edition of Power BI. It is a different kind of product, and only specific reasons justify it:
- Your calculation rules cannot be expressed in a dashboard: cost per production batch, true margin including customs duty, currency and landed costs, indirect cost allocation through a formula only your business uses.
- You need to write, not only read: enter a forecast, approve a budget, correct a record, trigger a workflow. A BI tool is a mirror, not a form.
- You publish outside the company: one space per client or franchisee, with strict permissions. Per-user licensing stops making sense.
- You pass 80 to 100 viewers: cumulative licence fees eventually exceed the cost of building something.
- You have ten to fifteen sources to reconcile with matching logic specific to your industry.
The decision logic is the same one described in our comparison of SaaS versus custom software: build only what makes you different, buy the rest. Hybrids are common too — a custom data pipeline feeding a standard Power BI report.
Excel vs Power BI: the 2026 budget and timeline comparison
Orders of magnitude seen in Morocco, unusual cases excluded. These are reference points for a decision, not a quote.
- Structured Excel: no extra licence. Five to fifteen days of work to normalise sources, freeze definitions and document them. Around 10,000 to 30,000 MAD with outside help. Timeline: 2 to 4 weeks.
- Power BI or Looker Studio, properly deployed: 30,000 to 90,000 MAD for two to four dashboards that are genuinely fed with live data, plus annual licences. Timeline: 4 to 8 weeks.
- Custom BI: 150,000 to 450,000 MAD depending on the number of sources and the complexity of the rules, with hosting of 500 to 3,000 MAD per month. Timeline: 3 to 6 months.
- Maintenance and changes: budget 10 to 20% of the initial cost per year in all three cases. An unmaintained dashboard dies within eighteen months.
For a line-by-line breakdown, our BI dashboard cost guide for Morocco decomposes every item. The payback maths itself is simple: count the monthly hours spent producing reports, multiply by the loaded hourly cost, add what your last decision made on a wrong number cost you. Compare over twenty-four months.
The 5 signs it is time to change tools
- You spend more than two days a month building the report, before any analysis happens.
- Two people quote two different figures for the same metric, and the meeting is spent deciding which one is right.
- Decisions rest on data that is more than two weeks old, because the file is only ready on the 15th.
- Exactly one person knows how the workbook works.
- You hand-paste exports from at least three different systems.
Practical rule: three signs out of five, a standard BI tool pays for itself. Three out of five plus one of the situations in the previous section (business calculation rules, data entry, external publishing) — look at custom. Zero or one sign: keep Excel and spend the money on cleaning your data instead.
CNDP and your data: what changes depending on where it is processed
As soon as your dashboard contains personal data — customer names, contacts, HR records, payroll, absences — you are processing personal data, which falls under Law 09-08 and CNDP formalities. Spreadsheets escaped this in practice, not in law; a BI tool makes the processing visible and permanent.
Two points deserve a conscious decision. First, where your data is hosted: Power BI and Looker Studio keep your models in data centres outside Morocco, which is an international transfer and comes with its own formalities. Second, who sees what, which you handle through role-based permissions.
Three habits make all of this far easier: push only aggregated data into the dashboard where you can, pseudonymise what you can (an ID rather than a name), and write down your purposes and retention periods before you publish. Our CNDP Law 09-08 compliance checklist walks through the process. For your own situation, check the current forms and obligations on the CNDP website: we give you reference points, not legal advice.
How to migrate without breaking everything: a 3-step method
1. Freeze your definitions before choosing a tool. Take five to ten metrics, not thirty, and write down exactly what each one means. Is revenue net of tax, recognised at invoicing or at collection? How many months makes a customer active? That single page is the highest-return step of the whole project, and it costs nothing but meeting time.
2. Run a pilot on one perimeter. One domain, one team, four to six weeks, running alongside Excel. The point of the pilot is not to look good: it is to compare both systems line by line and explain every gap. As long as one gap is unexplained, you do not switch. This is also where you discover the true quality of your sources.
3. Switch, then wean people off. Automate the data loading, take the old file out of circulation — otherwise it will survive in parallel for two years — train people on their own screen rather than in a classroom, and appoint an internal data owner. Without that owner, your new dashboard becomes the new Excel within eighteen months. That governance is part of what we frame in our data and digital transformation services.
What now?
Arrowlancer works on the data side: ingestion pipelines, data quality, Power BI and Looker Studio dashboards, and custom BI development when your business rules genuinely require it. We always start by looking at your current files and sources, because that is where the cheapest option for you gets decided — including the option of staying on Excel.
Describe your current reporting through our contact form and we will tell you plainly which of the three options holds up.