Getting reporting out of Excel without a monumental project
Published · updated
Swiss exporting SMEs have spent a year with little control over their external conditions. After several changes to the tariff framework, SECO reported on 24 July 2026 a 10% additional US tariff and a general total rate of 12.5% on affected Swiss goods, next to exemptions and duties specific to some sectors. The negotiations and the measures may change again. A management team can’t choose the tariff or the exchange rate. But it can choose what it costs to produce its own information. In many companies of 10 to 100 people that I meet, that cost hides under “monthly reporting”: two or three days of copying data from the ERP to the accounting system to a forest of spreadsheets.
The real problem is not Excel
Excel is an excellent analysis tool. The problem is asking it to act as a database: several versions of the same file, formulas only one person understands, figures that differ depending on where you look. When the management meeting opens with twenty minutes of debate about which number is the right one, that isn’t a visualisation problem. It’s a data governance problem.
The external gap is documented. According to localsearch’s 2025 SME study, 36% of Swiss SMEs don’t have their own website, with the micro-enterprises weighing heavily on the result. That statistic says nothing directly about how mature internal reporting is. It’s a reminder to measure the gap rather than assume it.
The step-by-step approach
Making your figures reliable needs neither a new ERP nor a six-figure project. The approach has four steps, and each step is useful on its own:
- Inventory the sources: finance, ERP, CRM, point of sale and timesheets. Who retypes what, where and how often? Half a day may cover a simple flow; a badly documented or multi-company scope will take longer.
- Centralise and govern: feed a common store automatically from the systems you already have. PostgreSQL is often the economical option, but the ERP, the available skills, the support and the hosting all affect the choice. One database removes the discrepancies only when every metric has a definition, an owner, a lineage, a reconciliation rule and quality checks.
- Automate one flow first: the monthly report that costs the most time. Not ten dashboards. One report, correct, on time, every month, with nobody retyping anything. Trust gets built there.
- Deliver the result: only then choose a dashboard tool: Metabase self-hosted, for example, or Power BI in an established Microsoft environment. The display layer is now mainly a delivery choice, although the permissions, the definitions and the lineage stay architectural.
And Excel? It stays, for what it does well: one-off analysis, simulation, exploratory work. What disappears is the monthly retyping, and the doubt.
The detail that matters in Switzerland
A central database can run on your infrastructure or with a provider you select. For sales, margin and customer data, a Swiss option deserves evaluation. At a time when depending on US services has become a political question in the federal parliament, the location, the jurisdiction, the administrative access and the exit should be decided while you design the architecture. That choice isn’t always free. But it’s often more expensive to correct afterwards.
When margins tighten, every hour of retyping is an hour of margin lost. The lever isn’t spectacular. But it’s entirely in your hands.
Give every metric a contract
Before you build the dashboard, write down each metric’s definition, its unit and grain, its source system, its owner, how often it refreshes, its reconciliation test, and the confidentiality class it carries. That compact contract, not a database alone, creates a shared source of truth.
The SECO figures, and their date
Tariffs change from one quarter to the next. The timelines, the technologies and the savings in this article are starting assumptions to measure in the real environment.
SECO data taken on 14 August 2026.