The owner of a 30-person company spends Monday morning doing exactly what they did three years ago: opening four files, asking the bookkeeper for one number, calling the warehouse manager for another, scribbling in a notebook, and leaving for a meeting with the feeling that they roughly know what's happening. "Roughly" is usually not enough — and a few hours of manual work every week is even less.
The term "owner dashboard" sounds like something out of a BI vendor's brochure. In practice it's a dozen numbers on one screen that nobody has to compile by hand, that refresh at least daily, and that are trusted enough to drive real decisions. Everything else — Power BI, DAX, data warehouses — are tools, not the goal.
What an owner dashboard is, and what it isn't
It isn't a management report in the sense a bank or investor expects. A management report is 20 pages, answers backward-looking questions, and goes out once a month. The owner dashboard is one page, answers the question "do we have a problem right now", and is opened several times a week — sometimes daily.
It also isn't an operational dashboard for the production manager or a sales funnel for the sales team. Those people need detail they can act on while working. The owner needs a synthesis that lets them not get involved in operations until something starts to burn. That's an entirely different level of aggregation.
In practice, a good owner dashboard fits on a single monitor (or an A4 sheet if the owner prefers paper), carries 10–15 numbers, and follows a simple rule: every number on it, when it goes red, should trigger a concrete action. If the owner sees red and doesn't know who to call, that number doesn't belong there.
Which numbers actually belong on Monday
It depends on the industry, but the skeleton is surprisingly repeatable. For trade, manufacturing, and B2B services in the 20–50 headcount range, this set works well:
- Sales — week-over-week, month-to-plan. Two numbers, not twenty charts. "€4.8m vs €4.2m" says more than twelve bars.
- Gross margin on sales. Revenue without margin is misleading — you can sell a lot and lose money. A rounded percentage, compared to last month, is enough.
- Cash balance + receivables overdue > 30 days. One is easy, the other usually hurts. If receivables grow faster than sales, that's already a signal.
- The three to five largest customers for the month. Not to pamper them — to notice when one drops off the leaderboard.
- One operational KPI from the industry. Manufacturing: line output, scrap, OTIF. Trade: fill rate, stock below reorder point. Services: team utilisation, backlog.
- One sentence from IT or operations. Not a number, just a short note — what broke, what's coming. Often the most valuable field on the dashboard.
Dates, currencies, and percentages stay in raw form, without pretty icons and gradients. The dashboard has to read in 30 seconds, not in a minute.
Where the data really lives, and why it hurts
The problem with the owner dashboard isn't that "we don't have BI". The problem is that data lives in five places, each in a different format, and nobody knows which numbers are "the real ones". Sales live in the ERP (often Comarch Optima, enova365, Subiekt, or Microsoft Dynamics in Polish SMEs), receivables in the ERP or a separate module, margin — depending on who set it up long ago — may live in the ERP, in the financial controller's Excel, or in the chief accountant's head.
With one of our clients — a 25-person wholesaler of building materials from central Poland, running on Comarch Optima — the Monday compilation took 2 hours a week (Optima export, two Excel sheets with ageing receivables, a phone call to the warehouse for stock below minimum, manual gluing of numbers into a slide). After three weeks of work — a pre-fill from Optima into Excel via a simple ODBC script, one ageing tab, one "comment of the week" field filled by the controller — the same set of numbers appears at 7:30 on Monday without anyone's involvement. Compilation time: 10 minutes (just the comment). Savings: roughly 95 hours a year, more than two full working weeks for the controller. And, more importantly — the numbers are the same every week, because they don't pass through anyone's manual "ah yes, I should also have added…".
Showcase investment in tools: zero. The owner reads the exact same Excel as before — it just fills itself in.
From Excel to Power BI — when which makes sense
The temptation to "start with Power BI" is huge, because it looks nice and modern. But for most SMEs it's the second step, not the first. The decision in one paragraph:
- Excel with automatic data pull (ODBC, Power Query, a simple script). Works for a dozen or so numbers, refreshes once a day, everyone in the company can open it. For 90% of SME owner dashboards it's enough for years. Cost: 2–5 days of work.
- Google Sheets with Apps Script. The alternative for firms in the Google ecosystem. Easier sharing, harder pivoting.
- Power BI / Looker Studio. Makes sense once the number of report consumers grows above 3–5 (directors, managers, the board) and each has different questions. Or when the data outgrows what Excel can reasonably hold (hundreds of thousands of order lines daily). Implementation cost: 2–8 weeks, depending on data quality.
- A dedicated data warehouse. Above 50 people, several data sources, and a real need to blend with external datasets. In a 20–50 person SME this is practically never the first move.
Rule of thumb: build it in Excel or Google Sheets, live with it for a year, see what actually clicked. Then decide on Power BI. The reverse order ends with a dashboard nobody looks at, because it shows the wrong things.
What not to put on it
Dashboards that don't survive six months share three traits. It's worth guarding against them up front:
- Vanity metrics. Follower counts, page views, likes. If you don't know what decision this number drives, don't put it on.
- On-the-fly figures from systems without a defined source of truth. If the same margin shows two different values in two reports, the dashboard loses credibility in a week. Better one agreed-upon number, even if imperfect.
- Too much granularity. Not 50 products — top 5. Not all customers — the biggest ones. Not every change — only variances above 10%. The owner won't scroll.
Fourth: avoid "colouring" numbers without concrete thresholds. "Red" without an answer to "so what" is visual noise.
How to start this week
You don't need a BI licence to begin sensibly:
- Step 1 — write down on paper 10 numbers you'd want to see on Monday. That takes 15 minutes and will expose which ones you don't actually know well.
- Step 2 — for each number, agree on one source of truth. ERP? Bank? Excel? A person? If the answer is "it depends", you've just found the first problem to solve.
- Step 3 — build the first version in Excel, by hand, for two weeks. You, not a vendor. Until you've filled it in yourself, you don't know what's hard about it.
- Step 4 — after two weeks, talk to someone who connects ERP to spreadsheets daily. What you've been doing by hand can usually be automated in a few days.
- Step 5 — only after six months should you think about Power BI. By then you'll know what belongs on it — which is the whole hard step.
The payback on this kind of dashboard — measured by the owner's reclaimed time and by catching problems earlier (e.g. a single customer's receivables ballooning) — usually shows up in the first quarter. Not because "the dashboard improves processes". Because the owner finally knows what's going on.
An owner dashboard isn't Power BI. It's 15 numbers that are trusted — and no more two-hour Monday mornings.