Power BI accounting dashboard: three routes to live figures
A Power BI accounting dashboard can be built three ways: the ledger API, a connector, or a reporting layer that consolidates first.
A Power BI accounting dashboard can be built in three ways: straight onto the API of your accounting system, through a ready-made connector, or on top of a reporting layer that consolidates and structures the figures first and then hands them to Power BI. The right route depends less on the tool than on how many administrations have to end up in one view.
For a controller this is a maintenance decision, not a technical one. Building the connection yourself is the cheapest way to start and the most expensive way to continue: you own the data model, you rebuild it per administration, and every change in the chart of accounts lands on your desk. With one legal entity and a clean ledger that is fine. With five entities, intercompany traffic and two currencies, you are recreating a consolidation engine in DAX.
What to settle before you build
- Number of administrations. One administration is one data stream. At five, every modelling choice multiplies by five, mistakes included.
- State of the chart of accounts. A ledger designed for statutory accounts has to be rebuilt for management information, and a visualisation tool is the wrong place to do that.
- Consolidation and intercompany. Eliminating internal positions, non-controlling interests and sub-consolidations are calculation rules, not visuals. Ask per route who maintains those rules.
- Budget and forecast. Power BI has no data entry. If the budget lives in Excel, you hold two versions of the truth and a manual step at every revision.
- Release and access rights. Decide whether management should watch live figures while the close is still running, with unmatched receivables and uninvoiced revenue.
- Ownership after go-live. A model only its author understands is a risk, in Excel or in DAX. Ask who maintains it two years from now.
Three routes to a Power BI accounting dashboard
| Route | What you build yourself | Suits |
|---|---|---|
| Ledger REST API | Authorisation, data model, history and all reporting logic in DAX | One administration, in-house BI capacity |
| Third-party connector | Nothing on the extraction side, everything on the reporting side | One or two administrations, standard ledger |
| Reporting layer before Power BI | The structure and consolidation setup, once | Groups with several entities or currencies |
The API route is the most open and the heaviest. You handle OAuth authorisation, pull ledger entries per division, respect the call limits the vendor imposes and build up history yourself. The bill arrives later: every new entity means another authorisation, another division and another slice of data model. The trade-off between wiring systems together yourself and putting a layer in between is covered in this article on interfacing financial systems.
The Exact Online Power BI connector question
In the Benelux these projects usually start on Exact Online, and the first search is for an Exact Online Power BI connector. Such connectors exist, and they do solve the extraction problem. What they leave untouched is the reporting problem: you inherit the supplier's field layout, so anything they skip is absent and anything they group differently has to be regrouped downstream. For a single profit and loss statement that is enough. For a consolidated balance sheet across divisions it rarely is, because a connector moves rows, not accounting logic.
What a Power BI accounting dashboard cannot do alone
Power BI is a modelling and visualisation layer, not a financial system. It has no notion of a chart of accounts, no elimination logic, no currency translation and no period close. That shows up at the second entity: internal revenue and costs have to be removed and kept visible where they fail to net to zero, as set out in the article on intercompany transactions. A ledger built for statutory reporting also needs its own subtotals, gross margin and EBIT percentage before it reads as management information, which is best fixed at the source. See splitting ledger accounts.
Foreign currency is the sharpest example. The profit and loss statement should be translated per period at that period's rate, the balance sheet at the closing rate. That methodological difference always produces a translation difference that has to be posted somewhere. Smartbooks performs the translation at trial balance level and posts the difference automatically, so the current year result on the balance sheet keeps matching the year-to-date profit and loss. A subsidiary with 30,000 USD of first-quarter result, translated monthly at 0.92, 0.91 and 0.93, totals 27,560 EUR; at the closing rate of 0.93 the year-to-date figure is 27,900 EUR. That 340 EUR gap is a posting, not a rounding error. In DAX it is a project of its own.
Finance KPI dashboard: which figures belong on it
A finance KPI dashboard works when every tile traces back to a ledger account or a documented formula. Gross margin, EBITDA, days sales outstanding, solvency and the cash position cover most management questions, each shown against budget and against the same period last year. Operational metrics such as headcount or units sold belong beside them, and that is where Power BI earns its place. What it needs from upstream is a stable account structure, a released set of figures and a budget in the same shape as the actuals. For the reporting discipline behind that, see best practices in reporting.
Where Smartbooks fits, and where it does not
Smartbooks is neither a Power BI replacement nor a connector. It is the layer in between: connecting Exact Online is self-service, as it is for Twinfield, AFAS, QuickBooks and Odoo, while Exact Globe is set up by the team. Per entity you record which administration feeds it and which business dimension applies, cost centres or cost units. Data refreshes several times per hour and the integration screen shows when each source last synchronised. On the output side there are API connections towards BI platforms and data warehouses, plus an Excel add-in for anyone who prefers a worksheet. A feature-level comparison sits in Smartbooks vs Power BI.
It fits when there is more than one administration, when intercompany traffic or a second currency is involved, or when budget and forecast have to sit beside actuals without a spreadsheet in between. It does not fit a single entity with a clean ledger, no consolidation question and a Power BI specialist already on the payroll: a connector is cheaper there. It also does not fit if your main goal is blending operational sources such as web shop or time-tracking data.
How the switch runs
- Create a company per entity, linked to an external data source, and pick the accounting system. During setup you select the division and one business dimension; only one can be active.
- The environment goes live with history already loaded and a first set of reports in place. From there you walk through reporting periods, reporting structure, dimensions and KPIs.
- Create a consolidation company and fill in the intercompany matrix: for each pair of entities, which debtor or creditor represents the other. Then flag the elimination accounts.
- Check the Intercompany Difference line on both statements. Where an amount appears, right-click it and choose Show intercompany transactions to reach the cause. The wider sequence is detailed in mastering financial consolidation.
- Point Power BI at the structured layer instead of at raw ledger entries. Structure, consolidation and budget comparison are already inside it.
Pitfalls when choosing
- Starting at the visual. A polished chart on a skewed ledger is an expensive way to spread wrong numbers faster.
- Forgetting the second entity. A pilot on one administration says little about the work five will cost. Test the route against next year's group.
- Treating live as quality. Real time only helps after the close. Agree which figures are released and which are still being worked on.
- Leaving budget outside the model. A dashboard without budget and forecast is a mirror, not a steering instrument. See management reporting.
Frequently asked questions
Can Power BI read an accounting system directly?
Yes, through its REST API. You then handle authorisation, division selection and the data model yourself, and you stay within the vendor's call limits. For one administration that is manageable; from the second onwards it grows into a data model project.
How often does a Power BI accounting dashboard refresh?
That depends on the route. With your own connection you set the refresh schedule yourself, within the limits of the source. Smartbooks pulls data several times per hour and shows per source when the last synchronisation succeeded, so you know what you are looking at.
Can several administrations sit in one dashboard?
Technically yes, and that is where the real problem starts. Adding up is not consolidating: internal positions have to come out, non-controlling interests shown separately and foreign currencies translated. Build that in DAX and you have built a consolidation system.
What a Power BI accounting dashboard costs is rarely a licence question; the weight sits in build hours, the data model and the consolidation logic. Weigh that one-off build against the annual maintenance load. To see consolidated figures before they reach a dashboard, visit the consolidation software page or the one on the Exact Online connection, and book a demo to find out in half an hour which of the three routes suits your group.
Related articles
AI CFO software: what to look for before you buy
November 18, 2026
How to Calculate Goodwill in an Acquisition
November 4, 2026
How to Prepare Financial Statements: Components, Steps and Deadlines
October 21, 2026
Intercompany Transactions: Eliminations, Reconciliation and Pitfalls
October 7, 2026
Indirect Method Cash Flow Statement: Format and Example
October 16, 2026
Absorption costing: formula, example and variable costing compared
October 19, 2026