Power BI shows your figures.
You still have to have them.
If your organisation already runs Power BI, the question is fair: why another tool? The short answer is that these are different trades. A business intelligence tool reports data you own. A material flow analysis starts where the data runs out — and there the figures have to be computed, not displayed.
The dividing line
On a dashboard, every figure on screen exists somewhere in a database. In a flow study it is the other way round: you know a national output, two yield coefficients and an import tonnage, and three quarters of the rest is missing. The work is to derive the missing values by requiring that the accounts add up — what enters a node must leave it, a parent must equal the sum of its children, a year-end stock opens the next year.
That is constrained optimisation, not aggregation. Power BI offers no solver: DAX can
fit a line by ordinary least squares (LINEST), without constraints and
without weighting by uncertainty. To reconcile, you must compute elsewhere and
import the result.
What each one does
Power BI facts verified against Microsoft documentation, August 2026.
| Power BI | OpenSankey / MFASankey | |
|---|---|---|
| Starting point | Data you have | Data that is incomplete, redundant, sometimes contradictory |
| Computing missing values | No solver; LINEST does unconstrained least squares | Constrained reconciliation, uncertainty propagated |
| Mass balance | Not a concept — to be rebuilt in DAX, with no guarantee | The founding equation: at every node, in = out |
| Sankey | No native visual: a custom one, flat source→destination pairs | The diagram is the model: nodes, flows, stocks, hierarchies |
| Node hierarchies | Absent from the Sankey visual; cycles resolved by duplicating a node | Collapse, expand, and reconcile across two levels at once |
| Reading a published result | A Pro licence at $14/user/month, unless on an F64 capacity or above | A public link — no account, no licence for the reader |
| Truly public publishing | “Publish to web” exposes the underlying model, not just what is shown | Only the selected views make it into the render |
| After the subscription | 30-day grace period, throttled on day 31, all requests refused on day 91 | Static files, readable and archivable indefinitely |
| Home ground | Sales, HR, finance, operations — thousands of connectors | Material, energy and money flows of a territory or a value chain |
Publishing: two costly misunderstandings
The first is about confidentiality. Power BI's public publishing mechanism, “Publish to web”, makes a report readable without an account — which is useful, and it is the only native way to reach a general audience. But Microsoft attaches an unambiguous warning to it: anyone can access the underlying data of your model, even if the report does not display it. A published report ships the model, not just the view.
The same trap exists with audiences, the feature that shows different pages to different groups: the documentation states that hidden content stays reachable to anyone holding a direct link. Hiding is not protecting.
We hold the opposite rule, and it strikes us as the only tenable one: restricting means not shipping. Filtering happens at render time — a published site contains only the selected views. What is not meant for an audience is not in the file that audience downloads.
The second misunderstanding is about time. A flow study is a deliverable that gets cited in a report, reopened by a partner three years later, archived by a public body. A Power BI report lives as long as the subscription: thirty days of grace, then throttled access, then refusal. Frozen exports lose interactivity, and an uncertified custom visual — potentially the Sankey itself — is not rendered in PDF export.
Our published diagrams are static files. No service to maintain, no licence to read them, no account to create. They will stay readable whatever becomes of us.
What about using both?
That is the most common scenario, and the most sensible. Keep Power BI for what it does better than anyone: pulling your information systems together, tracking indicators, serving dashboards to the whole organisation.
Leave us the upstream — the part where the data runs out. We reconcile the mass balance, propagate uncertainty, and produce a consistent, traceable set of values in which the collected figure stays visible next to the computed one. That result is a dataset like any other: it feeds your dashboards as readily as it feeds our diagrams.
So the question is not “Power BI or you”, but “who produces the figure”. Once it exists and it is right, display it wherever you like.
The other comparisons
Power BI is not a competitor — it is a neighbour. If you are looking to replace a flow analysis or Sankey drawing tool, the useful comparisons are elsewhere: STAN, e!Sankey and SankeyMATIC.