Steering by assumptions

Why assumptions?

A business plan is not a prediction: it is a chain of assumptions — about the market, prices, costs, timelines, adoption. The numbers it produces are merely the arithmetic consequences of those bets. As long as the assumptions stay implicit, the plan looks solid; that is precisely where portfolios drift.

When a project fails, it is rarely the calculations that were wrong. It is assumptions that were never written down, never confronted with data, never re-examined when the context changed. Conversely, a team that knows at any moment what has to be true for its plan to work — and where each of those conditions stands — can make decisions quickly and without drama.

At the start of any exploration there is a question to answer — sometimes reduced to a single assumption.

The three phases of the method

1. Surface

We start by extracting the assumptions from the plan: the ones that appear in the financial model, and above all the ones that don't. Each is formulated in falsifiable terms — "the market will accept this price", "going live will take six months", "this channel will generate 30 % of sales" — then ranked by its impact on the outcome and its degree of uncertainty.

2. Test

Each critical assumption is then confronted with the available data: yours (sales, operations, past projects) and the market's. Some can be checked in a spreadsheet in a day; others call for a pilot, a study or a prototype. What matters is separating what is proven from what is hoped — and spending validation effort where the stakes justify it.

3. Track

Finally, assumptions are tracked over time, the way a budget is. At every project or portfolio review, the question is not only "how is the schedule?" but "do our assumptions still hold?". Decisions to continue, pivot or stop become factual, not political.

In practice

The method applies to three types of engagement:

• Project portfolio review — a map of your projects and their assumptions, explicit selection criteria, and a review governance that lets you arbitrate continuously rather than once a year.
• Business plans and investment decisions — a financial model where every number traces back to an identified assumption, and scenarios that show what happens when an assumption breaks.
• AI framing and implementation — every use case is treated as a value hypothesis to test: an instrumented pilot, success criteria defined before starting, and a clear decision at the end.

Where this approach comes from

Antheos Data comes from data: years spent organizing, analyzing and making sense of data across the healthcare ecosystem, from the pharmaceutical industry to the pharmacy counter. That experience taught us one thing: data is not there to decorate slides — it is there to test ideas.

That bridge between analytical rigor and investment decisions is what makes us different: we can build the financial model and the data infrastructure that will confront it with reality.