Decision infrastructure

The systems, ownership and thresholds that determine how a company moves from information to a decision. Dashboards sit one layer below it, as reporting infrastructure.

The part that costs money

Every company already has decision infrastructure. Almost none of it was designed. The threshold exists as a habit, the ownership exists as whoever spoke last, and the escalation path exists as a direct message to the founder at nine on a Sunday. Undesigned infrastructure still determines outcomes — it just does it inconsistently, and nobody can point at the part that failed.

The layer below it is reporting infrastructure: warehouse, pipeline, transformation, BI tool, dashboard. That stack answers what happened. None of it holds a threshold, carries an owner, or contains a rule about what happens when a number crosses a line. This is why a company can spend three years and a large budget upgrading the reporting layer and find that decisions did not get faster — nothing in that project touched the layer where decisions live.

Three components make it real: thresholds (what size of change is worth a decision), ownership (whose call it is before the meeting starts), and a review loop (when we find out whether the call was right). Most companies can produce none of the three in writing for their ten most repeated decisions, which is a more useful finding than any dashboard those decisions were made from.

What it looks like

A company reforecasts when the CFO starts feeling uneasy. There is no written trigger for what pipeline movement causes a reforecast, no named owner for the call before the meeting begins, and no record of whether the last reforecast turned out to be right. The warehouse is excellent, current to the hour, and modelled properly. The decision infrastructure is a feeling.

How you would actually measure this

What it is not

Related terms

Frequently asked questions

What is decision infrastructure?

The systems, ownership and thresholds that determine how a company moves from information to a decision. Dashboards sit one layer below it, as reporting infrastructure.

How is decision infrastructure different from a data stack?

The data stack is reporting infrastructure: it establishes what happened and presents it. Decision infrastructure sits one layer above and holds the thresholds, the ownership and the review loop — what change is worth a decision, whose call it is, and when we learn whether it was right. Upgrading the reporting layer does not create the decision layer, which is why better dashboards so rarely produce faster decisions.

What are the components of decision infrastructure?

Three: thresholds, which define what magnitude of change deserves a decision; ownership, which names who holds the call before the meeting starts; and the review loop, which records the outcome so the next decision of that type is better informed. A company that can write all three down for its recurring decisions has decision infrastructure. One that cannot has habits.

Do small companies need decision infrastructure?

A five-person company’s decision infrastructure is the founder’s attention, and that works genuinely well until it does not. The failure point is roughly where the founder stops being present for every decision — usually somewhere between twenty and forty people, or earlier if the company runs across several systems and time zones. Building it before that point is cheap; building it after a bad quarter is not.