The operating cadence of the discipline. Four phases, four owners, four cadences. The loop never closes — each turn is sharper than the last.
The discipline runs as a four-phase loop. Drag the nodes, then click any phase to open its inputs, outputs, cadence, and named instruments.
Scan is the discipline of reading early, ambiguous indicators that may precede change. A healthy signal portfolio is balanced across domains and time horizons, and deliberately includes signals that could disconfirm your thesis. The skill is separating signal from noise without filtering out the weak ones that matter most.
Phase one of the Predictive Planning Loop.
Build a live signal layer from inside and outside the organization. Mix system data, paid feeds, public data, and human signal — what your sales team is hearing, what your suppliers are quietly preparing for, what your competitors' job postings reveal. The Scan is continuous, not episodic. The discipline is in the cadence, not the dashboard.
Phase two of the Predictive Planning Loop.
Translate the signal layer into a small set of pressure-tested scenarios. Three, named distinctly, living on different cleavages. Each scenario has an indicator set — what we would expect to see in the next sixty days if this scenario is the one converging. A scenario without indicators is a story without consequences.
Phase three of the Predictive Planning Loop.
Resource decisions under uncertainty, sized to scenario, hedged where the cost of being wrong is structural. The Stake Sizing Matrix records four fields — Scenario, Size, Reversibility, Conviction; Probe, Hedge, Build, and Bet are the Institute's working names for where a stake lands. The discipline is making real commitments instead of producing another option set.
Phase four of the Predictive Planning Loop.
Continuous correction as scenarios converge or collapse. The Scan loop continues, but indicators are now tracked against active scenarios. When an indicator triggers, the team reviews the implication for the commitment. Adjust, don't restart. Steer is incremental correction, not periodic relitigation.
The phases run in sequence, but they do not stop and restart. The Scan never pauses; Steer is constant. Story and Stake punctuate the cycle on their own cadences.
The waterfall mental model is the most common installation failure. A team finishes Scan, finishes Story, finishes Stake, and then declares the cycle done. What they have done is build a quarterly planning ritual with new vocabulary. The discipline never closes the loop because the operating environment never stops moving.
In practice, the four phases overlap. The Signal Steward is already feeding Wednesday's intake into next month's Story read while last quarter's Stake is still being steered. The diagram looks like four points on a circle for the same reason a clock face does — they are all true at once. Position on the circle is a function of cadence, not phase completion.
The accent arc on the diagram — Steer returning to Scan — is where the discipline pays for itself. Every Steer cycle re-seeds the next Scan with what the world actually did, which indicators triggered, which scenarios collapsed. The fifth revolution of the loop is not the same loop running again. It is a sharper instrument with a memory.
The person who runs the Loop on behalf of the leadership team is the Predictionist.
