Flow Distribution

Definition

Flow Distribution is "the proportion of each flow item within a value stream, adjusted depending on the needs of each stream to maximize business value." Because the four flow items (features, defects, risks, debts) are mutually exclusive and collectively exhaustive, their distribution is a forced trade-off: increasing one share necessarily decreases the others. Rather than letting this ratio emerge as an accident of whoever shouts loudest in planning, the Flow Framework treats it as a dial that leadership and teams set on purpose, and re-set as the product's maturity, market position, or life-cycle stage changes.

In the Book

Kersten contrasts a new product's value stream — which should skew heavily toward Feature flow to hit launch, since there are few customers to generate defects and low public exposure to create risk work — against a legacy back-end service kept alive only to support other systems, whose flow distribution should shift almost entirely to defect fixes and risk reduction with minimal feature investment. He ties the dial explicitly to Geoffrey Moore's Zone Management (Incubation, Transformation, Performance, Productivity zones): each zone implies a different target distribution, and making that target explicit lets managers and teams share an understanding of what a given value stream is currently optimized for. The clearest illustration is Bill Gates's 2002 Trustworthy Computing memo at Microsoft, which the book reads as a company-wide flow-distribution reset toward risk and security work, later followed by a deliberate pivot back toward feature flow to fight the web-platform disruption threat. Chapter 6 uses the concept diagnostically again: the industry-wide rise in software-related automotive recalls is evidence that flow distribution across the automotive sector has drifted too far toward features relative to defects and risk, and needs to be consciously rebalanced.

Why It Matters

Flow Distribution names a decision that most organizations make unconsciously and retroactively — discovering their real priorities only by looking backward at what got shipped — and turns it into a forward-looking, adjustable setting tied explicitly to a system's current life-cycle stage. Any system with a fixed capacity and multiple categories of competing, non-fungible demand (a factory splitting output between new product lines and existing-customer support, a research group splitting time between new results and replication, a government agency splitting budget between new programs and enforcement) faces the identical tuning problem: the ratio should track the system's current strategic posture, not calcify around whatever it happened to be last quarter.