Marginal Reallocation Review¶

Recurring review — instantiates Pareto Focus

Periodically asks whether the next unit of effort still earns its keep on the current few, or should move to the next tier, the tail, or elsewhere.

A Marginal Reallocation Review is the recurring ritual that keeps a focus decision from ossifying. Its distinguishing feature is that it reasons at the margin, not the total: it does not ask "are these still our biggest contributors" — they usually are — but "does the next unit of effort on them still beat the next unit spent elsewhere?" Because focused effort improves the vital few and the returns from piling on more eventually flatten, even the top contributor should surrender the next increment once its marginal return drops below an alternative's. Run on a cadence, and re-measuring the distribution each time, the review re-opens the allocation the rest of the archetype tends to freeze.

Example¶

A growth team has poured spend into its two best-performing acquisition channels for two straight quarters. The quarterly marginal reallocation review does not relitigate which channels are best — it asks whether the next $10k on them still beats the next $10k somewhere else. The leaders are saturating: cost-per-acquisition is climbing as spend scales, a textbook case of diminishing returns. So the review shifts the marginal dollar to a mid-tier channel and to retention, without abandoning the leaders that still carry the base. The point is not to declare new winners but to place the increment where it now earns the most.

How it works¶

The review turns on the margin rather than the average. Each cycle it re-measures the current distribution — yesterday's top driver may have been improved or saturated — and then tests, for the current critical few, whether the next increment of effort still returns more there than at the best alternative destination. When it does not, the increment moves. The recurrence is essential: a one-time analysis cannot see saturation that only appears after months of concentrated effort.

Tuning parameters¶

Review cadence — how often the review runs. Too frequent and the allocation thrashes on noise; too rare and effort keeps piling onto a saturated few long after it stopped paying.
Marginal-return threshold — how much more the next unit must earn to justify moving it. A tight threshold makes reallocation hair-trigger; a loose one makes it sticky.
Reallocation destinations — the menu of places the increment can go (the next tier, tail protection, a new bet). A narrow menu just doubles down on what is already known.
Increment size — how large a chunk of effort is reconsidered each cycle, which sets how granular the rebalancing can be.

When it helps, and when it misleads¶

Its strength is that it stops the archetype's most common failure — over-serving a winning few well past the point of diminishing returns — and keeps the focus rule a live decision rather than a frozen 80/20 cut.^[1]

Its failure modes are reallocating on noise (a single bad month is not saturation) and, if the destination menu omits the tail, simply chasing the next-hottest thing while the long tail still goes unprotected. Its classic misuse is running it to rubber-stamp existing spend — "reviewed, no change" — without actually computing the marginal return against a real alternative. The discipline that keeps it honest is to compare marginal returns against concrete alternatives, deliberately including tail protection, and to smooth over normal variation before acting.

How it implements the components¶

remeasurement_loop — it re-runs the ranking each cycle so effort tracks the current distribution rather than the one that justified the original focus.
diminishing_returns_trigger — its core test is whether the next increment on the current few has fallen below the return available elsewhere, which is the signal to move effort on.

It does not compute the concentration curve or find the knee — that is the Cumulative Contribution Curve, whose output it re-reads each cycle; it does not establish the service tiers it reallocates across — that is the Tiered Support Model; and it does not watch the tail for harm — that is the Long-Tail Monitor; it only decides whether the next increment should go there.

Instantiates: Pareto Focus — keeps the focus allocation current as concentrated effort changes the distribution it was based on.
Consumes: Cumulative Contribution Curve — supplies the refreshed distribution each review reads.
Sibling mechanisms: Cumulative Contribution Curve · Long-Tail Monitor · Tiered Support Model · Top-Driver Analysis · Top-Cost-Source Intervention · Pareto Chart · Defect-Cause Prioritization · High-Risk Targeting List · Key Account List

Notes¶

This is where Pareto Focus borrows from its neighbor, Marginal Reallocation. A static 80/20 cut answers "where is the mass" once; this review answers "where should the next unit go" repeatedly. It is the marginal-thinking guardrail that stops a one-time concentration decision from hardening into a permanent one after the returns that justified it have flattened.

References¶

[1] The law of diminishing marginal returns holds that, past some point, each additional unit of input yields less additional output. Applied to focus, it means concentrated effort on the vital few eventually saturates, so the next unit is better spent elsewhere — the exact condition this review is built to detect. ↩