Skip to content

Preserve–Prune–Recombine Review

Consolidation review — instantiates Opportunity-Gated Adaptive Diversification

The consolidation decision after the fan-out — sorting proven lineages into keep, kill, or merge, retaining the winners and folding their best parts back together.

When the saturation signal fires, someone has to act on it. Preserve–Prune–Recombine Review is the decision that ends the diversification: given the evidence that niches have filled, it assigns each lineage a fate — preserve the proven fits, prune the rest, and recombine complementary parts of the losers into the survivors. Its defining move is that pruning is not deletion: a line that failed as a whole may carry one component worth salvaging, so the review is built to harvest value on the way out through a shared interface rather than simply switching lines off. It is the action counterpart to detection — it decides what to keep, not whether the space is full.

Example

A carmaker that spawned six niche EV platforms during an all-out land-grab — city runabout, long-range sedan, work van, performance, budget, luxury — now has a saturated space and too much fragmented engineering to sustain. The Preserve–Prune–Recombine Review works through them. Two platforms have clearly found their niches and are preserved. Three underperformers are pruned — but not blindly deleted: the review keeps a rollback option on one in case its niche reopens. And the best-in-class battery module from a pruned performance platform is recombined into a surviving platform through the shared mechanical and software interface all six were required to keep.

The result is fewer, stronger lines that inherit the best of the ones cut — not a cull that throws away everything a losing platform learned. Six niche bets consolidate into two robust platforms plus a salvaged module, and the space stops fragmenting engineering it can no longer afford to spread so thin.

How it works

The review runs four moves, and the order matters:

  • Apply the retention rule. Score each lineage against an explicit keep bar — fit evidence and the difficulty of the niche it faced — and preserve those that clear it.
  • Apply the prune rule. Wind down the rest deliberately, deciding for each whether the exit is a hard kill or an archived, reactivatable one.
  • Open the consolidation pathway. Fold survivors into fewer, stronger lines rather than leaving a thinned but still-fragmented set.
  • Recombine through the interface. Use the shared compatibility interface to transfer salvageable parts from pruned lines into survivors, so value leaves with the loser instead of dying with it.

Tuning parameters

  • Retention-bar height — how many lineages survive. A high bar concentrates strength in a few lines; a low one hedges but keeps fragmentation.
  • Prune reversibility — a hard kill (cheap, final) versus an archived, reactivatable exit (costlier to hold, but preserves a rollback option if the niche reopens).
  • Recombination ambition — a light lift-and-reuse of a component versus a deep re-engineering that fuses the best of several lines; more ambition salvages more value but costs more and can destabilize a survivor.
  • Consolidation speed — a fast, clean cut versus a gradual sunset that lets dependents migrate.

When it helps, and when it misleads

Its strength is that it ends sprawl without throwing away what the fan-out bought: it retains genuine winners, salvages the best parts of the losers, and — done well — preserves a floor of truly distinct survivors rather than collapsing to a single line.

Its failure modes are pointed. The first is survivorship bias — preserving the lineages that happened to win in easy niches while pruning ones that merely faced harder tests, so the review rewards luck of assignment rather than fit.[1] The second is the classic backwards run: convening the review to rubber-stamp a survivor already chosen by politics, with the keep/kill rationale written afterward. And over-aggressive consolidation destroys option value and the hidden resilience of common-mode diversity — collapsing to one line is the very premature lock-in the whole pattern set out to avoid. The discipline that guards against all three is to judge against the pre-set fit criterion and niche difficulty, keep a rollback option on prunes whose niches might reopen, and protect a minimum diversity rather than optimizing to a single winner.

How it implements the components

Preserve–Prune–Recombine Review fills the archetype's consolidation-and-retention components — the ones that act on a saturated space:

  • selective_retention_rule — the explicit bar deciding which lineages are preserved.
  • exit_or_prune_rule — how losing lines are wound down, hard-killed or archived with a rollback option.
  • consolidation_pathway — how surviving lines fold into fewer, stronger ones.
  • recombination_interface — the shared connection through which a pruned line's best parts transfer into a survivor.

It does not detect when consolidation should begin (that's Saturation and Crowding Review), measure each line's fit (that's Lineage–Niche Fit Dashboard), or shelter lines during exploration (that's Protected Pilot Lane).

Notes

Consolidation must not over-consolidate. The archetype warns against premature lock-in at both ends — collapsing a saturated space to a single line reintroduces exactly the fragility the fan-out was meant to avoid — so this review should preserve a floor of genuinely distinct survivors, the minimum diversity that a mechanism like Diversity-Floor Rate Boost exists to protect, rather than optimizing to one winner.

References

[1] Survivorship bias — reasoning only from the cases that made it through while ignoring those that didn't, which distorts what "success" looks like. In a consolidation review it shows up as crediting the lineages that won easy niches and pruning those that faced harder tests, so retention tracks the difficulty of the assignment as much as the merit of the line — which is why niche difficulty belongs in the keep bar alongside raw fit. ↩