Feature Admission Rubric¶
Scoring rubric — instantiates Second-System Complexity Restraint
Scores each proposed addition on value, evidence, complexity, support, migration, and tier-fit, so admission to the successor is a decision the item earns rather than a default it inherits.
In a second system, the default answer to "should this go in?" quietly flips from no to yes — the successor is the once-in-a-generation chance, so every deferred wish arrives assuming it belongs. Feature Admission Rubric flips the default back. It is a per-item scorecard: each proposed addition is scored on a fixed set of axes — the value it delivers, the evidence that value is real, its complexity cost, its support and maintenance burden, its migration risk, and which release tier it fits — and admission is granted only when the scored profile clears a bar. Its defining idea is that admission is earned per item, on evidence, not conferred by the excitement of the moment. It judges one thing at a time, coldly, against the same axes every time.
Example¶
A photo-editing app's second version is drowning in requests after the first version's success: AI background removal, RAW support, cloud sync, a curves tool, social sharing, animated stickers. Rather than argue each in a meeting where the loudest advocate wins, the team runs every candidate through the same rubric. AI background removal scores high on value and on evidence (top support request, measured in-app searches) but high on complexity and support load — admitted, but to the stabilization tier, not launch. Animated stickers score low on evidence (nobody asked) and moderate on complexity — not admitted; parked in the backlog. RAW support scores high on value for a small, vocal pro segment but carries heavy migration and support cost — held for a later tier pending evidence the segment will pay.
The rubric's output is not a ranking of good ideas but a set of admission decisions with reasons attached. Two features the founders loved score below the bar and are declined — and because the score, not a personality, made the call, the decline sticks.
How it works¶
- Fix the axes before the candidates. Value, evidence, complexity, support burden, migration risk, tier-fit — chosen and weighted up front so no item is judged on axes invented to favor it.
- Score each item independently against those axes; the rubric is deliberately item-level, giving each addition one honest profile rather than a relative popularity contest.
- Apply an admission bar, not a sort. A candidate is admitted, deferred to a named tier, or declined based on whether its profile clears the threshold — evidence is a gate, so a beloved feature with no evidence does not pass on charm.
- Send everything not-admitted to the backlog with its score, so a decline is a documented, reversible state — not a loss.
Tuning parameters¶
- Axis weights — how much value, evidence, complexity, and burden each count. Weighting evidence heavily protects against hype-driven scope but can starve genuinely novel bets that have no track record yet.
- Admission bar height — how high the composite must clear. A high bar keeps the successor lean but risks excluding worthwhile additions; a low one lets the default-yes creep back in.
- Evidence standard — what counts as evidence (measured demand, pilot data, support-ticket volume, mere assertion). Raising it is the rubric's main defense against relabeling wishes as requirements.
- Scorer independence — whether the proposer scores their own item or a neutral party does. Independent scoring resists optimism but costs coordination.
When it helps, and when it misleads¶
Its strength is depersonalizing admission: a shared scorecard lets a team decline a powerful stakeholder's pet feature on the record rather than in a corridor, and it directly attacks the second-system symptom of deferred ambitions being relabeled as non-negotiable requirements. Structured prioritization scoring — the discipline behind approaches like RICE — turns "everyone wants everything" into a comparable, defensible set of decisions.[n1]
Its failure mode is the illusion of objectivity: the composite score is only as honest as the inputs, and a determined advocate can inflate the value estimate or dredge up thin "evidence" until the number clears the bar — precision laundering a preference. The classic misuse is gaming the axes, or treating a high score as an obligation to admit rather than permission to consider. The guarding discipline is to keep the evidence axis strict and auditable, to have someone other than the proposer challenge the scores, and to remember the rubric ranks admission-worthiness, not truth — a low score is a not now, revisited when real evidence arrives.
How it implements the components¶
ambition_triage_gate— the rubric is the triage gate: it turns an undifferentiated backlog of wishes into per-item admit / defer / decline decisions on named criteria.deferred_ambition_backlog— every declined or deferred candidate is written back to the backlog with its score, keeping ambition visible and its status justified.
It does not decide the ordering or timing of the tiers those admitted items flow into — that schedule is the staged_successor_release_ladder, owned by Staged Release Ladder; and it does not cap the successor's total complexity, which is the aggregate job of Complexity Budget Review.
Related¶
- Instantiates: Second-System Complexity Restraint — supplies the per-item admission gate that keeps the default answer at no.
- Consumes: Constraint-Release Inventory — populates part of the backlog the rubric scores.
- Sibling mechanisms: Architecture Decision Record with Future-Option Cost · Complexity Budget Review · Constraint-Release Inventory · Parity-Then-Expansion Gate · Postlaunch Ambition-Thaw Review · Rewrite Scope Firewall · Second-System Premortem · Staged Release Ladder · Successor Charter
Editorial Notes¶
Form Classification¶
Form family: Decision, Gate & Allocation
Rationale: The rubric scores a proposed feature and makes a bounded admission or exclusion disposition for the successor based on value, evidence, cost, support, migration, and tier fit.
Nearest alternative: Assessment, Review & Assurance — Scoring evaluates the proposal, but the mechanism's operative output is whether the feature is allocated a place in the successor.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Organizational & Management Science
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Portfolio governance and stage-gate management provide the lineage for scoring additions before admission.
Related originating lineages:
- Computer Science & Software Engineering — Software product and architecture governance materially shaped complexity, migration, and support criteria.
- Human-Computer Interaction — Usability practice materially shaped the value and user-surface dimensions of admission.
Review resolution: Both reviewers agree that organizational_management is primary. I retain computer_science, human_computer_interaction only as formative origin lineage(s), without treating every later application as an origin. cross_disciplinary_synthesis is appropriate because the exact artifact combines contributions from multiple professional lineages. Reach is multi_domain as a separate applicability judgment: it does not widen or narrow the recorded provenance. Encyclopedia synthesis is true because the exact generalized artifact is an encyclopedia-authored combination or refinement. The secondary differences are reconciled with no unresolved primary-provenance ambiguity.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
[n1] RICE scoring — a prioritization method (popularized by Intercom) that rates each candidate on Reach, Impact, Confidence, and Effort to produce a comparable score. The admission rubric borrows its core idea: force every proposal through the same weighted axes, especially a confidence/evidence term, rather than deciding by advocacy. ↩