Design Downselection Review¶
Evaluation review — instantiates Progressive Narrowing
Converges a portfolio of design concepts through feasibility and prototype evidence to a single committed design, deliberately keeping one structurally different concept alive until the evidence justifies letting it go.
A Design Downselection Review is the recurring evaluation event where an engineering or product team reduces a portfolio of competing design concepts toward the one it will build — judging each on feasibility, prototype and user evidence, cost, risk, and strategic fit. Its defining commitment is that convergence is earned by evidence and paced deliberately: rather than picking a favorite early, the review advances a slate, tests the survivors harder each round, and protects at least one structurally different concept from elimination-by-familiarity until the data forces the call. Where a scoring tool ranks options in the abstract, this review turns real prototype results into a committed selection.
Example¶
A team is deciding how to land a rover on Mars and carries five architectures into the first review: airbags, legged propulsive descent, a sky-crane, an inflatable decelerator, and a crushable-structure lander. Round one asks only feasibility against mass and entry-heating limits; the inflatable decelerator survives on paper and is explicitly named the "structurally different" concept worth keeping despite low familiarity. Round two demands drop-test data: airbags fail at the rover's mass, and the crushable structure's peak decelerations exceed the instruments' limits. Round three pits the sky-crane against legged descent on landing-accuracy prototypes; the sky-crane wins on the evidence and becomes the committed design, while the inflatable decelerator is retired with a documented engineering reason — immature at this mass — rather than quietly dropped.
The discipline that mattered was deepening the evidence bar each round and shielding the odd-one-out until it genuinely lost. That is what keeps a team from converging on the comfortable option before the hard data has arrived. (Illustrative — not a claim about any specific mission's real trade study.)
How it works¶
- Advance a slate, not a favorite. Each review passes forward multiple concepts until the evidence discriminates among them.
- Deepen the evidence bar each round — feasibility analysis → prototype/drop test → user or mission-representative test. A concept may be eliminated only on evidence appropriate to its stage.
- Name and protect a wild card. At least one structurally different concept is explicitly retained past the round its unfamiliarity would otherwise kill it.
- Commit once, explicitly. The final review freezes the selected design and formally retires the losers with reasons.
Tuning parameters¶
- Slate width per round — how many concepts survive each review. Wider slates delay convergence and cost prototyping effort but reduce the chance of dropping a late-blooming design.
- Evidence-bar escalation — how fast the required evidence deepens. Steep escalation converges quickly but can retire a concept before its evidence matured; gentle escalation is thorough but slow.
- Wild-card protection strength — how firmly the structurally different concept is shielded. Strong protection guards against groupthink but can keep a dead concept on life support.
- Commitment reversibility — whether the final selection is truly frozen or can be revisited. A hard freeze concentrates effort; a soft freeze preserves safety but invites relitigation.
When it helps, and when it misleads¶
Its strength is that it forces convergence to be paid for in evidence, and the wild-card discipline is a direct antidote to converging on the familiar. It fits exactly the case where concepts are internally generated and their real merits only become visible once something is built.
Its failure mode is that prototype evidence is expensive, so a team under schedule pressure cuts the slate to one too early and then backfills a review to bless a decision already made — a review as ceremony. The classic misuse is the predetermined winner, where the review exists to make a favored concept look inevitable. The guarding discipline borrows from set-based concurrent engineering:[n1] keep several designs alive and defer commitment until the trade space is genuinely resolved, rather than fixing a point design early and reworking it under pressure.
How it implements the components¶
evidence_requirement— each round specifies the evidence (feasibility analysis, prototype, drop test, user study) a concept must produce to advance, so concepts leave on data, not taste.diversity_retention_rule— the named-wild-card discipline holds a structurally different concept in the slate past the point familiarity would eliminate it.comparison_frame— surviving concepts are judged in a shared trade frame (mass, risk, accuracy, cost, fit) so each round's comparison is apples-to-apples.finalist_commitment_rule— the final review freezes the selected design and formally retires the alternatives.
It commits to a design but keeps no fallback_or_reserve_set of ready backups and runs no bias_and_fairness_check on protected groups — those are procurement_shortlisting and hiring_shortlist_process; a downselection review defends its choice to engineers, not to losing bidders or applicants.
Related¶
- Instantiates: Progressive Narrowing — staged, evidence-deepening convergence on a single built design.
- Consumes: Weighted Scoring Matrix — a review often uses a scoring matrix as the comparison arithmetic within a round.
- Sibling mechanisms: Diagnostic Narrowing Protocol · Research Hypothesis Elimination · Hiring Shortlist Process · Procurement Shortlisting · Successive Screening · Legal Issue Narrowing · Candidate Disposition Log · Weighted Scoring Matrix · Funnel Process
Editorial Notes¶
Form Classification¶
Form family: Decision, Gate & Allocation
Rationale: Successive evidence gates retire concepts, preserve a structurally different wildcard, and ultimately freeze one committed design with reasons, so the mechanism's defining output is staged design selection.
Nearest alternative: Assessment, Review & Assurance — Feasibility and prototype evidence are assessed at every round, but the review exists to choose which concepts advance and which are retired.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Engineering & Design
Origin pattern: Single lineage
Present-day reach: Multi-domain
Rationale: Set-based engineering cohered evidence-gated reviews that carry several concepts in parallel and eliminate them only as feasibility and prototype evidence arrives.
Related originating lineages:
- Innovation & Entrepreneurship — Product portfolio governance supplied strategic-fit, cost, and staged investment criteria.
Review resolution: Set-based engineering cohered evidence-gated reviews that carry several concepts in parallel and eliminate them only as feasibility and prototype evidence arrives. The retained alternate lineages materially shaped the mechanism's form.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
[n1] Set-Based Concurrent Engineering — the Toyota-associated practice of developing several design alternatives in parallel and narrowing only as evidence eliminates them, rather than committing to a single point design early and iterating on it. It is the canonical counter to premature design convergence, and the reason a downselection review deliberately carries a slate rather than a favorite. ↩