Skip to content

Stage-Gate Concept Review

Gated investment review — instantiates Divergence-Convergence Cycle Orchestration

Advances a concept through a sequence of go/kill gates where each gate releases only the next increment of investment, and an independent reviewer — not the concept's champion — decides whether the evidence earns it.

The danger in backing a promising concept is spending big on it before the evidence justifies the size of the bet. Stage-Gate Concept Review manages that by never committing more than one increment at a time. Between each stage of work sits a gate: a go / kill / hold / recycle decision at which an independent reviewer — deliberately not the person selling the concept — checks the accumulated evidence against pre-set criteria and decides whether to release the next, larger tranche of investment. Its defining logic is a commitment ladder: because each stage costs more and is harder to unwind than the last, the evidence bar rises with the spend. Where a screening funnel narrows how many options survive, this mechanism governs how deep to invest in the ones that do — and it puts that escalation under arm's-length judgment rather than under the momentum of advocacy.

Example

A packaged-foods company is developing a new snack concept. It doesn't fund a production line on the strength of a good tasting session; it walks the concept up a ladder of gates. An idea-screen gate passes it cheaply into a small concept test. A concept gate, seeing decent sensory results, releases a modest development budget. At the development gate — the one just before an expensive scale-up — an independent review group, separate from the brand team, finds that the shelf-life stability data simply isn't there yet, and holds the concept rather than letting the team's enthusiasm carry it onto a costly pilot line. The spend never got ahead of the proof. A weaker concept in the same portfolio is killed outright at the concept gate, freeing budget that advocacy would otherwise have kept alive for another year.

How it works

  • Put a gate between every stage. Progress is not continuous; it is a series of explicit go / kill / hold / recycle decisions, each a checkpoint that must be actively passed.
  • Escalate commitment, and the evidence bar with it. Each stage is larger and less reversible than the last, so the proof required to unlock it rises in step — cheap steps on thin evidence, expensive steps only on strong evidence.
  • Judge at arm's length. The gate decision is made by reviewers independent of the concept's champions, against criteria set before the stage began, so optimism and sunk cost don't grade their own work.
  • Decide on the evidence in front of the gate. A gate weighs the results the prior stage produced; it does not generate them.

Tuning parameters

The dials that adapt this review to a specific portfolio:

  • Gate count and spacing — many small gates give tight control but add overhead and slow cadence; few big gates are fast but let more ride on each decision.
  • Evidence bar per gate — how much proof each increment of investment demands; set too high it strangles ambitious concepts, too low it lets weak ones climb.
  • Evaluator independence — champion-led self-review through fully arm's-length gatekeepers. More independence curbs optimism but risks bureaucracy and a bias toward safe, incremental concepts.
  • Kill-versus-recycle disposition — whether a failed gate ends the concept or returns it a stage for rework, which trades decisiveness against second chances.
  • Reversibility weighting — how tightly commitment depth is tied to how hard the next step is to undo, which is what sizes the bet to the risk.

When it helps, and when it misleads

Its strength is keeping spend proportional to evidence: weak concepts die cheaply at early gates, strong ones earn their escalation, and independent gatekeeping blunts the advocacy, optimism, and sunk-cost pressure that otherwise carry a favourite past the point where it should have stopped. It is worth its overhead when the later stages are genuinely expensive or hard to reverse.

Its failure modes are the well-known pathologies of gated processes. Gates can decay into rubber stamps — "success theatre" in which nothing is ever actually killed and the review only ratifies decisions already made upstairs.[1] Heavy gating slows everything and can bias a portfolio toward timid, incremental concepts that clear every gate cleanly. And the process is easy to corrupt by softening a gate's criteria the moment a treasured concept would fail them. The discipline that keeps it honest is to keep the gatekeepers genuinely independent, fix the kill criteria before the stage runs, and preserve a real willingness to kill — a gate that never says no is not a gate.

How it implements the components

Stage-Gate Concept Review fills the investment-governance components of convergence:

  • phase_transition_gate — each gate is a transition checkpoint between stages, with an explicit go / kill / hold / recycle decision.
  • reversible_commitment_ladder — the escalating stages tie the size of each commitment to accumulated evidence and to how reversible the next step is.
  • independent_evaluator_role — the gate is judged by reviewers separate from the concept's advocates, so escalation is authorised at arm's length.

It does not produce the evidence a gate weighs — that is Prototype Tournament — nor does it preserve the concepts it kills or defines the post-launch conditions for reopening; those are the Reserve Option Board and Reopening Trigger Review respectively.

Notes

It is easy to confuse with the Concept Screening Funnel, and the two are often used together, but they answer different questions: the funnel decides which options survive by progressively cheaper-to-costlier filters; a stage-gate decides how much to invest in a survivor and does so under independent, escalating review. A funnel narrows the field; a gate sizes the bet.

References

[1] The discipline of moving a new product through defined stages separated by go/kill decision gates is the Stage-Gate system articulated by Robert G. Cooper. Its most-documented failure in practice is the gate that never kills — a checkpoint that has become a formality — which is why independent gatekeeping and pre-set kill criteria are load-bearing rather than ceremonial. ↩