ChecklistBusiness Building
Stage-gate funding for new ventures: a gate-by-gate kill criteria checklist
Metered funding releases money to a new venture in tranches, each tied to evidence agreed in advance, so a venture board can limit its downside without micromanaging the team. This checklist sets out five gates from thesis approval to the scale decision, the evidence each should require and what each tranche is for, then shows how to write kill criteria before anyone becomes attached to the idea.
On this page
- Why tranche funding beats an annual venture budget
- Five funding gates from thesis to scale
- Gate zero: approving the venture thesis
- Gate one: confirming the problem is real
- Gate two: validating the solution with users
- Gate three: proving early commercial traction
- Gate four: the decision to scale
- Writing kill criteria before attachment sets in
- Hypothetical kill criteria for an AI claims-triage venture
- Running the venture board
- Behavioral traps that keep weak ventures alive
- Questions and answers
- Sources
Why tranche funding beats an annual venture budget
A full year's budget handed over up front asks a new venture to defend a detailed plan before it has evidence for any of it, and the money then gets spent on the plan whether or not the plan survives contact with customers. Metered funding, also called tranche or stage-gate funding, works the other way round: each release buys answers to the next set of questions, and the next release depends on what those answers were.
That caps the downside, because a venture that fails early has used only its early tranches, and it keeps the team focused, because the evidence needed for the next tranche is written down. ColdAI's studio method follows the same logic: core assumptions are tested before significant resources are committed1, and the capital structure and material commitments of each studio company are approved at group level2.
This page covers internal money released by a parent company or venture board. Raising money from outside investors is a separate discipline, covered under fundraising.
Five funding gates from thesis to scale
- Thesis approved
Releases a small budget for a validation sprint.
- Problem validated
Releases money to test solution concepts with real users.
- Solution validated
Releases the MVP build and priced pilots.
- Early traction
Releases a go-to-market build-out in one segment.
- Scale decision
Releases scale funding or prepares an outside raise.
Gate zero: approving the venture thesis
Releases a small, time-boxed budget for a validation sprint. Ask for:
Gate one: confirming the problem is real
Releases money to test solution concepts, such as a prototype or a concierge version, with real users. Ask for:
Gate two: validating the solution with users
Releases the MVP build and priced pilots. Ask for:
Gate three: proving early commercial traction
Releases money to build out go-to-market in the strongest segment. Ask for:
Gate four: the decision to scale
Releases scale funding or prepares the venture for outside capital. Ask for:
Writing kill criteria before attachment sets in
Kill criteria are the conditions under which a venture stops. They work only if written before the evidence arrives, ideally at the previous gate while the team is still neutral about the outcome. Written later, they bend to fit whatever the results turned out to be.
A good criterion names the assumption, the evidence source, the threshold and the date. “If fewer than half of the target buyers we interview by the end of the sprint describe the problem unprompted, we stop” is a kill criterion; “if the market seems lukewarm” is not. Decide in advance who may override a criterion, require the override and its reasons in writing, and allow it rarely.
Hypothetical kill criteria for an AI claims-triage venture
Running the venture board
Set the membership
A sponsor, a finance lead and at least one member with venture experience from outside the sponsoring business. Venture leads present but do not vote on their own venture.
Meet at gates, with a standing check-in
Gate reviews happen when a venture is ready, plus a short regular review so that stalled ventures surface.
Use one board paper format
Evidence against each criterion, spend against the tranche, what remains unproven and the ask for the next gate, short enough to read in the meeting.
Choose one of four decisions
Proceed, proceed with named conditions, pivot with a smaller tranche, or stop. Record the decision and the reasons.
Fund from a ring-fenced pool
Keep venture money apart from the annual budget so no gate waits for the planning cycle, and money returned by stopped ventures stays available for new ones.
Behavioral traps that keep weak ventures alive
Sunk-cost reasoning
Early signal“We have already spent so much” comes up in the board discussion.
MitigationJudge each gate only on evidence and the next tranche; past spend is not a criterion.
Sponsor attachment
Early signalThe sponsor argues the venture's case instead of questioning it.
MitigationGive a board member with no stake the explicit job of challenging the evidence.
Vanity metrics
Early signalBoard papers lead with sign-ups, page views or pipeline value rather than commitments and retention.
MitigationFix the gate metrics in the kill criteria so substitutes are visibly off-plan.
Moving goalposts
Early signalThresholds are reinterpreted after the results arrive.
MitigationShow results beside the criteria as originally written in every paper.
Zombie ventures
Early signalA venture never fails a gate but never clearly passes one.
MitigationAllow one conditional pass per gate; a second becomes a pivot or a stop.
Fear of stopping
Early signalTeams avoid recommending a stop because they fear for their careers.
MitigationRecognize well-run stops openly and give stopped teams first call on the next venture.
Questions and answers
How much should each venture funding tranche release?
Enough to answer the next gate's questions with some margin, and no more. Work out the tests, build work and team time the next stage needs, add a contingency and set a date by which the evidence should exist. Tying tranches to the full business plan defeats the purpose, because the plan will change once real evidence arrives.
Can a venture skip a funding gate if the evidence is already strong?
Yes, if the evidence for both gates exists and has been reviewed against both sets of criteria. A venture built from an existing product line may already have proof of the problem, for example. What should not be skipped is the written record, because later investors and auditors will want to see what was known at each decision.
What happens to the team when a venture is stopped?
Ideally the team moves to the next thesis or into roles where its new skills are useful, and the stop is recorded as a decision made well. If people expect a stop to damage their careers, they will avoid recommending one and keep weak ventures alive. Publishing what was learned also saves the next team from repeating the same tests.
How is stage-gate funding different from raising a venture capital round?
Stage-gate funding is internal: a parent or venture board releases its own money as evidence arrives and can stop at any gate. A venture capital round brings in outside investors who buy equity on negotiated terms, usually funding a longer stretch of work at once. Many ventures use gates until the scale decision and then raise outside capital, which is covered under fundraising.