← All entries
·13 min read·decision gates

90 Minute Decision Gates for Leaders: Kill Soft Greenlights With Decision Logs

Run 90 minute decision gate meetings that stop soft greenlights and record decisions in a timestamped decision log.

On this page

A decision gate is a formal checkpoint where leaders decide whether a project or initiative gets more funding, more time, and more resources, or gets stopped. Every gate ends in one of three outcomes: Go, Conditional Go, or Stop. Treat each one as an investment decision measured against fixed criteria, not a status update dressed up as a meeting.


TL;DR:

  • Decision gates should be used only for high-stakes projects where reversing course is costly, such as product launches or market entries.
  • Effective gates require clear, binary decision criteria, a comprehensive evidence package, and a designated veto holder to avoid soft greenlights.
  • Reviewing evidence and making decisions in a structured 90-minute meeting, with private deliberation, prevents drift and ensures accountability.
  • Tracking outcome calibration and using decision journals like Betlog improves gate accuracy by capturing hypotheses and confidence levels upfront.
  • Simplify gate criteria documents to four essential dimensions and enforce deadlines to prevent criteria shifting and ensure decision quality.

Betlog
Make Decision Reasoning Visible
Betlog records hypotheses, confidence, trade-offs, metrics, and outcomes so teams can review decisions without rewriting the past.
Explore Betlog

What Decision Gates Are and Why They Matter

Decision gates exist to interrupt momentum on purpose. Projects have a way of drifting forward on sunk cost and optimism alone, and a gate forces someone with real authority to ask whether the money and people committed so far still make sense given what’s now known. That’s the governance function: not micromanagement, but structured friction at the moments it counts most.

Done well, gates deliver four concrete benefits. McKinsey’s research on gate discipline makes the case bluntly: a gate that doesn’t release or block budget isn’t governance. It’s theater with a calendar invite.

  • Risk control — bad bets get caught before they consume another quarter of runway.
  • Portfolio discipline — capital and talent flow toward the projects earning them, not the ones with the loudest champion.
  • Resource allocation — teams know exactly what evidence unlocks the next phase of funding.
  • Executive accountability — someone signs their name to the decision, which changes how carefully it gets made.

Not every initiative needs this machinery. A two-week internal experiment with no capital risk doesn’t need a scorecard and a signed record. Gates earn their overhead on decisions where reversing course is expensive: new product launches, market entries, six-figure budget commitments. Match the rigor to the stakes, or the process itself becomes the risk.

Anatomy of an Effective Decision Gate

Most weak gates fail for the same reason: they’re missing a structural piece, not because the team lacked judgment. A gate that actually works has four parts, and skipping any one of them is how “soft greenlights” creep in.

  1. Decision criteria and scorecards. Set the pass/fail thresholds before anyone sees the evidence. Binary criteria (yes/no, met/not met) beat vague ones like “sufficient customer validation” every time, because vague language is where politics sneaks back in.
  2. The evidence package. This is the data, customer research, financial model, and risk assessment the team brings to the table. It should exist as a document, not a slide deck improvised the night before.
  3. Decision rights. Name who attends, who actually votes, and who holds veto authority. A structured go/no-go framework works specifically because the facilitator and decision makers are locked in before the meeting starts, not negotiated in the room.
  4. The decision outcome and record. Go, Conditional Go, or Stop, written down and signed by the people who made the call.

PMI’s guidance on gate philosophy frames the review itself around five questions: does the business case still hold, how has the project performed, what risks remain, what’s the plan for the next phase, and will resources actually be committed. Answer those honestly and the rest of the gate mostly takes care of itself.

Pro Tip: Give every gate a named “no” vote. If a single reviewer can’t stop a launch on a legitimate critical concern, you don’t have a gate. You have a rubber stamp with a meeting attached.

Common Gate Models: Stage-Gate, Phase-Gate, and Agile Hybrids

Most organizations don’t invent their own gating structure from scratch. They adapt one of a handful of established models, and the right choice depends heavily on how long the project runs and how much can go wrong along the way.

  • Classic Stage-Gate alternates stages of work with gates that check readiness before the next stage starts, originally built for physical product development where redoing work is expensive.
  • Five-stage/five-gate models compress this into a leaner cycle, common in software and smaller product teams where speed matters more than exhaustive documentation.
  • Seven-gate models, used in formal government project governance, add checkpoints most commercial teams skip: strategic assessment, project approach, business case readiness, detailed plan readiness, deployment readiness, and a post-implementation review. The extra gates earn their keep on large, long-duration, high-risk projects where a missed readiness check is costly to unwind later.
  • Agile-Stage-Gate hybrids shrink the distance between gates and let conditional approvals carry lightweight learning milestones instead of full-blown deliverables. Robert Cooper’s work on next-generation Stage-Gate models describes this shift toward what he calls Adaptive, Agile, and Accelerated gating, designed to keep governance intact without slowing sprints to a crawl.

Pick based on reversal cost, not tradition. If a wrong call costs you a week, run lean gates. If it costs you a year and a facility lease, run the full model.

How to Design Effective Decision Gates for Your Organization

Most gate processes aren’t broken because the wrong people show up. They’re broken because the criteria were never actually agreed on before the evidence started flowing in. Fix the sequence and most of the drama disappears.

  1. Run a criteria session before any evidence gets collected. Get the actual decision makers in a room and force them to agree, in writing, on what “pass” looks like. Doing this after the evidence package is already assembled just invites people to retrofit criteria to whatever number they already like.
  2. Define your dimensions explicitly. Most credible gates score across four areas: customer evidence (do real users want this), technical readiness (can it actually be built or scaled), financial credibility (do the numbers survive scrutiny), and strategic alignment (does this still fit where the company is headed).
  3. Make criteria binary wherever you can. “Customer interviews show at least 60% willingness to pay at the target price” beats “strong customer interest” because it can’t be argued with in the room.
  4. Set a hard evidence deadline and enforce it. Predefined, binary criteria paired with an enforced submission deadline is what stops criteria from quietly shifting mid-review to accommodate whatever the team actually delivered.

Once criteria exist, codify them into a one-page gating playbook: which gates apply to which project types, what evidence is mandatory at each one, and who the standing reviewers are. Keep it short enough that people actually read it before a gate, not after they’ve already missed a requirement.

Government project offices tend to tie the rigor of a gate to project complexity, and the same principle scales down. Guidance on tailoring gate rigor to risk recommends at least one independent review annually for large, long-running projects, precisely because internal teams lose objectivity about their own work over time.

Pro Tip: If your criteria document is more than two pages, nobody is reading it before the gate. Cut it down to the four dimensions and the pass thresholds. Everything else belongs in the evidence package, not the playbook.

Running an Effective Gate Meeting: A 90-Minute Playbook

A gate meeting has one job: produce a clean, defensible decision. Most gates that go sideways do so because the meeting structure itself invites drift, not because the underlying evidence was weak.

Timed 90-minute decision gate process

Five business days before the meeting: distribute the full evidence package and get written confirmation that reviewers have read it. A gate where the decision makers see the numbers for the first time in the room isn’t a review, it’s an ambush on everyone involved, including the project team.

The 90 minutes break into three blocks:

  1. Evidence presentation (40 minutes). The project team walks through the scorecard against each predefined criterion. No new criteria introduced here, no moving goalposts.
  2. Question and answer (30 minutes). Reviewers probe gaps, but no decisions get made out loud during this block. That rule alone eliminates most of the social pressure that turns a Stop into a reluctant Conditional Go.
  3. Private deliberation (20 minutes). The project team steps out. Reviewers score independently against the scorecard, then discuss, then reach a decision without the team’s presence shaping the room.

The team gets called back in only to hear the outcome, not to negotiate it. A structured framework built around private deliberation and a signed record exists specifically to prevent conditional drift, where a Stop slowly softens into a Conditional Go because nobody wants to be the person who kills the project in front of its champions.

Within 48 hours: the decision record gets signed and distributed to the full team. Conditional Go decisions need specific, dated conditions attached, not vague language like “continue building momentum.” A Stop decision needs a clear statement of what would need to change for the idea to come back to a future gate.

Measuring Gate Effectiveness and Fixing Common Failure Modes

A gate process that never says no isn’t working. Track how often actual outcomes match the confidence the team expressed going in. That’s calibration, and it’s the single best signal of whether your gates are catching real problems or just rubber stamping momentum.

Watch for three recurring failure patterns:

  • Soft greenlights — a Conditional Go with conditions so vague nobody can later say whether they were met.
  • Missing reviewers — the person with the strongest incentive to say no wasn’t in the room.
  • Portfolio blindness — gates evaluate each project alone, with no check against what else is competing for the same budget.

Independent reviewers help catch what internal teams miss about their own work, and portfolio-level checks catch the resourcing conflicts a single project gate never will.

Pro Tip: *Track the percentage of Conditional Go decisions that actually close their conditions on time.

How Decision Journals Complement Gate Reviews

A gate is only as good as the evidence and reasoning behind it, and most teams lose the reasoning the moment the meeting ends. A decision journal like Betlog fixes that by capturing the bet before anyone knows how it turns out.

Before a gate, the team logs the hypothesis, a confidence level expressed as a probability rather than a gut feeling, the trade-offs they’re accepting, and the specific metrics that will settle the question. That entry becomes part of the evidence package itself, and it doubles as the canonical decision record once the gate outcome gets signed.

  • Confidence stated upfront makes it possible to check calibration later: did your 70% bets actually land 70% of the time?
  • A timestamped record protects the post-mortem from hindsight bias, since nobody can quietly rewrite what they “really thought” after the outcome is known.
  • Separating decision quality from luck in the post-mortem stops teams from crediting every win to skill and blaming every loss on circumstance.

What I’d Fix First in a Broken Gate Process

Most gate processes don’t need a redesign. They need three things in the next 30 days: written binary criteria, a named veto holder, and an enforced evidence deadline. Get one senior sponsor to publicly back a Stop decision early. That single act does more to fix a rubber-stamp culture than any new template ever will.

— Cesar

Capture the Reasoning Behind Every Gate Decision With Betlog

Betlog exists for the moment right before a gate, when someone has to commit to a hypothesis and a confidence level instead of a hunch dressed up as certainty. It captures the same artifacts a good gate needs anyway: the trade-offs accepted, the metrics that will decide the outcome, and what would prove the bet wrong.

Betlog

Where most gate documentation dies in a slide deck nobody reopens, a Betlog entry stays as the permanent, timestamped record behind the decision. That’s what makes the post-mortem honest later: you can check whether a 70%-confidence Go actually won 70% of the time, instead of reconstructing what everyone “really believed” after the fact. Betlog runs one plan at $39 per month or $390 per year. Start logging your next gate decision before the outcome tells you what to believe about it.

Sources

FAQ

What Is a Decision Gate?

A decision gate is a formal checkpoint where designated reviewers decide whether a project continues, continues with conditions, or stops, based on evidence measured against pre-set criteria. It’s fundamentally an investment decision, not a progress update, according to McKinsey’s research on gate discipline.

What Are the Stages in a Typical Stage-Gate Process?

Stage-Gate models vary by organization, but a common structure runs discovery, scoping, business case, development, testing, and launch, with a gate separating each stage. Larger or higher-risk projects sometimes extend this into a seven-gate model that adds strategic assessment and post-implementation review.

What Should a Gate Decision Record Include?

A gate decision record should state the outcome (Go, Conditional Go, or Stop), the criteria evaluated, any conditions attached with deadlines, and the signatures of the decision makers. Pairing that record with a timestamped decision journal entry, the kind Betlog captures, protects the reasoning from being rewritten in hindsight.

Who Should Have Veto Authority at a Gate?

Veto authority should sit with whoever owns the risk the gate is meant to catch, whether that’s a finance lead on budget gates or a technical lead on readiness gates. Without a named veto holder, gates tend to drift into soft greenlights where nobody feels empowered to say stop.

How Much Does Betlog Cost?

Betlog offers one plan at $39 per month or $390 per year. Both options include the same decision journal features for capturing hypotheses, confidence levels, and post-mortems.

Keep readingMore from the ledger
Your move

Put the ideas on the record. Log your next bet.

Create your free workspaceFree while in beta · No credit card required