Stage Gates Are Broken — Here's How to Make Them Actually Work

Stage gates are meant to protect projects — but in most organisations they've become rubber stamps that create the illusion of oversight without any of the substance. This episode breaks down exactly why they fail and what it takes to make them genuinely rigorous.

Show Notes

Stage gate reviews should be the moments when a project earns the right to continue — a structured, honest reckoning between what was promised and what has actually been delivered. In practice, most gate reviews have drifted into something far less useful: polished presentations, softened risks, and committees that approve everything because stopping feels too drastic. This episode of ProjectManager.co unpacks the three core failure modes that turn stage gates into theatre, and offers concrete fixes for each one.

Here's what the episode covers:

• The wrong framing kills the gate before it starts. When teams prepare by asking "what do we need to show to get approved?", the review becomes a sales exercise. The fix is to define specific, verifiable exit criteria before the phase begins — conditions that are either demonstrably true or they aren't, with no room for spin.
• Late information sharing makes scrutiny impossible. Sending the gate pack the morning of the meeting guarantees the first half is spent summarising slides nobody has read. Separating information sharing (at least three working days out, with written questions required in advance) from decision-making transforms the quality of the conversation.
• Binary decisions create a false choice. When the only options are "proceed" or "stop", committees almost always choose proceed — even when real concerns exist. A third option, the conditional proceed, lets reviewers approve progress while binding the team to specific, time-boxed conditions that have a named owner and a hard deadline.
• Gates are not a substitute for ongoing oversight. Problems that fester between formal checkpoints will still arrive at the gate fully formed. Effective project governance treats the gate as the formal checkpoint and everything else — regular status reporting, sponsor engagement, live escalation paths — as the early warning system.
• One change to start with today. Before any project enters its next phase, write down the verifiable conditions that must be true for the gate to pass. Not a deliverable list — a conditions list. That single discipline sharpens every other part of the gate process almost automatically.

The episode also touches on how AI risk management tooling can help surface the concerns that gate packs traditionally bury, and why teams working within hybrid delivery models often struggle most with applying stage gates consistently across workstreams. Listeners who want to go deeper on a related discipline should check out The Risk Register Nobody Reads (And How to Fix It) — a previous episode that tackles the same pattern of governance theatre from a different angle.

ProjectManager.co

What is ProjectManager.co?

Project management for people who deliver, not people who certify. Estimation that survives contact with reality, scope and dependency management, status reporting nobody has to decode, and where self-hosted and AI-assisted tooling genuinely helps.

Each episode takes one practice — a re-estimate mid-project, a dependency nobody owns, a status meeting that produces nothing — and works through a better way to handle it. Written for delivery leads and PMs running real projects, with no methodology evangelism. Five or six minutes, one practice at a time.

Topics include estimation that survives reality, dependency and critical path management, scope negotiation, status reporting nobody decodes, resource planning, retrospectives that change something, and self-hosted versus SaaS tooling.

Produced by ProjectManager.co, self-hosted AI project management. Full details, services and further reading at https://projectmanager.co