5.10 Project Charter Reviews and Toll-Gate Governance
Key Takeaways
- A charter review re-validates the problem, scope, goal, benefit, resources, and risks against what the team has learned since the last gate.
- A toll gate is a decision point where the sponsor approves the phase, returns it with named gaps, or cancels the project.
- Cancelling a project at a toll gate is a legitimate and valuable outcome, not a failure of the team.
- Charter drift is silent scope change that no sponsor approved, and it is what produces disputes at closure.
- Benefit estimates should tighten at each gate, moving from an order-of-magnitude Define figure to a finance-validated number at Control.
Why charters must be reviewed
A Define-phase charter is written with the least information the team will ever have. Almost everything in it is provisional:
| Charter element | Why it changes |
|---|---|
| Baseline | Measure-phase data usually differs from the initial estimate, sometimes materially |
| Scope | Stratification often shows the problem is concentrated in a narrower slice, or the root cause lies upstream |
| Goal | Entitlement analysis in Analyze reveals what is actually achievable |
| Benefit | The estimate tightens as the solution becomes concrete |
| Resources | Requirements shift as the analysis and then the implementation dominate |
| Risks | New risks appear; assumed risks fail to materialize |
| Team | Members join or leave as phases change |
A charter never reviewed becomes a historical document that the team quietly stops referring to. That is charter drift: the project people are actually doing diverges from the project the sponsor approved, and nobody notices until closure, when the sponsor asks about a benefit that stopped being in scope four months ago.
What a charter review covers
Run it as a short standing item at each toll gate rather than as a separate ceremony.
- Is the problem statement still accurate? Has the measured baseline changed the numbers?
- Is the scope still right? Anything added, removed, or discovered outside the boundary?
- Is the goal still both achievable and worth achieving?
- Has the benefit estimate changed, and in which direction?
- Are the resources still committed and sufficient?
- Have the risks changed?
- Is the sponsor still the right sponsor? If the root cause moved into another function, the answer is often no.
Every answer that is "no" produces either an updated charter signed by the sponsor, or an explicit decision to continue unchanged. Both are acceptable; silence is not.
Toll gates
A toll gate (sometimes called a phase gate or tollgate review) is a decision point, not a status presentation. The sponsor has exactly three options:
- Approve: the phase criteria are met; proceed and release the next phase's resources.
- Return: named gaps must be closed; the gate reconvenes on a stated date.
- Cancel: the project should stop.
Cancelling matters. A deployment where no project is ever cancelled is not exercising judgment, and Black Belt capacity is the scarcest resource in the system. Legitimate reasons to cancel at a gate: the measured baseline shows the problem is far smaller than believed; the root cause is outside any feasible scope; the process is being replaced by a capital project; a regulatory or business change has removed the need. Cancelling at Measure after six weeks is a good outcome compared with discovering the same fact at Improve after five months.
Phase-gate criteria
| Gate | Entry criteria to pass |
|---|---|
| Define | Signed charter; problem statement with baseline; scope with in-frame and out-of-frame; SIPOC; customers identified and VOC translated to CTQs; team and resources committed; benefit estimate agreed with finance |
| Measure | Operational definitions written; measurement system validated (MSA passed); data collection plan executed; baseline performance and capability established; problem statement refined with actual data |
| Analyze | Potential causes identified and screened; root causes confirmed with data or statistical evidence; entitlement understood; goal validated as achievable |
| Improve | Solutions generated and selected against criteria; risk assessed (FMEA); pilot run and results verified statistically; full implementation plan approved |
| Control | Control plan in place and owned; process owner has formally accepted; training completed; documentation updated; benefit validated by finance; lessons learned captured; project closed |
The single most useful discipline is that Measure does not pass without a validated measurement system. Every subsequent phase depends on the data, and an MSA failure discovered in Analyze invalidates everything after it.
Who attends
- Sponsor: mandatory. If the sponsor cannot attend, reschedule rather than proceeding without a decider; a gate without the decider is a status meeting.
- Process owner: mandatory from Measure onward, because they will inherit the control plan.
- Master Black Belt or deployment lead: methodology quality assurance.
- Finance representative: at Define to agree the benefit basis and at Control to validate it.
- Affected function leads: where the phase's decisions touch them.
Running the gate
Circulate a short pack in advance -- typically the updated charter one-pager, the phase evidence, the decision being requested, and the risks. Then:
- Ten minutes on what the data showed, not on what the team did.
- The explicit decision request.
- Risks and what the team needs from the sponsor.
- The decision, recorded in the room with a date.
A gate that ends without a recorded decision has not happened. The most common failure is a gate that runs long on methodology explanation, leaves no time for the decision, and defaults to "carry on", which is approval without scrutiny.
Benefit tightening across gates
| Gate | Benefit basis | Typical accuracy |
|---|---|---|
| Define | Estimated from the gap and a closure fraction | plus or minus 50% |
| Measure | Recomputed from the validated baseline | plus or minus 30% |
| Analyze | Bounded by entitlement from the root-cause analysis | plus or minus 20% |
| Improve | Based on measured pilot results | plus or minus 10% |
| Control | Validated by finance from actual post-implementation performance | Actual |
Showing this progression at each gate keeps the sponsor's expectation calibrated and removes the awkward conversation where a $400,000 Define estimate closes at $150,000 with no explanation of when or why it changed.
At the Measure toll gate, data shows the true baseline defect rate is 1.1% rather than the 4% assumed at Define, making the estimated benefit far too small to justify a full-time Black Belt. What is the appropriate sponsor decision?
Why does the Body of Knowledge emphasize periodic charter reviews with stakeholders?
Which criterion must be satisfied before a project can pass the Measure toll gate?