In many organisations, the project review is a meeting that ends without a single decision having been taken. The schedule has been commented on, the indicators have been read aloud, and everyone leaves with the reading they had when they came in. The meeting was held; the project, for its part, still does not know what to do about its delay, its budget or its request for reinforcements.
Yet that is precisely what a project review is for. It is an internal checkpoint, held at a regular pace or when a milestone is passed, where the team reports on the state of the project to those who have the power to decide on it: sponsor, steering committee, programme management. It is judged by what it produces, namely a record of dated decisions. To get there, the committee must obtain an answer to five questions, and recognise the moment when a sixth arises: that of the audit.
Is the project still worth what remains to be spent?
The question seems settled since the launch, which is why it is readily skipped. Between two reviews, however, the context moves: a regulatory deadline is shifted, another project delivers part of the expected benefit, the remaining work has been re-estimated upwards.
The committee therefore compares two present-day quantities: what the project can still bring in and what remains to be paid to obtain it. Past spending does not enter the calculation. The document to request is the updated business case, in which the assumptions that have changed since the previous review are flagged. The PRINCE2 method makes this a principle: the justification of the project is re-examined at the end of each stage, before the next one is authorised.
Where do we stand, with evidence?
Progress is measured on what is finished and verified. A budget three-quarters consumed does not mean the work is three-quarters done, and a task “in progress” for three reviews is not moving forward. The earned value technique formalises this comparison between planned work, completed work and actual cost; failing that, a list of accepted deliverables is enough.
The committee asks for three items. First, what has been delivered and accepted since the last review, backed by a demonstration or acceptance test results. Next, the remaining work, in an estimate redone by the teams responsible for producing it. Finally, the forecast date and cost at completion, presented alongside their values from previous reviews: a forecast that slips by the same gap at every session is information in itself.
Coloured indicators only make sense if the rule that moves them from green to amber was written before the review. Ensuring that measures remain comparable from one session to the next is one of the tasks of the project office, whose role is described on our page on setting up a PMO.
What threatens what comes next?
A review does not reread the risk register line by line. It examines a short list: the risks whose last useful decision date falls before the next review, dependencies on another project or a supplier, the plan assumptions that nobody has yet verified.
For each one, the committee wants to know who is accountable for it, what response is under way and what signal would announce that it is materialising. Keeping this list is detailed in our article on project risk management.
What decisions does the project expect from this committee?
This is where the review changes direction: having reported, the team asks. Scope trade-off, reinforcements, additional budget, priority conflict with another project: each request is presented as a choice, with its options, the price of each and the consequence for the project if the committee does not decide.
Two conditions make an answer possible. Requests are sent before the session, with the pack, so that members arrive having read them. And those who sit on the committee genuinely have the authority to decide: when the person who decides is represented by someone who will “pass it on”, the decision is postponed by one review.
The record of decisions sets down what was decided, by whom and for what date. The next review opens by going back over it. A project that asks for nothing for several sessions in a row deserves a question: either it needs nothing, or requests are no longer coming up.
Can the milestone be passed?
The question does not arise at every session. It distinguishes the periodic review, which follows the project at a fixed pace, from the milestone review, held before a commitment: launch of the build, entry into acceptance testing, go-live.
The passing criteria are written in advance, when the milestone is planned. On the day, three answers are possible: passed, passed with reservations, refused. The second is the most convenient, and the most fragile. A reservation only has value if it carries an owner, a deadline and a check placed on the agenda of the next review. Without that, the milestone has been passed without conditions, and nobody decided it.
Project review or audit: when to change instrument?
A project review works on what the project presents to it. It holds as long as the committee trusts this information. The audit responds to a different situation.
| Project review | Project audit | |
|---|---|---|
| Who examines | The sponsor or the steering committee, with the team | A party independent of the team and its line management |
| When | At a regular pace and at each milestone | On a one-off basis, at the decision of a sponsor |
| On what material | The information reported by the project | Facts verified at their source and cross-checked |
| What comes out of it | Steering decisions | Substantiated findings, options and a recommendation |
The vocabulary fuels the confusion: people speak of an “independent review” or a “health check” for examinations entrusted to a third party. The criterion is the independence of whoever examines, whatever name is given to the exercise.
Several signals indicate that the review has reached its limit. The same questions remain unanswered two sessions in a row. Forecasts slip back every time. The team, the supplier and the users no longer describe the same project. Or a heavy commitment is approaching, and the committee wants the opinion of someone who has nothing to defend in the project. Our page on the project audit describes the possible formats and how they unfold.
If the review can no longer answer one of the five questions on its own, the decision to enter in the record is this one: commission the audit, appoint its sponsor and set the date by which its conclusions are expected.