An extension of time is for delay to completion, which means delay to the chain of activities that governs completion. Work delayed off that chain may cost real money and still not justify any extension at all. Establishing which is which is the first task in extension of time claim preparation.
The term carries more authority than it deserves. A critical path is an output of a plan, and when the plan changes the path changes with it.
What the term means
The longest chain of dependent activities through the works. Its length sets the completion date, and anything that lengthens it moves that date.
Activities off the chain have slack: they can be delayed, within limits, without affecting completion.
Which chain governs depends on durations and on the logic linking activities, both of which are judgements made when the programme was prepared.
Why it moves
As durations change and activities complete early or late, the governing chain changes with them.
Work that had slack can become critical. Work that was critical can cease to be.
So a claim that identifies the critical path as though it were fixed for the whole project is describing something that never existed.
Concurrent paths
Where two chains are of similar length, both effectively govern, and delay to either delays completion.
Near critical work matters for the same reason, which is why claims that consider only the single longest chain frequently miss the real cause.
A programme with several paths of similar length is fragile, and that is a planning observation as much as a claims one.
What the programme has to contain
Activities at a level of detail that reflects how the work will actually be carried out, rather than twenty bars covering a two year project.
Logic links between them, so that the chain exists at all. A programme without links is a bar chart and shows no path.
And durations that were arrived at from resources and output rather than from the date the work was wanted by.
Programmes that cannot support a claim
No logic, so no path can be identified and no effect calculated.
Logic that forces a predetermined end date, which produces a path that follows the constraint rather than the work.
Or so little detail that the affected work does not appear as an activity, which makes its effect impossible to show.
Delay off the path
Work delayed within its slack costs money and does not move completion, which means it belongs in a cost claim rather than a time one.
That is frequently the real answer on disrupted projects, and it is an answer worth giving early.
Pressing for an extension where the delay was off the path wastes the entitlement argument and weakens the cost claim beside it.
Where the chain runs through others
On many projects the governing chain runs through design information, statutory approvals or long lead procurement rather than through site work.
Those activities belong on the programme for exactly that reason, and leaving them off is what makes late information so hard to claim for.
A programme that shows information requirements with dates is also the most useful management document on the project, quite apart from any claim.
What the assessor is looking at
Which activities governed completion at the time the event happened, and what the event did to them.
Whether the programme relied on was the one in use at the time or one prepared afterwards for the purpose.
And whether the effect claimed follows from the logic, or has been asserted and left for the reader to accept.
Software and its limits
Planning software calculates the path from the logic and durations entered. It does not check whether either is sensible.
A result produced from poor logic looks exactly as authoritative as one produced from good logic, which is why outputs are worth reading rather than quoting.
Most of the useful work in this area is in the inputs, and it is done by people who know how the work was built.
Keeping it simple where the facts allow
On many projects the cause of delay is obvious and a short narrative with a marked up programme establishes it adequately.
Reaching for a detailed analysis where the facts are plain adds cost and invites argument about the method rather than the facts.
Proportionality is the test, and it is worth applying before any analysis is commissioned.
Where expert analysis belongs
Complex delay analysis on a heavily disrupted project is a specialist discipline with its own practitioners and its own literature.
Where a dispute genuinely requires it, saying so is better than producing an approximation and presenting it as more.
For most projects the question is narrower: which chain governed, what hit it, and what the contract says about the consequence.
What to establish first
Which programme was in use, what it showed as governing, and whether it was accepted or submitted under the contract.
That takes a day or two and it determines whether a time claim is available at all.
Where the answer is that no usable programme exists, that is worth knowing before anything is drafted rather than after.
Where it fits
The critical path question comes before both the entitlement argument and the cost claim, because it governs what either can achieve.
Where the delay turns out to be off the path, the money route is loss and expense rather than an extension.
The programme that makes either possible is maintained during the works as part of cost control and variations, not reconstructed afterwards.
Which is the recurring point in this whole subject: the work that makes a claim short is done while the project is running, by people who are not thinking about claims at all.
Running late with nobody agreeing why?
Extension of time claim preparation sets out the events relied on, their effect on the programme and the contractual route to relief, in a form that can be assessed rather than argued over.