Construction Shutdown Weeks: Programme the Closure Before It Programmes You
The programme says the facade phase runs from the end of November to the middle of February, and somebody has written "eleven weeks" in the duration column. Everyone in the room knows the site closes for Christmas. Nobody has taken it out of the number. That is the whole problem with a planned shutdown: it is the most predictable thing on the job, known months before anyone breaks ground, and it still ends up being discovered halfway through January when the phase is already three weeks behind a date that was never achievable.
This is a different animal from the days you lose to weather. A wet fortnight is found in the site diary after the fact and counted as a variance. A shutdown is a fact you hold in your hand at tender stage. It should be priced into the baseline programme, not explained afterwards.
Planned closure is a programming input, not a variance
The Christmas and New Year industry shutdown, the Easter break, a builders' holiday week, a trade-specific closure written into an enterprise agreement — all of these are on the calendar before the contract is signed. That changes what you do with them. Unplanned loss gets documented; planned loss gets programmed. The general mechanics of stripping non-working days out of a date range are covered in the companion piece on counting construction working days when rain days are in play, which also deals with why a raw weekday span flatters your capacity and why the endpoints and the site's holiday jurisdiction have to be set deliberately. Take that as read here.
The contrast worth holding onto is the one about entitlement. Weather loss is the classic ground for an extension of time, and the sibling article covers how a documented working-day count works as delay evidence. A planned shutdown generally does not work that way at all. You knew about it, you had the chance to programme around it, and a contract administrator is entitled to ask why a closure that was on the calendar in March turned into a claim in January. The shutdown is not something you recover afterwards. It is something you get right beforehand or wear.
A two-week closure costs more than ten working days
This is where most shutdown planning goes wrong. The site closes for ten working days, so ten days come out of the count, and the arithmetic feels finished. It is not, because a site does not stop and restart like a light switch.
Demobilisation eats the days before the closure. The site has to be made safe and secure: scaffolding checked and tied, excavations covered, loose materials shifted or strapped, temporary works signed off for a period when nobody is walking the site. Plant gets off-hired or the hire keeps running through a dead fortnight — either way somebody has to organise it, and the last day before a shutdown is rarely a full production day. Crews are chasing their own leave, tools go home, and the final afternoon is spent tidying rather than building.
Remobilisation eats the days after it. Crews come back staggered, not all on the Monday — subcontractors stretch their break, some people do not come back at all, and January is when labour moves between employers. Plant has to be re-delivered and recommissioned. Inductions get redone for new faces. Temporary works and scaffolding need re-inspection before anyone climbs. The first few days back run at maybe half pace, and on a big job the ramp is longer than that.
So the honest figure for a two-week closure is not ten days. It is ten days plus a partial day at the front and something like two to three days of lost output at the back. If your programme books the closure at its nominal length, you are already short before January starts. Put the ramp in the number, or at least put it somewhere visible, rather than pretending the site resumes at full pace on the first morning back.
Where the closure lands matters more than how long it is
A three-week shutdown that falls at a clean handover point — structure complete, building watertight, fit-out not yet started — costs you roughly its own length and very little else. The same three weeks landing in the wrong place costs multiples of that.
Think about what a closure does to work that cannot be paused cleanly. A pour sequence broken across a shutdown leaves a construction joint nobody designed and a formwork cycle that has to be struck, stored, and reset. An inspection cycle interrupted means the inspector is on leave too, so the hold point that should have cleared on the 22nd clears on the 12th and every following trade waits. Waterproofing part-installed and unprotected across three wet weeks may simply have to be redone. Wet trades mid-cycle, temporary propping that was meant to be short-term, a crane booked for a lift that now sits idle either side of the break — each of these turns a fifteen-day closure into a thirty-day hole.
So the question to ask of the programme is not just how many working days survive the shutdown. It is which trade is on the critical path across the closure boundary, and whether that trade can stop and restart without penalty. If it cannot, the fix is usually to move the boundary rather than absorb the cost: pull the pour forward a week and accept overtime, or push it back and let the closure fall in a genuine gap. Shifting the work relative to a fixed closure is cheap when you do it in the programme and expensive when you do it on site. That option only exists because the closure date was known in advance — which is precisely the advantage a planned shutdown gives you and an unplanned loss never does.
Your shutdown is not everyone else's shutdown
The site calendar and the supply chain calendar rarely line up, and the mismatch runs in both directions.
Your site is shut but the supplier is not. A delivery arrives at a locked gate. Precast turns up with nobody to take it and no crane, and now you are paying demurrage or a re-delivery charge, or worse, it gets left somewhere it should not be. Somebody has to actively hold those deliveries, and "somebody" usually means a person who is on leave.
The supplier is shut but you are not. This one is quieter and more damaging. A specialist fabricator closes for four weeks when your site only closes for two, so the fortnight you were counting on as productive has no material behind it. Imported items are worse again, because the factory shutdown at the far end of the chain is on a calendar you never looked at.
Then there are lead times that simply span the closure. A ten-week lead item ordered in November does not arrive in the first week of February the way a naive count suggests, because the fabricator's own working days have a hole in them too. Order dates that need to clear a shutdown have to be pulled forward by the length of the closure plus the ramp at both ends, and they need to be pulled forward against the supplier's calendar, not yours. The practical habit is to ask every critical supplier for their shutdown dates at procurement stage and put those dates on the programme next to your own.
A worked example
A precast facade phase is programmed Monday 30 November to Friday 12 February. The raw span holds 55 weekday dates, and the duration column says eleven weeks.
The site closes Monday 21 December to Friday 8 January — three weeks, 15 weekdays, gone. That leaves 40. Two public holidays fall in the working segments either side, so 38. The three-day ramp back in January costs roughly another two days of real output, so call the effective figure 36 against a headline of 55. That is a third of the phase that was never there.
But the more useful output is not the single total. It is the split: 15 working days before the closure and 23 after it. Those two numbers are what you actually programme against, because they tell you what can be finished before the gate shuts. If the facade sequence needs 18 days to reach a weathertight line, 15 does not get you there, and you are leaving the building open across the closure. Now you know in November, and the options are real ones — add resource, resequence, or move the boundary. Discover it in January and the only option left is an apology.
Commit the net figure, do not just record it
Because the closure is known in advance, the output of the count is a commitment rather than a record. The number belongs in the contract programme: a net working-day figure per segment, stated per phase, with the shutdown shown as a hard boundary rather than a stripe running through the bars. That is what makes the durations defensible when someone asks where the eleven weeks went — they did not go anywhere, they were never counted as eleven.
It also changes what the number is for. The rain-day count is retrospective and argumentative; it exists to explain. The shutdown count is prospective and contractual; it exists to commit. Same arithmetic, opposite direction in time, and the shutdown one is the only one you get to be right about before the fact.
Run the segments through the tool
The Construction Shutdown Calculator takes a date range, your work-week pattern, and the site's holiday calendar, then lets you enter the closure dates as custom exclusions so the shutdown comes out of the total like a run of holidays. Run it once per programme segment — the stretch before the closure and the stretch after it — rather than straight through the break, and you get the two figures the programme actually needs instead of one span that hides the gap. Add your ramp allowance, put the resulting numbers in the baseline, and the shutdown stops being the thing that ambushes the job every January.