The Bar Says Three Months; the Site Has to Resource Thirteen Weeks
A construction programme is drawn in months because that is the only way an eighteen-month build fits legibly on one sheet for a client. Nobody builds in months. Subcontractors are booked by the week, plant is hired by the week, the look-ahead pinned up in the site office covers the next few weeks, and progress is claimed against weekly production rates. Every bar on that Gantt chart has to be broken back down into weeks before a single crew can be committed to it.
That translation is where programmes quietly lose time — not through bad sequencing, but because a bar labelled "3 months" gets resourced as twelve weeks when it is really a little over thirteen.
Where the Two Views of a Programme Meet
Months are a drawing convention
Crews are committed by the week
Look-aheads live in a two-to-six week window
Lost time is counted in days
Breaking a Programme Bar Into Crew Weeks
The routine takes a few seconds per bar and produces the numbers a resource schedule, a subcontract order and a look-ahead all need.
Type the bar duration in months
Enter the figure on the left and the week count lands on the same keystroke. Fractions work, so a two-and-a-half-month first fix comes back as 10.8703125 wk. A value pasted from a schedule export with spaces or a comma decimal is read without editing.
Strip out the weeks nobody will be working
The result is calendar weeks. Take off the Christmas shutdown, public holidays and the weather allowance for the season before turning it into crew weeks — a thirteen-week winter bar can easily be eleven productive ones.
Flip it for the client report
Press swap (↔) to run wk → mo when the site has given you a week figure and the progress report wants months: a six-week delay reads as 1.37990513 mo. Either box takes input, so a week count typed on the right returns the month figure directly.
Drop the number into the resource sheet
Copy gives digits with no unit and no grouping, so a histogram column multiplies it by a gang size without complaining about text. Ctrl + C in the field behaves the same way.
A Mid-Rise Programme Expanded Into Crew Weeks
Indicative phase durations for a mid-rise frame-and-fit-out job, each month figure converted at 4.348125 wk. The phases overlap on the programme, so the column deliberately does not sum to the eighteen months they sit inside.
| Programme phase | Months on the bar | Calendar weeks | Crew note |
|---|---|---|---|
| Enabling works and site set-up | 1 mo | 4.348125 wk | Front-loaded; hoarding and welfare absorb the first two weeks |
| Piling and substructure | 2 mo | 8.69625 wk | Rig hire is booked whole-week; a part week still costs a week |
| Superstructure frame | 3 mo | 13.044375 wk | Crane period must match this, not the twelve weeks the bar suggests |
| Envelope and cladding | 4 mo | 17.3925 wk | Weather-exposed; carry a seasonal allowance inside the week count |
| MEP first fix | 2.5 mo | 10.8703125 wk | Follows the frame floor by floor; sequenced in weekly zones |
| Internal fit-out | 6 mo | 26.08875 wk | The longest continuous labour draw on the job |
| Commissioning and handover | 1.5 mo | 6.5221875 wk | Witness testing dates are fixed; float here is usually gone |
| Whole programme | 18 mo | 78.26625 wk | Roughly 78 weekly progress reports and look-ahead cycles |
Read the frame row twice. Three months is thirteen weeks and a fraction, so a crane booked for twelve comes off hire while the last floor is still going up — the classic way a programme that looked fine on the bar chart turns into a claim on site. The same rounding applied across substructure, envelope and fit-out is where five or six weeks of unbooked crew hide inside an otherwise sensible eighteen-month plan.
What This Pair Does in the Planning Office
Bar durations expanded one after another
The week figure follows each keystroke, so a whole column of phase durations can be worked through in one pass without submitting anything between bars.
Site figures turned back around for the client
Both boxes accept input, so a delay or a recovery measured in weeks on site returns a month figure for the monthly progress report without restarting the sum.
Fractions of a week survive the conversion
Up to eight decimals are kept, so the 0.044375 separating 13.04 weeks from 13 stays visible instead of being rounded away on every phase in turn.
Days and quarters in the same list
The dropdowns hold days for a weather allowance and quarters for a client reporting period, so one phase can be checked against all three without leaving the page.
Questions From the Site Planning Meeting
Why is the master programme in months when everything on site runs on weeks?
Because the two documents have different audiences. The master programme has to show an eighteen-month sequence, its overlaps and the critical path on a sheet a client can read in a meeting, and at week resolution that becomes 78 columns of unreadable bars. The site needs the opposite: a short window at enough detail to book gangs, plant and deliveries. The usual arrangement is a monthly master programme, a weekly resource schedule derived from it, and a rolling look-ahead cut from that. The conversion is the joint between them, which is why it is worth doing exactly rather than by eye.
The programme shows a month of float — how much is that in weeks on site?
Four and a third calendar weeks, or about 21.7 working days on a five-day site. That sounds generous until you notice what usually consumes it: two wet weeks in the envelope phase, one late information release and a delayed inspection will absorb a month of float without any single event looking serious enough to report. Convert float to weeks the moment it is granted, park it against named risks rather than leaving it drifting at the end of a chain, and review it weekly — float that is only checked monthly is normally spent before anyone notices it has gone.
How do I build weather days into a phase measured in months?
Convert the bar to weeks first, then subtract, because weather is lost in days and days only make sense against a week count. A four-month envelope phase is 17.3925 calendar weeks; if the contract or your own records allow fifteen non-working weather days across it, that is three working weeks gone, leaving roughly 14.4 productive weeks to resource. Stated back in months, the allowance alone is 0.68995257 mo. Winter phases justify a larger allowance than summer ones, so apply it per phase against the season it falls in rather than as a flat percentage across the job.
How far ahead should the look-ahead schedule run?
Two to six weeks is the working range, which is 0.45996838 to 1.37990513 mo of the master programme expanded into constrained, day-level detail. Two weeks suits fast fit-out sequences where crews change zone constantly; six weeks suits phases with long-lead materials or heavy plant coordination, because that is roughly the notice a supplier needs. Anything beyond six weeks stops being a look-ahead and turns back into the programme, since the assumptions underneath it will have moved before the window arrives. Whatever length you choose, refresh it weekly — a look-ahead updated monthly always lags the site.
A milestone has slipped two months — how much of the sequence do I re-plan?
Start by converting: two months is 8.69625 wk of re-sequencing, not eight, and that near-week difference decides whether a following trade returns before or after a shutdown. Then work out how much of the slip the following activities can absorb through their own float before the completion date moves. Where a lag was written into the logic — a curing period, a drying-out allowance, a testing hold — check it in weeks too, because those lags do not compress however the labour is rearranged. Only then reissue the programme, and reissue the weekly resource schedule with it; changing one without the other is how a re-planned job ends up double-booking gangs.
No comments yet. Be the first to comment!