Cash Flow Forecasting for Australian Construction: The 13-Week Rolling Model

May 27
7 min read

Why 13 weeks is the right window for construction cash forecasting

Most construction businesses run cash forecasting in two unhelpful modes. Either they forecast 12 months out at the start of every financial year and never touch the file again — useful for the board pack, useless for daily cash decisions. Or they reconcile yesterday's bank balance against today's payment commitments — useful for survival, useless for planning ahead.

The middle ground that works for AU construction operators between $2M and $50M revenue is the 13-week rolling cash forecast. Three months out, refreshed weekly, granular by week. Long enough to see the cash low-point three to four weeks before it arrives. Short enough that the underlying assumptions stay defensible.

This is the practitioner guide. The structure, the inputs, the refresh cadence, and the integration with retention, progress claim timing, and bank facility headroom. Builds on the Construction Accounting Australia cornerstone piece.

What goes into a 13-week forecast

The model has six rows running across 13 weekly columns:

Opening cash balance — Monday morning, week 1
+ Inflows — progress claim receipts, retention releases, daywork invoices paid, other income
− Outflows — subcontract payments, materials, payroll, GST/BAS, tax, finance, overheads
= Net movement
= Closing cash balance — Friday evening, week 1 = Monday opening week 2
Working capital minimum line — the floor below which you trigger covenant or facility issues

Each weekly column captures three to seven days of expected activity. The "low point" — the lowest forecast closing balance over the 13 weeks — is the metric the founder, the bank, and the auditor all care about.

Inflows — the construction-specific complications

Three categories of inflow need separate modelling:

Progress claim receipts

For each active project, work out when the next claim is expected and when payment typically arrives:

Claim issue date (per the head contract — typically monthly or by-stage)
Head contractor's standard payment terms (30 days, 45 days, 60 days from claim approval)
Approval lag (claims often take 5-10 days to be approved before the payment clock starts)

The total: claim issued 5 May, approved 14 May (9 days), payable 14 June (30 days net). That's 40 days from issue to cash. Model it as a known cash inflow on the week containing 14 June.

Each progress claim becomes a separate row in the forecast inflow stack — labelled by project + claim number.

Retention releases

From the active project register, retention release windows are typically:

50% at practical completion of the project
50% at end of defects liability period (typically 12 months after PC)

For each project nearing PC or end-of-DLP, model the retention release as a CONDITIONAL inflow — colour-code it differently from confirmed claims. Some operators discount conditional retention inflows by 10-20% to be conservative.

Daywork invoices and variation invoices

Less predictable timing than progress claims. Model based on the historical run-rate of daywork generated + invoiced. If the variation register shows $80k of approved-but-unbilled work, that's a near-term inflow once invoiced.

Outflows — the rhythm and the spikes

Subcontract payments

Match the SoPA timing for each state (see Subcontractor Compliance & PTRS). Most subs paid 30 days from claim approval. The trick: model subs by HEAD project so that when a head contract's payment receipt is forecast, the matching subcontract outflow is forecast in the same window. Mismatched timing = working capital strain.

Materials

Either 30-day suppliers (model as cluster-paid on the 15th of each month) or COD/cash-on-delivery for project starts. Major material orders (steel, concrete, fitout) often have large lump-sum cash outflows tied to project milestones.

Payroll

Weekly or fortnightly. Predictable. Include super guarantee (paid quarterly), PAYG withholding (paid monthly to ATO), and any leave provisions building up.

GST/BAS

Monthly or quarterly depending on turnover. The trap covered in GST on Variations and Retentions — you pay GST on the gross supply value, including retention you haven't received yet. Model the BAS payment timing precisely; surprise GST bills break cash forecasts.

Tax (PAYG instalments + company tax)

Quarterly PAYG instalments per ATO schedule. Annual income tax post-lodgement balance.

Finance costs

Loan repayments, lease repayments, line fees, interest. Mostly fixed and predictable. Forecast precisely.

Overheads

Office rent, insurance (monthly or annual lump), utilities, software subscriptions, professional services. Mostly small ticket but they aggregate.

The refresh cadence

Weekly — Monday morning

Every Monday before 10am:

Roll the model forward one week (drop the past week, add a new week 13)
Refresh opening cash balance against actual bank balance Friday evening
Refresh inflow expectations — confirmed claim receipts in the past week clear; expected timing of upcoming claims confirmed or adjusted
Refresh outflow commitments — confirmed payments past week clear; scheduled payments next 1-2 weeks confirmed
Compare new low-point to last week's. Material shift > $50k investigated.
If low-point dips below working capital minimum line within the 13-week window, escalate to founder + facility planning.

Monthly — alongside management pack

Once a month the model gets a deeper review:

Reconcile last month's actual cash movements to last month's forecast — variance analysis
Identify systematic over/under-forecasting (consistent errors that need pattern adjustment)
Stress test — what if a key receipt slips 30 days? What if a major contract pauses? Build downside scenarios
Update assumptions if the project portfolio has changed materially

The bank narrative — your 13-week forecast is your facility's protection

When a bank reviews your facility, the 13-week forecast is the single most-read document in the pack. Per Bank Covenants for Construction SMEs, bankers care about:

Forecast low-point — is it above the minimum working capital line?
Facility headroom — how much undrawn capacity exists at the low-point?
Material assumptions — what drives the receipt timing and outflow rhythm?

A clean 13-week forecast with documented assumptions + a track record of forecast accuracy keeps bank reviews trivial. A 13-week forecast where the low-point shows below the minimum line — without a mitigation narrative — triggers facility review.

Worked example — $12M EPC contractor, May-July 2026

Sample forecast snapshot, week 1 of 13:

Line	Week 1 May 5	Week 2 May 12	Week 3 May 19	Week 4 May 26	Week 5 Jun 2
Opening cash	$680,000	$610,000	$525,000	$420,000	$890,000
+ Progress claim receipts	$0	$0	$0	$580,000	$0
+ Retention release	$0	$0	$0	$0	$45,000
+ Daywork receipts	$15,000	$0	$0	$8,000	$0
− Subcontract payments	$25,000	$45,000	$60,000	$48,000	$32,000
− Materials	$28,000	$0	$0	$24,000	$0
− Payroll + super	$22,000	$0	$22,000	$0	$22,000
− BAS / GST	$0	$40,000	$0	$0	$0
− Overheads + finance	$10,000	$0	$23,000	$46,000	$18,000
Closing cash	$610,000	$525,000	$420,000	$890,000	$863,000
Working capital min line	$300,000	$300,000	$300,000	$300,000	$300,000
Headroom	$310,000	$225,000	$120,000	$590,000	$563,000

Reading this: the low-point is week 3 ($420,000 closing). Still well above the $300k working capital minimum, so $120k of headroom — comfortable. The $580k progress claim receipt in week 4 reinflates the position.

If the progress claim slipped — say, head contractor disputed it and approval got pushed to week 5 — the model would re-run with that delay and the new low-point would be week 4. Whether that's still above $300k depends on the subcontract outflow timing.

Common cash forecasting mistakes

Mistake 1 — Optimistic receipt timing

Operator models receipt date as day-of-invoice + 30 days. Reality is day-of-invoice + 5-10 day approval + 30 days. Forecast over-states cash inflows by 1-2 weeks.Fix: model based on head contractor's historical actual payment behaviour, not their contract terms.

Mistake 2 — Retention treated as confirmed

Retention release shown as definite inflow in the forecast. Defects rectification slips; release delayed 90 days. Forecast misses the gap.Fix: retention release as CONDITIONAL inflow (separate row or colour-coded). Don't count on it until released.

Mistake 3 — GST liability under-forecast

GST modelled at 10% of cash received. Actual: 10% of supply value (including retention). Surprise BAS bill breaks the forecast.Fix: GST on the gross supply, forecast in the BAS payment period.

Mistake 4 — Weekly granularity ignored

Forecast in monthly buckets, not weekly. Hides the within-month cash low-points (e.g., last week of month before head contractor pays first week of next).Fix: 13 weekly columns. The granularity exposes timing risk.

Mistake 5 — No working capital minimum line

Forecast shows cash position but no comparison to the minimum line that triggers covenant or operational issues.Fix: explicit minimum line row. Headroom calculated each week. Flag any week where headroom drops below $X (define $X based on facility size + risk tolerance).

Mistake 6 — Forecast set-and-forget

Built at year start. Never refreshed.Fix: Monday-morning refresh as a non-negotiable. 30-45 min weekly.

The downloadable template

The 13-week rolling cash forecast we install with every Financial Controller tier client is built as an Excel template:

Active projects register feed (rolls in claim timing automatically)
Subcontract payment matching by head project
BAS / GST / PAYG schedule built in
Sensitivity scenarios (base / upside / downside) toggle
Bank narrative auto-generation

Request the template via the Finance Function Maturity Audit form — it auto-sends on request.

When this becomes a specialist conversation

Operators below $5M usually start with a simplified monthly cash forecast. As you cross $5M-$10M and multiple concurrent projects appear, the weekly granularity becomes necessary. At Financial Controller tier ($4,950/month) we run the 13-week as a weekly process, integrated with the management pack and the bank narrative.

Above $25M revenue or with capital events approaching, Head of Finance tier ($8,500/month) adds 12-month forward forecast + capital structure modelling alongside the rolling 13-week.

TL;DR for the busy founder

13-week rolling cash forecast — Monday morning refresh, weekly granularity.
Forecast progress claim receipts based on actual historical payment timing, not contract terms.
Retention release as CONDITIONAL inflow — don't count on it until released.
GST on gross supply value (including retention), payable at BAS for the period of supply.
Working capital minimum line + headroom calculation each week.
If forecast low-point drops below minimum within 13 weeks, escalate to founder + facility plan.
Document forecast assumptions — banks want to see them during facility reviews.