Cash Flow Forecasting for Construction: A PMC Handbook

Construction is a cash-flow business wearing a project-management costume. Profitable projects fail because the money arrives later than it leaves; healthy-looking portfolios seize up because three clients paid late in the same month. Yet most cash-flow "forecasting" in construction is a spreadsheet updated in a panic at month-end. This handbook lays out how a PMC builds a forecast that actually predicts, not just records.

Why construction cash flow is uniquely hard

Three structural features make it harder than most industries:

• Front-loaded costs, back-loaded receipts. You mobilise, buy long-lead items and pay subcontractors before the client's payment certificate clears. The gap is financed by you.
• Retention. A slice of every certified amount is held back — often 5–10% — and released only at practical completion and end of the defects period. That's real earned money you can't spend for months or years.
• Certification lag. Work done ≠ cash received. There's a valuation, a certification SLA, and a payment period between them, and each hop is a place for delay.

The three-month rolling model

The workhorse of PMC cash management is the three-month rolling forecast, refreshed weekly or fortnightly. It answers one question: will we have cash to meet obligations over the next 13 weeks?

Build it in three layers:

• Committed outflows. Subcontractor payments due, material POs, payroll, plant hire, overheads. These are the most certain; anchor the model here.
• Expected inflows. Certified amounts by their contractual payment date (valuation date + certification SLA + payment period), not the optimistic "we invoiced, so it'll come." Apply a realistic collection assumption per client based on history.
• Retention schedule. Model retention held and its release triggers (practical completion, DLP end) as separate, dated line items. It's cash, just deferred.

Payment-milestone planning

The single biggest lever on construction cash flow is when milestones are defined to trigger payment. A milestone worth 15% of contract value that lands in week 20 vs week 14 is six weeks of financing you either carry or don't.

• Align payment milestones to genuine value points — mobilisation, foundations complete, superstructure, MEP first fix, handover — so certification is objective and disputes are fewer.
• Front-load fairly where the contract allows (mobilisation advances, materials-on-site payments) to offset early outflows.
• Map each milestone's payment to the forecast so a slipped milestone visibly moves the cash date, not just the programme.

Variance tracking: forecast vs actual

A forecast you never check against reality is a wish. Every cycle, compare:

• Forecast inflow vs actual received. By client. A persistent negative variance for one client is a collections problem to act on, not absorb.
• Forecast outflow vs actual paid. Catches under-forecast commitments.
• Days Sales Outstanding (DSO). How long certified money actually takes to arrive. Rising DSO is an early liquidity warning.
• Forecast accuracy over time. If you're consistently 20% optimistic on inflows, bake that into the collection assumption.

Connecting cash flow to the project data you already have

The reason cash-flow forecasting is a month-end scramble is that the inputs live in different places — the programme, the budget, the certificates, the retention register. Pull them together and the forecast largely builds itself:

• Certified amounts and payment terms come from your payment certificates.
• Retention comes from the certificate mechanics.
• Timing comes from the programme (milestone dates) and the client's payment regime.

A practical operating rhythm

• Weekly. Refresh the 13-week rolling forecast from actuals; flag any inflow slipping past its date.
• Per certificate. Update the forecast the moment a valuation is certified — that's when the payment date becomes real.
• Monthly. Review variance by client, DSO trend, and retention due for release; escalate slow payers.
• At milestone slip. Re-forecast the linked receipt immediately, so the cash impact of a programme delay is visible the same day.

The bottom line

Cash-flow surprises are a data-latency problem, not a finance-skill problem. Build a three-month rolling model anchored on committed outflows and realistically-dated inflows, tie payment milestones to the programme so slips move the cash date automatically, and track forecast-vs-actual until your assumptions are honest. Do that and month-end stops being a scramble — because you already knew.