Your P&L and your bank account answer different questions
A profitable business can run out of cash. This is not a paradox, it is arithmetic. Revenue lands on the income statement when you earn it; cash lands in the account when someone pays. Between those two moments sit progress payments, retainage, customer terms, payroll cycles, and a tax calendar that does not care what your receivables look like.
Contracts that look healthy are exactly where this bites hardest, because the bigger the job, the longer the gap. You have front-loaded costs — labor, materials, subs — and back-loaded collections. That is a financing requirement, whether or not anyone called it one.
Job costing tells you whether a job made money. A cash forecast tells you whether you can fund it while it does. You need both, and they are not the same report.
A quarter, in weeks
The number is not arbitrary. Thirteen weeks is the shortest horizon that covers a full operating cycle and the longest one you can still forecast with any honesty.
Near weeks are actionable
In the first few weeks you know what is scheduled, what is invoiced, and what is due. That is precision you can move money against — shift a payment, time a draw, delay a purchase. Monthly forecasting is too coarse to see the week payroll and a vendor run collide.
Far weeks are early warning
A shortfall thirteen weeks out is a planning problem — you can line up a line of credit, slow a hire, or accelerate billing. The same shortfall next Tuesday is an emergency. The horizon is what converts one into the other.
Weekly grain, not monthly
A monthly forecast averaged everything into a number nobody could act on. Weekly columns show you the specific week the account dips, which is the only version of the information that changes a decision.
It rolls, it does not end
Every week you record what actually happened, explain the differences, and push the window forward one week. It is a discipline, not a document — which is exactly why one built once and filed away tells you nothing.
Direct method, or it is fiction
There are two ways to build a cash forecast, and only one of them works. The indirect method starts from net income and adjusts for non-cash items. It is fine for a year-end statement and useless for deciding what to pay this week, because net income is an accrual concept — it includes revenue you have not collected and expenses you have not paid.
The direct method ignores the income statement entirely. You list money actually coming in and money actually going out, week by week, from the documents that create the movement: open invoices, scheduled payments, payroll calendars, the recurring-payment register, and debt service dates. Cash in, cash out, ending balance. Nothing in between.
The test is simple. If a line in your forecast cannot be traced to a specific invoice, contract, or calendar date, it is an assumption wearing a number's clothing.
What goes in, in order of reliability
Weeks 1–2 — scheduled payments. Checks cut, ACH queued, payroll runs dated. This is the only part of the forecast that is closer to a fact than a guess.
Weeks 2–5 — approved open invoices with realistic payment timing, not stated terms. If your customers pay in 58 days while the terms say 30, forecast 58.
Weeks 3–6 — received but unapproved invoices, with an approval-lag assumption. The lag is real and usually longer than anyone admits.
All weeks — the recurring register. Rent, insurance, software, loan payments, lease payments. These are boring and they are the ones that break a forecast when forgotten.
Weeks 5–13 — pipeline conversion. Open quotes and POs modeled at a conversion rate you can defend. This is the softest layer, and it should be labeled as such.
Payroll, taxes, and debt service come from their own calendars — never from an average. A quarterly estimated tax payment does not belong in a weekly smoothing formula.
Most 13-week guides assume a treasury department you do not have
The standard advice is written for companies with a controller, a weekly liquidity call, and a revolving credit facility whose availability they forecast alongside cash. That is a real discipline and it is genuinely useful — at a company with a treasury function.
An owner-operated business is a different shape. There is no revolver to draw on. No one is running a payment-timing committee. The person reading the forecast is the same person running the crews, quoting the jobs, and signing the checks. Enterprise forecasting guidance quietly assumes away the constraints that define your actual problem.
What that means in practice: the forecast has to survive being built by someone who has ninety minutes a week, from data that lives in three different systems, with no analyst to clean it up. Anything more elaborate gets abandoned by week three — and an abandoned forecast is worse than none, because it produced false confidence before it died.
So the version worth building is the smallest one that still tells the truth: a starting balance, expected collections by week, committed outflows by week, and the resulting low point. That is enough to see the collision coming. Everything beyond it is sophistication you will not maintain.
Three things make trade cash flow harder than most
Retainage is earned money you cannot spend
Five or ten percent withheld on every pay application is real revenue that sits outside your reach until the job closes out — often long after. Inside ordinary receivables it makes your collections look healthier than your account will ever be. Forecast it separately or the forecast lies to you.
Progress billing lags the labor
You pay your crews weekly and your subs on their terms. Your customer pays on the pay-application cycle — submit, review, approve, disburse — which runs weeks behind work already performed. The gap is not a mistake; it is the structure of the industry, and it has to be financed deliberately.
The tax calendar does not follow the job calendar
Estimated taxes, payroll deposits, and state filings land on fixed dates regardless of whether a draw came in that week. In a heavy month those obligations can stack against payroll in the same seven-day window. Seeing that collision thirteen weeks early is the entire point.
This is where job costing and cash forecasting have to meet. Job costing tells you a job earned its margin; the forecast tells you whether you can carry it to collection. A job that is profitable and unfundable is still a job that can take the business down.
Variance review is where the forecast actually improves
Building the first version is the easy half. The value comes from the weekly loop: record what actually hit the account against what you projected, then explain the differences before rolling the window forward one week.
You are not doing this to grade yourself. You are doing it because your misses are systematic, not random. If collections consistently land two weeks later than modeled, that is not variance — that is a wrong assumption you keep writing down. The review is how it gets corrected.
Cap it. Pull the ten largest misses, ask why each one moved, and stop. A review that tries to reconcile every line takes a day and gets abandoned; ten lines takes twenty minutes and survives.
After a few cycles the forecast stops being a projection and becomes a control — you can see a problem forming in week nine and deal with it in week nine, while it is still just a decision.
A forecast that only ever looks fine is not being read
The first honest forecast usually shows a squeeze somewhere. That is the forecast working, not failing — you have found a problem while it still has options attached.
The levers, roughly in order of how little they cost you: accelerate collections on work already done (bill the underbilled, chase the aged), re-time discretionary outflows that week, stage larger purchases around the dip rather than into it, and make sure a tax or debt date is not landing on top of a payroll run when it did not have to.
What you do not do is treat the operating account as a reserve and discover the reserve was customer deposits. That failure mode is covered on the job costing page, and it is the most common way a busy contractor finds trouble.
Three tiers, one flat monthly fee
Cash forecasting sits in the Advisory and CFO tiers. Assessment call → financial review → custom plan. Elev8 does not perform audit, review, or compilation. See if we're a fit →
Questions, answered
Why 13 weeks and not a full year? +
Thirteen weeks is one quarter in weeks, and it is the shortest horizon that covers a full operating cycle while staying accurate enough to act on. Beyond a quarter, precision collapses into guesses. Inside a quarter, you know what is invoiced, what is scheduled, and what is due — which is the detail that lets you move money deliberately instead of reacting.
How is this different from my profit and loss statement? +
They answer different questions. Your P&L is accrual-based and reports revenue when earned, whether or not it has been collected. A cash forecast reports money actually arriving and leaving. A profitable business can still run short of cash — that is normal arithmetic in any business with a lag between doing work and getting paid.
I do not have a controller. Can I realistically maintain this? +
Yes, if you build the small version. Starting balance, expected collections by week, committed outflows by week, and the resulting low point. That takes about ninety minutes a week and covers the decisions that matter. Elaborate models that need an analyst get abandoned by week three.
What happens when the forecast shows a shortfall? +
That is the forecast working. A problem nine weeks out is a planning decision — accelerate billing on work already done, re-time a discretionary purchase, line up a facility. The same shortfall next Tuesday is an emergency. The horizon is what converts one into the other.
Should I include revenue I have quoted but not won? +
Only in a clearly labeled layer, at a conversion rate you can defend from your own history. Pipeline belongs in weeks five through thirteen at best, never in the near weeks. Mixing hoped-for work into weeks one and two is how a forecast becomes a wish.
How accurate is a 13-week forecast supposed to be? +
The near weeks should be close; the far weeks should be directionally honest. If weeks one and two are not reliable, the problem is your data plumbing, not the model. Expect to be wrong in the far weeks — the variance review is what teaches you how you are wrong, and that is what corrects it.
We do retainage-heavy work. Anything different? +
Track retainage separately from receivables, aged by job. It is revenue you have earned and cannot collect yet, so inside ordinary AR it makes collections look better than they are. In the forecast it belongs on its own line with the close-out timing you actually expect, not the terms you were given.
Do you build these forecasts for clients? +
Yes. Cash forecasting sits in the Advisory and CFO tiers, built from your real payables, receivables, payroll, and debt service rather than a template. Elev8 is a service-area practice — remote and on-site across Texas, New Mexico, and Colorado.