Your Cash Forecast Is Reading the Wrong Terms

A trades company's thirteen-week cash forecast would not reconcile to its own revenue. The problem was not the model.

Invoices had been recorded at net 30 and net 45 when the terms actually being sold were net 15. The forecast was reading the system correctly. The system was wrong, and had been for long enough that nobody thought to check it.

Key Takeaways

• Your forecast pulls payment terms from the invoice record, and nobody audits whether those match the terms you actually sell on
• Large purchases reach finance when the payment is due, not when the decision is made
• Your aged receivables include balances that were never collectible in the first place
• You check the bank balance because you do not trust the accounting system, and you are half right

What Do Financial Review and Analysis Services Actually Do?

They make a company's own numbers trustworthy enough to decide from, and then build the forward view on top of that. The forward view is the visible part; the reconciliation underneath it is the part that determines whether any of it is true.

This matters more in trades than in most industries because the operating system and the accounting system are usually different products, maintained by different people, with different ideas about when something happened. A forecast is a downstream consumer of both. It inherits every disagreement between them without flagging that it has done so.

So the honest description of the first month of this work is not forecasting. It is auditing the inputs a forecast would otherwise trust.

Why Doesn't My 13-Week Cash Forecast Match Reality?

Usually because it is reading inputs that were set up once and never revisited. Payment terms, vendor records, and bank accounts get configured at some point in the past and then drift from how the business actually runs, silently, because nothing in a normal month forces a comparison.

A trades client of ours could not reconcile its thirteen-week forecast to recorded revenue. Invoice terms had been entered inconsistently across the customer base — net 30 and net 45 against a net 15 standard — and one significant account's terms had been changed to net 15 several months earlier without the forecast assumptions ever following. The correction was not a better formula. It was to model collections from actual collection history rather than from the terms stored on the invoice.

That distinction holds up well beyond this one company. Stated terms describe an agreement. Collection history describes a behaviour. Only one of them is a forecasting input.

How Do You Forecast Spending Nobody Told You About?

You cannot, so the forecast needs a rule rather than a better guess. Vehicle purchases and similar commitments typically reach finance when payment comes due, which is far too late for a thirteen-week view to be worth anything.

The same trades client put a documented process on purchases above $1,000, with approval layers scaled to the amount, and added a purchase cutoff in the final days of each period so that material buying stopped distorting both cash and the close. The intent was explicitly visibility rather than blocking spend — justified purchases still happen, they just stop arriving as surprises. Branch-level inventory reporting was added alongside it so that month-end buying could be checked against what was already on the shelf.

It is worth naming the context, because it changes how seriously this gets taken. A bank had recently declined to extend the company credit. At that point cash discipline stops being housekeeping and becomes a financing question with a deadline attached.

Why Is My Accounts Receivable Balance Not Real Money?

Because an aged AR report mixes three different things: debts customers owe, amounts still being earned, and artifacts of how the system was configured. Only the first is collectible, and the report does not distinguish them.

A plumbing client of ours found roughly half of its open invoices aged past 120 days. A single $29,000 balance traced back to a billing-location setup that had routed invoices to the wrong customer record rather than to anyone refusing to pay, and about $10,300 of the total belonged to active work that was not yet billable at all. There was also an unresolved question about when the clock on net-45 terms started — invoice date, approval date, or the date a document was accepted into a customer portal — which is the entire difference between a receivable that is late and one that is not. Late fees were printed on every invoice and had been charged almost never.

None of that is a collections problem, and treating it as one costs a company credibility with customers who do not actually owe anything.

Should You Trust QuickBooks or the Bank Balance?

Neither, until they have been reconciled to each other — and that reconciliation is the work, not a formality before the work. Owners who check the bank instead of the accounting system are usually right that the system is wrong and wrong that the bank balance is a substitute for it.

An automotive services client of ours ran a thirteen-week forecast against a target of roughly $205,000 a week in net revenue while actual receipts ran closer to $150,000 to $160,000. Underneath, the accounting was running about two weeks behind, a duplicated bank account was overstating cash, and individual transactions were landing in the wrong place — a detailing charge recorded as parts, a lease missing from the books entirely, marketing understated. The owner had stopped reading the reports and was judging the cash position from the bank and his own familiarity with the business.

The first assignment there was not a better forecast. It was getting the month closed and the bank reconciliation produced, because that is the document that makes every other number arguable rather than merely asserted. Trades-specific financial leadership tends to start in exactly this unglamorous place.

What Does a Forecast Change Once the Inputs Are Right?

It stops being a document and becomes a short list of real choices. At the automotive client, weekly inflows around $155,000 against a requirement near $205,000 made the options explicit: raise revenue, cut cost across materials, overhead, marketing and owner distributions, or both.

It also made one specific decision easy, which is the better test of a forecast. A large lump-sum sales tax payment was on the table. The model showed it could not be absorbed at current receipts, so the decision was to spread the paydown across months instead. That is a forecast doing its job: not predicting the future accurately, but making one week's choice obvious.

Thirteen weeks is the working horizon because it is long enough to reveal a developing shortfall and short enough that the commitments inside it are still concrete. Treat the first four weeks and the last nine differently — the near weeks are mostly known obligations, and presenting the later ones with the same apparent confidence trains everyone to discount the whole document.

What Should a Trades Business Expect in the First Quarter of This Work?

Expect the first month to be unglamorous and slightly annoying. It is bank reconciliations, coding corrections, terms audits, and a list of things that were recorded wrong for longer than anyone is comfortable admitting.

Expect the second month to produce the first genuinely useful forecast, and expect it to say something you did not want to hear. In every one of the cases above, the corrected picture was worse than the reported one, which is the normal direction — errors that flatter a business survive longer than errors that do not, because nobody investigates good news.

Expect the third month to be about discipline rather than analysis: who approves spend, who codes what, and what happens on a fixed day each week. That is where the forecast stops degrading, and it is the part most companies skip.

Your Next Step

A Forecast Built on Unverified Terms Is a Confident Guess

If you could not say today whether the payment terms in your system match the terms you actually sell on, your cash forecast has a problem no model will fix. Cash visibility is one of the dimensions the Financial Control Score Quiz scores directly. Seven questions, about a minute, no cost.

Take the Financial Control Score Quiz Book a Discovery Call

Common Questions

Frequently Asked Questions

Why Doesn't My Cash Flow Forecast Match My Bank Account?

Most often because the forecast trusts inputs nobody has audited: payment terms stored on invoices, duplicated or stale bank accounts, and transactions coded to the wrong category. The model is rarely the problem. The records it reads usually are.

How Accurate Should a 13-Week Cash Forecast Be?

The first four weeks should be close to exact, because they are mostly known obligations. The last nine are estimates and should be presented as such. A forecast that shows uniform confidence across all thirteen weeks teaches the team to discount the whole thing.

What Is the Difference Between Reactive Reporting and Proactive Forecasting?

Reactive reporting explains a period that has closed. Proactive forecasting changes a decision that is still open. The practical test is whether anything in next week's schedule, spending, or collections changes because of the document you just produced.

How Do You Know Which Receivables Are Actually Collectible?

Work the aged report line by line rather than by bucket, and separate three categories: genuine debts, work still being earned, and system artifacts such as invoices routed to the wrong customer record. Buckets hide exactly the balances that most need a decision.

Should a Trades Business Hire a Bookkeeper or a Fractional CFO First?

If the books are late or unreliable, hire the bookkeeping capability first, because forward analysis built on an unreliable close produces confident wrong answers. Once the close is dependable and on time, the constraint moves to who decides what to do about it.