Quarterly forecasting, explained

The short answer

Quarterly forecasting means rebuilding the rest of the year once a quarter. The quarter that just closed is replaced with actuals, the remaining quarters are re-estimated in the light of what those actuals taught you, and the output is a single line everyone can use: where the year lands if the business keeps doing what it is doing now.

The annual budget cannot do this job, because it is written once, before the year starts, when every number in it is a guess. By April the guesses have met reality. The budget stays useful as the promise you made, but steering by it is steering by a photograph, and the quarterly forecast exists to put a live picture next to it.

Budget and forecast are two different documents

Most of the confusion around forecasting comes from treating these as one thing. The budget is the commitment. It is written before the year begins, it carries the targets pay and plans were set against, and it does not move. The forecast is the current expectation. It is rewritten every quarter and it is allowed to disagree with the budget, because that disagreement is the point.

The gap between the two is the most useful number in the pack. A forecast tracking 8 percent under budget in April is nine months of time to do something about it. The same gap discovered in December is just a bad year, fully formed.

How to run a quarterly reforecast

On clean books this is about a working day, four times a year.

1. Close the quarter first. A forecast built on unreconciled numbers inherits their errors, and a proper monthly close is what makes this step already done.
2. Drop actuals into the quarter just finished. Leave the budget column exactly as it was written.
3. Re-estimate the remaining quarters from drivers, not from last quarter's forecast. Customers won and lost, pricing, headcount, capacity: change the causes and let the numbers follow.
4. Shape the remaining quarters with your seasonal pattern rather than dividing the year by four. Reading seasonality in your numbers shows how to work that shape out.
5. Write down what changed since the last forecast and why, in five lines or fewer. Next quarter, that list is how you find out whether your forecasting is getting better.

Keep the output small: revenue, gross profit and cash for each remaining quarter, plus the year-end landing. A forecast with three hundred rows is not more accurate than one with twelve, it is only harder to update, and the forecast that is hard to update is the one that quietly stops happening.

Roll it forward four quarters

The calendar-year version of this has a design flaw: every reforecast looks at a shorter window. By the fourth quarter you are forecasting three months and planning nothing, and January arrives unexamined.

The fix is to make the window roll. Always forecast the next four quarters, whatever the calendar says, so a reforecast in September already holds a view of next spring. Hiring, leases, borrowing and pricing decisions all reach further than December, and a rolling window means the forecast always has a full year to stand on.

Which tool answers which question

Quarterly forecasting sits in a stack, and it helps to know which layer you are asking for.

The monthly close says what actually happened, reconciled and final. It is the floor the other two stand on.

The 13-week cash flow forecast is the survival document: cash by week, one quarter out, built to find the week the balance gets tight before that week arrives.

The quarterly forecast is the direction document: revenue, profit and cash by quarter, built to say where the year lands and what to change while changing it is still cheap.

A growing business wants all three, on their own clocks: the close monthly, the cash grid weekly, the reforecast quarterly. Skipping the middle one is how a business hits a profitable year and still misses payroll.

Four traps that ruin it

1. Rewriting the budget to match the forecast. It makes the variance report tidy and deletes the one number that mattered, the gap. The budget stays frozen; the forecast does the moving.

2. Nudging last quarter's forecast by a percentage. That is not forecasting, it is anchoring. If nothing about the drivers changed, say so and keep the number; if something changed, size that thing rather than adjusting the total until it looks plausible.

3. Forecasting the revenue without the costs that produce it. Growth arrives with hires, stock and equipment that get paid for before the revenue lands. Can I afford to make this hire? walks the sharpest version of that timing.

4. Optimism as a policy. A forecast that always says the second half will fix it teaches everyone to ignore the forecast. The reforecast note from step five is the cure: when the last three all over-promised, the next one has to say why this time is different.

What it changes

Decisions move earlier. The hire, the price move, the cost cut: each gets decided against a live year-end landing instead of a December surprise.

The bank conversation improves. A lender asking for forward numbers gets this quarter's forecast instead of a stale budget, which reads as a business that can see where it is going.

The board conversation changes subject. With a budget alone, meetings argue about whether the plan was wrong. With a forecast beside it, they discuss what to do next, which is the conversation the meeting was for.

Bad news gets a date while it is still cheap. Most expensive problems were visible two quarters early to anyone holding a current forecast. The document's whole value is buying back that time.

How ClarIQ handles it

In a ClarIQ engagement the reforecast rides the rhythm that already exists: the books are closed monthly, so quarter-end forecasting starts from reconciled numbers the same week, and the memo carries the year-end landing next to the plan with the gap named in plain language. The forward cash view runs continuously underneath it as the 13-week grid.

If you want the backward half answered on your own figures first, the free month check takes five numbers and comes back in about two minutes.

Published September 1, 2026.