Spend forecast calculator

Spent so far is the actual spend for this period to date, across whatever the forecast covers — one campaign, one channel or the whole account — in one currency.

Days elapsed and days in the period are the real day counts of the flight, not a calendar month. A flight that ran nine days of thirty is what the example uses; a flight that started mid-month and ends on the 30th has its own numbers, and dividing by thirty when the period is nineteen days produces a forecast that is wrong by a third.

Budget is optional and produces the variance, which is the line that turns a projection into a decision.

The forecast assumes the remaining days resemble the elapsed ones. That is a strong assumption stated plainly, and it is right often enough to be useful.

It breaks in four predictable places. A period that has so far been all weekdays and ends over a weekend; a public holiday inside the remaining days; a payday or a sale that reliably moves demand; and the first two or three days of any period, when a single unusual day dominates the average.

None of those is a reason to distrust the arithmetic. They are a reason to read the forecast as the neutral case, and to weight the remaining days by hand where the calendar is telling you something the average cannot know.

A forecast of 42,000 against a budget of 40,000 is not an alarming percentage. It is 2,000 of real money, and stating it that way changes what happens next.

It also sizes the correction. Two thousand over, with twenty-one days left, is about a hundred a day — an adjustment nobody will notice. The same two thousand discovered with four days left is five hundred a day, which is more than a third of the daily rate and will be felt in delivery, in cost per result, and in whether the campaign is still learning when the period ends.

That asymmetry is the argument for checking early. The cost of a pacing correction rises sharply as the days run out, and the correction that is cheapest is the one made when it was smallest.

An overspend forecast gets attention because it costs money. An underspend forecast is usually read as caution and is more often a failure, because the budget was approved to buy something and the unspent money bought nothing at all.

The causes are finite and none of them is fixed by raising a daily ceiling: a bid below what the auction is charging, an audience too small to absorb the money, a schedule that excludes most of the day, an ad still in review, a payment problem, or a bid strategy delivering conservatively because it has not seen enough conversions to settle.

The order is always delivery first, ceiling second. A campaign that is trying to spend and cannot will keep not spending at a higher ceiling, and the period will end with the forecast exactly where it was.

The one honest exception is a ceiling that is genuinely binding: a campaign hitting its daily cap every day by mid-afternoon and stopping is being limited by the number in the box, and there the forecast is telling you something real about demand the account is refusing to buy.

A daily ritual. Automated bidding needs stable conditions, and a material budget change can return a campaign to a learning state. A forecast checked and acted on every day produces worse delivery than one checked weekly and acted on once, because each correction disturbs the process the previous one was meant to improve.

A reason to level the days. Even pacing is the honest default, not the best plan. Demand is not flat across a week or a month, and a plan that deliberately weights the days that convert — while holding the period total — will beat a flat one. The forecast still works as the control, because it is comparing against the total rather than against an assumption of evenness.

A substitute for a reserve. Holding a few per cent of the period back covers an auction that turns out more expensive than planned. A period run with the entire budget committed to a schedule has no answer when the market moves, and the answer it is forced into is a stop.

Pacing is a budget question and not a performance one, and it is worth saying because the two get conflated in the same weekly meeting. A campaign running over budget while producing results below its ceiling is a campaign to fund, not to cut; one running under budget while producing excellent results is money being left on the table. The forecast says where the spend lands. What the spend is buying is a different number, and the decision needs both.