Budget split calculator

Total budget is the money being divided, for one period, in one currency.

First, second and third share are percentages of that total. The labels are deliberately blank of meaning: the lines might be channels, campaigns, funnel stages, markets or months, and the arithmetic does not care which. What it cares about is that they come to a hundred.

Every other calculator on this site answers a question. This one also refuses to answer a question that is not yet well formed.

Shares that total 90 leave a tenth of the money unassigned. That is either a line somebody has not named, a figure typed short, or a plan that quietly intends to hold something back — and each of those needs a different conversation. Shares that total 110 are a plan to spend money that does not exist, which is a mistake nobody notices in a percentage column and everybody notices at the end of the quarter.

So the amounts stay blank until the shares are a split, and the sum is shown above them so the page says what is wrong rather than going silent. It is the same principle as a ratio with no denominator rendering as a dash: printing a number that was not really computed is worse than printing nothing.

The check catches one particular error often enough to be worth naming. A plan built by adjusting shares one at a time — taking five points off one line and giving them to another — drifts away from a hundred the moment somebody adjusts a third line without taking the points from anywhere. The percentages still look reasonable individually, the column still looks like a plan, and the total is the only thing that says otherwise.

The temptation is to allocate in proportion to what the reporting says worked, and reported performance is not a neutral guide. The lines that capture demand people already had — the ones that intercept somebody who has decided — report the best numbers in nearly every account, because the revenue was closest to already happening. Allocating strictly on those figures moves money toward harvesting and away from creating demand, and the account gradually shrinks its own future while every dashboard improves.

A defensible split usually answers three questions rather than one.

What is already saturated. A line where the next unit of money reaches people the previous unit already reached is finished, whatever its average return says. The marginal figure decides this, not the average.

What has never been tested. A share set aside for things that might work is not a rounding error; it is the only way anything new is ever found. It should be small, deliberate, and written down as an experiment with a date rather than as a line that quietly becomes permanent.

What the business needs to learn. Sometimes the right allocation is the one that answers a question — whether a market works, whether an audience exists, whether a price holds — and the answer is worth more than the incremental return.

Set once and inherited. A split written eighteen months ago is a set of assumptions about a market that has moved. If nobody can say why a line has its share, the share is a habit.

Proportional to last period's results. As above: it compounds toward whatever measures best rather than toward whatever works best, and the two are not the same thing.

Too many lines. Nine lines at the top level usually means the first decision has not been made. Three forces it into the open, and any line can be split again by taking its amount as a new total.

Even splits as a default. Three equal shares is occasionally the right answer and usually a way of avoiding the argument.

Changed too often. A line needs long enough to produce a readable result, and automated bidding needs stable conditions to work. A split revisited weekly is a plan nobody is allowed to execute.

Confused with pacing. A split decides how much each line gets. It says nothing about when that money is spent inside the period, which is a separate decision with its own arithmetic — and a line that is on plan for the quarter while spending nothing for three weeks and everything in the fourth is not executing the split it was given.

Once the shares are a split, the amounts are the inputs to everything else on this site. Each one divided by the days in its flight is a daily budget. Each one against the outcomes it is expected to produce is a cost per result to plan against. Each one against what a customer is worth says whether the line can work at all before it runs.

Write the split down with its reasoning, and review it on a stated interval. The reasoning matters more than the numbers: a split whose shares are recorded without a sentence each is a plan that will be argued from memory in three months, and memory will favour whichever line had the best month.