Variance Screen Calculator

A report that explains 40 variances explains nothing. Enter your lines with the range each one normally moves within, and the screen separates the few that are genuinely signal from the many that are just a business being a business. Nothing is sent anywhere, the maths runs in your browser.

Filter 1: Materiality
In dollars, tied to decisions rather than to a percentage. A variance that cannot change an action you would actually take does not deserve investigation, however large the percentage looks.
Filter 2: The Lines and Their Normal Range
Line Type Budget Actual Normal range ±%
Normal range is how much the line usually swings month to month for ordinary reasons, estimated from history. It is the whole foundation: without it, every variance looks equally suspicious. Leave a row blank to ignore it.
What actually needs a person
1 to investigate

The month, screened
Line Variance Variance % Normal range Material Signal Verdict
Within normal range
Outside normal range
Material dollars
1Monitor. Large but normal.
1Investigate. This is the signal.
Immaterial dollars
1Ignore.
2Note, low priority.

Where the attention goes
Investigate

The trap this avoids
Reading by dollar size alone

The thinking behind this tool: Which Variances Deserve Your Attention →
It works through a real month where the second largest miss in the report is noise, and the line that matters is not the biggest number on the page.

The hard part is not the screen, it is the normal range. Establishing a defensible band for each line, and keeping the report to exceptions once the bands exist, is where this changes how a month actually gets read. Start a conversation →