Ask most finance teams what counts as a significant error and you get a number somebody chose years ago. It is usually round, it is rarely written down, and it is the same for a trading company with thin margins as for a manufacturer with thick ones. That is not materiality. It is a habit.

What ISA 320 actually asks for

International Standard on Auditing 320 defines materiality in terms of the decisions of the people who read the financial statements. A misstatement is material if it could reasonably be expected to influence those decisions. The standard does not give a number. It gives a discipline: pick a benchmark that reflects how the business is judged, apply a percentage that reflects its risk, and revisit the result when the business changes.

  • For companies judged on profit, a percentage of profit before tax, commonly around five percent.
  • For companies judged on scale or in a loss year, a percentage of revenue, commonly around one percent, or of total assets.
  • A lower performance materiality below that figure, so that several small errors cannot add up unnoticed.

Why one figure for everyone fails

A threshold that is right for one entity is wrong for the next. Set it high and a growing company reviews too little: duplicate payments, near-duplicate vendors and cut-off errors slip under the line month after month. Set it low and the team drowns in exceptions and stops reading them. Both failures look like diligence from the outside.

How we calibrate it

At onboarding we compute the threshold from the client’s own last closed year: the benchmark, the percentage, and a performance materiality below it. We write the calculation down, we review it every quarter, and we recompute it when revenue or profit moves by more than the percentage itself. The Auditor agent in MFT Intelligence carries exactly this number, so every transaction it flags is flagged against your threshold, not a default.

A threshold you cannot explain to your board in one sentence is not a threshold. It is a guess with a decimal point.

What changes for your team

Two things. First, the review becomes explainable: this entry was flagged because it exceeds a documented threshold derived from your own numbers. Second, approvals become proportionate: below the line one registered approver, at or above it a second person. That is the same logic MFT applies when an agent drafts a correcting entry and a person approves it.