How to Audit a Budget or Forecast in Excel

Budgets and forecasts carry weight: people commit resources based on them. Auditing one means confirming the numbers are not just plausible but actually flow from defensible inputs. This guide walks through a structured audit, from reconciling totals down to tracing the formulas that build them.

Reconcile the Totals First

Begin where the stakes are highest: the summary totals. Confirm the grand total equals the sum of its parts and that period subtotals roll up correctly.

A budget that does not internally reconcile has a structural problem worth finding before you check anything else.

1. Select each subtotal and confirm its SUM range covers exactly the right rows
2. Check that subtotals add up to the grand total
3. Confirm annual figures equal the sum of their periods
4. Note any total that is a hardcode rather than a formula

Identify the Drivers

Every budget rests on a few key drivers: headcount, unit volume, a growth rate, a price. Find them, and confirm they live in clearly labeled input cells rather than buried in formulas.

If a driver is typed inside a formula, it cannot be flexed and is almost certainly undocumented. That is your first finding.

Trace the Drivers Through to the Totals

Pick a headline line, like total operating expense, and trace it back through its precedents to the drivers. This confirms the forecast is actually built from the inputs, not pasted over them.

Trace Precedents under Formula Auditing can show the links, but the arrows crowd a dense budget grid. Walking the chain one layer at a time keeps it legible.

1. Select a key output line in the budget
2. Trace its precedents to the line items feeding it
3. Step into each line item and trace it to its driver
4. Confirm the path ends at a labeled input, not a hardcode

Check the Period Logic

Forecasts live and die on period logic. Confirm that each period correctly carries forward from the last, that the opening period is handled deliberately, and that no column silently breaks the pattern.

Watch especially for the first forecast period, where the formula often differs from the rest of the row.

• Opening balances reference the right prior period
• Each period column uses the same formula as its neighbors
• Seasonality or ramp logic applies consistently
• No period accidentally references a fixed cell instead of a relative one

Pressure-Test the Forecast

Finally, flex the key drivers and watch the totals respond. A 10% change in volume should move revenue by a sensible amount. If it does not, a link is broken or a hardcode is overriding the logic.

This last step is what separates an audit from a glance. It proves the model behaves, not just that it adds up.