KPI and Variance Analysis Built on Verified Actuals

Variance analysis has a trust problem that has nothing to do with the math. It’s easy to compute “actual minus budget.” It’s much harder to know, at a glance, whether the number sitting next to it is a verified actual, an unaudited plan figure, or a forecast someone typed in last quarter and forgot to update. Rexfin’s KPI and variance views are built to make that distinction unmissable, not to bury it in a footnote.

The grid, and what each column actually is

The core view lines up Actual against Budget for the current fiscal year, the variance between them, and Forecast for the year ahead. The budget and forecast columns carry a visible UNAUDITED tag, always. That’s not a formality: plan figures are never persisted as verified cells, so they’re structurally excluded from the same citation and export checks that apply to actuals. Only actuals get the full treatment: click-to-source, verified status, the works. The point is that you should never be able to mistake a plan number for a reconciled one just because they’re sitting in adjacent columns.

Variance itself gets a direction and a color, and the color accounts for what “good” means for that line: a cost line running under budget is favorable, so the same arithmetic sign that would flag revenue as a miss reads as good news on an expense. And variance only renders when the actual column is genuinely the budget year being compared: Rexfin won’t quietly compute a variance between mismatched periods just because two numbers happen to be in the same row.

When the plan data isn’t there

Sometimes a plan layer is missing, malformed, or rejected outright by the checks that guard it. Rexfin doesn’t let the column just disappear: a labeled alert explains that the plan layer is unavailable and why. A blank space where a number should be looks like a bug or, worse, like nothing was ever budgeted. A visible message says plainly that data is missing, which is a very different fact than “the budget was zero.”

Every actual traces back, and the KPI math is defined once

Every actual figure behind a KPI or a variance calculation traces to a specific source-verified data point: the same provenance chain that backs the statements themselves. Flag a line as a material variance and it’s still click-to-source, not a summary number floating free of where it came from. The KPI math underneath is defined once and reused everywhere it appears, so the same metric can’t quietly mean something slightly different on two different screens.

Period comparisons respect the entity’s actual fiscal calendar rather than assuming a clean January-to-December year, and Rexfin refuses to compute a comparison across periods that don’t genuinely line up: a partial month isn’t measured against a full one, and a comparison doesn’t happen at all if the periods don’t match.

The multi-year trend view

Beyond the current-year grid, actuals from every filing an entity has submitted pivot into a multi-year trend, with a restated year giving way to whichever filing reported it most recently, so an old, superseded number doesn’t linger in the trend line after a correction. Year-over-year change is computed only between consecutive years, and compound growth is computed only where the endpoints are genuinely positive and the elapsed time is real, not assumed. Every year in that trend line is still clickable back to the atom it came from.

Who this is for

FP&A teams and controllers who’ve been burned by a variance dashboard that looked clean but couldn’t say where a number actually came from, or that quietly kept showing last year’s plan after the real one loaded. It’s for anyone who needs “cost ran over budget by 8%” to survive the follow-up question: over which budget, as of when, and traced from where.

See how budget approvals and segment-level detail feed into this same variance view, or book a demo to trace a real variance line back to source live. The rest of Rexfin’s reporting tools are indexed at the product tour.