What is Variance Analysis?
Variance analysis compares actual performance to baseline, quantifies the gap (cost variance, schedule variance, scope deviations), and — the part that matters — digs into why, so the response targets causes, not symptoms. A variance number without a cause narrative is trivia; with one, it's a steering input.
Thresholds keep it sane: variances inside the agreed band get noted; outside it, they trigger formal analysis and possibly corrective action. Chasing every 1% is how control teams stop controlling.
Formula
CV = EV − AC · SV = EV − PV
Worked example
Month 7: CV −$400K, beyond the ±5% threshold. The analysis splits it: $250K is a steel-price escalation (market — mitigate via early procurement of remaining tonnage), $150K is rework from a subcontractor's survey error (performance — backcharge and add verification holds). Two causes, two different responses — which is exactly what the single number −$400K could never have told anyone.