Key takeaways
- Variance analysis in the month-end close documents why every material account moved before the financial statements circulate. It records the explanation beside the figure, so the close ships already answered.
- A dual threshold performs the filtering. Flag an account only when it exceeds a dollar floor AND a percentage, so a $34,000 movement on a $180,000 balance surfaces while a $6,200 movement on a $190,000 balance stays quiet.
- For a modeled $40M SaaS company, roughly 12 of 340 accounts breach the threshold in a period. Documenting those 12 explanations consumes about 1.2 hours of close-window labor.
- Reconstructing identical explanations after a board meeting, once the ledger advances, requires about 3.6 hours at $40.23 hourly. The analysis is unchanged; the timing establishes the cost.
- The discipline reflects a regulatory expectation. SEC Regulation S-K Item 303 obligates public filers to describe the underlying reasons for material period-to-period changes in quantitative and qualitative terms.
Variance analysis in the month-end close is the discipline of explaining why every material account moved before the financial statements circulate. It measures each balance against a baseline, isolates the accounts that exceed a defined threshold, and documents the explanation beside the figure. A modeled $40M single-entity SaaS company maintains roughly 340 active general-ledger accounts. Only about 12 accounts fluctuate enough in a period to justify a written explanation. Documenting those 12 during the close, rather than inside a board meeting, distinguishes a defensible close from a provisional one.
What is variance analysis in the month-end close?
Variance analysis converts a set of accurate numbers into a set of explained numbers. Auditors describe the identical technique as flux analysis, an abbreviation of fluctuation analysis. The close produces a trial balance. Variance analysis interprets that balance against a prior period, an approved budget, or a current forecast, then documents the driver behind every movement significant enough to influence a decision.
The deliverable is a concise narrative attached to the financials. Each flagged account receives one sentence: the amount, the magnitude, and the underlying explanation. A marketing balance that increased because a conference invoice posted within the period reads differently from one that increased because a vendor renegotiated its rate. Both entries are accurate under GAAP. Only the documented driver reveals which movement signals a durable trend and which represents timing.
Which baselines do you compare against, and what does each catch?
Four baselines answer four separate questions, and a rigorous close applies several. Comparing exclusively against the prior month overlooks seasonality. Comparing exclusively against budget overlooks a stale plan. The table maps each baseline to the question it resolves and the movement it exists to surface.
| Baseline | Question it answers | What it catches | Typical trigger |
|---|---|---|---|
| Budget vs actual | Are we tracking the plan the board approved? | Overspending and revenue shortfalls against the committed figure | Above the dollar floor and above a percentage of the budgeted line |
| Month-over-month | What changed since the previous close? | Incremental or discontinued spending, one-time items, posting errors | Above the dollar floor and above a percentage sequentially |
| Year-over-year | How does this period compare to the equivalent month last year? | Seasonality, growth-rate shifts, structural change in the business | Above a percentage annually on seasonal accounts |
| Actual vs forecast | Are we tracking the most recent reforecast? | Drift since the prior re-forecast and forward-looking exposure | Above a percentage of the current forecast |
Most monthly closes lead with month-over-month and budget-to-actual comparisons. Seasonal accounts additionally warrant a year-over-year view, because a single month distorts them. Cash-intensive accounts interpret cleanly against budget. Revenue and headcount-driven expenses interpret cleanly against the prior year. An auditor subsequently applies identical logic under AICPA AU-C 520, which directs the auditor to investigate fluctuations that differ from an expected value by a significant amount.
How do you set a threshold that flags the right accounts?
Combine two conditions with AND: a dollar floor and a percentage. An account earns an explanation only when it satisfies both. The dollar floor eliminates immaterial accounts that swing dramatically on a percentage basis. The percentage eliminates substantial accounts that drift a rounding error in absolute terms. Together they isolate the movements that genuinely alter an executive’s judgment.
Consider the arithmetic on a single account. A marketing balance advances from $180,000 to $214,000. The absolute change equals $214,000 − $180,000 = $34,000, and the relative change equals $34,000 ÷ $180,000 = 18.9%. Against a policy of $10,000 and 5%, that account satisfies both conditions and demands a documented driver.
A contrasting account illustrates the filter. A balance moving $6,200 on a $190,000 base changes 3.3%, exceeds the dollar floor, but fails the percentage. It stays off the flux schedule. For the modeled $40M company, this dual condition compresses 340 active accounts into roughly 12 required explanations each period.
The threshold parallels how the audit standard frames materiality. Under PCAOB AS 2305, an auditor develops an expectation for an amount, defines the difference from that expectation acceptable without further investigation, and evaluates the significant unexpected differences that remain. A close that establishes its own threshold answers that examination before the auditor conducts it.
Why explain variances before the close locks instead of in the board meeting?
Because the explanation costs considerably less at the desk than in the room, and the underlying figure is identical either way. Documenting a driver while the entry remains fresh requires about six minutes. Twelve flagged accounts at six minutes each totals 72 minutes, or 1.2 hours of close-window labor, filed the moment the books lock.
Defer the analysis, and the identical 12 questions surface live, after the ledger advances into the subsequent period and the context deteriorates. Reconstructing each explanation then approaches 18 minutes: re-pull the support, reconstruct the timing, reconfirm the driver. Twelve accounts at 18 minutes totals 216 minutes, or 3.6 hours.
Price both paths against a wage. At the Bureau of Labor Statistics May 2025 median accountant wage of $40.23 hourly, the prepared path costs 1.2 × $40.23 = $48.28, and the reactive path costs 3.6 × $40.23 = $144.83. The differential is $96.55 per close, recurring across all 12 closes annually. The dollar figure understates the exposure. An executive who hears a number without a driver discounts every adjacent number, and the decision the meeting convened to reach slides to a follow-up.
How do you systematize variance analysis in a fast close?
Embed the flux review into the close calendar instead of appending it after the lock. On a five-day close calendar, variance analysis occupies day four: statements draft, and every material account receives its driver before the period locks on day five. Templating the baselines accelerates the review, because the comparison against budget, prior month, and prior year populates the instant the trial balance finalizes.
The habit endures only when the ledger remains current throughout the month. The Continuous Close Method™ keeps accounts substantially reconciled as transactions post, so the flux review interprets genuine movement instead of unreconciled noise. Puzzle surfaces the movement the moment a transaction posts; a controller supplies the explanation and the judgment a board relies on. Aaron Ressel reviews the flux narrative before Debit & Co. releases a board packet, the control layer this discipline protects.
Anchor the policy to the standard governing the reader. Public filers describe material changes under SEC Regulation S-K Item 303; a private company borrows equivalent rigor for its board and its lenders. For the day-by-day sequence the flux review inhabits, the month-end close process establishes the frame, and the close benchmarks by company size establish the pace it must sustain.
Frequently asked questions
What is variance analysis in the month-end close?
Variance analysis in the month-end close explains why every material account moved before the financial statements are released. It compares each account against a baseline, such as the prior month, the budget, or a forecast, isolates the accounts that breach a defined threshold, and documents a concise written driver for each one. The resulting financials circulate already explained, so an executive sees not only the figure but the reason it changed since the previous period.
What is the difference between variance analysis and flux analysis?
They describe an identical technique from two vantage points. Management calls it variance analysis; auditors call it flux analysis, an abbreviation of fluctuation analysis. Both compare a recorded amount against an expected amount and investigate the differences exceeding a threshold. AICPA AU-C 520 frames the audit version, directing an auditor to investigate fluctuations differing from expected values by a significant amount, the equivalent review a controller conducts internally during the close.
How do you set a variance analysis threshold?
Combine a dollar floor and a percentage with AND, so an account requires an explanation only when it satisfies both conditions. The dollar floor screens out immaterial accounts that swing sharply on a percentage basis, and the percentage screens out substantial accounts drifting only a rounding error in absolute terms. Under a policy of $10,000 and 5%, a marketing account moving $34,000 on a $180,000 base changes 18.9% and flags, while a $6,200 movement on a $190,000 base changes 3.3% and does not.
Which baseline should you compare against in a monthly close?
Apply several, because each baseline surfaces a different movement. Month-over-month captures incremental spending and posting errors; budget-to-actual captures performance against the approved plan; year-over-year captures seasonality and growth-rate shifts on accounts a single month distorts; actual-versus-forecast captures drift since the prior reforecast. A common approach leads with month-over-month and budget-to-actual, then adds year-over-year on the seasonal accounts where a one-month view misleads.
Why explain variances during the close instead of in the board meeting?
Because an identical explanation costs considerably less documented at the desk than reconstructed under questioning. Documenting roughly 12 flagged drivers requires about 1.2 hours during the close, filed when the books lock. Rebuilding them after a board meeting, once the ledger advances and the context deteriorates, approaches 3.6 hours, a difference of about $96.55 per close at a $40.23 hourly wage. The larger cost is credibility: a number presented without a driver discounts the numbers surrounding it.
What tools support variance analysis in a fast close?
The reporting layer inside QuickBooks Online or NetSuite builds the budget-to-actual and prior-period comparisons, a continuous-accounting platform such as Puzzle keeps the ledger reconciled so the comparison interprets genuine movement, and a documented flux template captures the driver beside each flagged account. Puzzle surfaces the movement as transactions post; a controller supplies the explanation and the judgment a board relies on. Tools compute the variance, yet the written driver still depends on human review under AU-C 520.
Auditing guidance cited from AICPA AU-C 520, Analytical Procedures, and PCAOB AS 2305, Substantive Analytical Procedures. Disclosure requirement from SEC Regulation S-K Item 303, Management’s Discussion and Analysis. Wage data from the U.S. Bureau of Labor Statistics, May 2025. Figures verified as of August 2026; the $40M company, thresholds, and account movements are anonymized illustration, not a benchmark.
