Most small businesses produce a variance report every month and then do nothing with it. The report compares actual results to budget, the numbers that are off get noted in a meeting, and the month closes. The information was produced; the decision-making did not happen. That gap, between variance data and variance decisions, is where the discipline of variance analysis lives.
Variance analysis is not the act of observing differences; it is the act of explaining them. Every variance above a meaningful threshold has a cause, every cause points toward a decision, and every decision has an owner. Without that structure, variance reports accumulate without compounding into learning.
The Three Categories of Variance
Every variance, at its root, is driven by one of three things: volume differed from plan, price differed from plan, or efficiency (how much input was used per unit of output) differed from plan. The corrective action depends entirely on which one.
| Variance Type | What It Means | Typical Corrective |
|---|---|---|
| Volume variance | Sold more or fewer units than planned | Sales pipeline review; revisit demand assumptions |
| Price variance | Sold or paid at a different price than planned | Pricing review; supplier renegotiation; mix analysis |
| Efficiency variance | Used more or less input per unit of output than planned | Process review; productivity investigation; capacity analysis |
A single variance line often blends all three. Decomposing a revenue variance into its volume and price components, rather than treating it as one number, is where most of the insight lives.
The Monthly Variance Walk
A usable variance review follows a sequence. The steps below are the minimum; additional depth (by customer, by product, by department) layers on once the base discipline is installed.
Set the Threshold
Not every variance needs explanation. A typical threshold is 10 percent or $5,000, whichever is greater. Variances below the threshold are noted but not analyzed; variances above are explained in detail. Without a threshold, the review either explains everything (and takes too long) or explains nothing (and produces no value).
Decompose Revenue Variance Into Volume and Price
Revenue variance = (Actual units × Actual price) − (Budget units × Budget price). Separating the two components tells you whether the issue is demand (volume) or pricing (rate). Combined revenue variance alone cannot distinguish the two, and the corrective for each is different.
Decompose Cost Variance Into Price and Efficiency
For direct costs, separate price variance (paid more or less per unit of input) from efficiency variance (used more or less input per unit of output). A labor cost overrun driven by higher wage rates is a different problem than one driven by lower productivity.
Assign a Cause and an Owner to Each Variance
Every variance above the threshold gets a one-sentence cause and a named owner. "Sales missed by 12 percent — Q1 pipeline weaker than projected; owner: VP Sales." Variances without owners are observations, not decisions.
Translate Cause Into Action
Each explained variance triggers one of three responses: (a) accept as noise and move on, (b) investigate further before deciding, or (c) take a specific action (reprice, adjust hiring plan, renegotiate supplier). The variance review is not complete until every material variance has been routed to one of the three.
The Four Most Common Variance Mistakes
Small business variance reviews tend to fail in predictable ways. The errors below are the ones that most consistently turn a variance process into a status ritual.
Variance Review That Produces Noise
- Every line reviewed at the same level of detail
- Variances stated in dollars only, no percentage context
- One-word explanations ('seasonality', 'timing')
- Same recurring variance explained identically month after month
- No threshold — tiny variances get the same treatment as large ones
Variance Review That Produces Decisions
- Threshold-based review; only material variances get deep analysis
- Variances stated in dollars, percentage, and trend (three-month view)
- Causes decomposed into volume, price, and efficiency components
- Recurring variances escalated into reforecast conversations
- Every material variance closed with an owner and an action
A timing variance is a legitimate explanation, the expense or revenue shifted between months but the full-year plan is intact. A permanent variance is a change to the full-year expectation. The two require different responses: timing gets a calendar adjustment; permanent triggers a reforecast conversation. Labeling every variance "timing" is one of the most common ways a variance review becomes meaningless.
What the Analysis Reveals Over Time
A single month of variance analysis produces a snapshot. Three or four months of disciplined analysis produces a pattern, and patterns are where the real management value lives.
A variance that appears once is information. The same variance that appears three months in a row is a plan that is wrong. The discipline of variance analysis is not the monthly review itself; it is the willingness to update the plan when the evidence accumulates. Businesses that review variances monthly but never reforecast are treating their budget as more trustworthy than their actuals, which inverts the relationship.
A Structured Variance Review Agenda
The meeting itself needs structure. Without an agenda, variance review collapses into whoever in the room has the loudest opinion about that month's worst line. The structure below runs in 45 to 60 minutes for a small business.
- 5 min — Headline results: revenue, gross margin, EBITDA vs. budget and vs. last year
- 10 min — Revenue variance: decomposed into volume and price; owner explains
- 10 min — Cost of goods variance: decomposed into price and efficiency
- 10 min — Operating expense variances above threshold, by department
- 5 min — Cash flow and balance sheet variance review
- 10 min — Action items: each material variance gets an owner and a next step
- 5 min — Reforecast trigger check: any variance large or recurring enough to warrant updating the full-year view
Keep a running log of variance explanations and their eventual resolutions. A variance that was labeled "timing" in February should be checked again in March, did the revenue actually arrive, or did "timing" turn into a permanent miss? Without the log, the same explanations recur without ever being validated.
Frequently Asked Questions
Detailed enough to decompose material variances into volume, price, and efficiency; not so detailed that the review takes more than an hour. The right granularity is revenue-by-segment and expense-by-department. Account-line-level detail is rarely necessary at the review; it can be requested for specific variances that warrant it.
If variances are consistently large in the same direction for three or more months, the budget is the problem, not the actuals. That triggers a reforecast. The point of variance analysis is not to defend the budget; it is to manage the business against the best available expectation.
Yes. A revenue variance that comes in 15 percent above plan is as informative as one 15 percent below — it may signal a pricing opportunity, a demand shift, or a one-time event. Treating only unfavorable variances as worthy of explanation is a common bias that hides upside as well as downside.
The business owner, the finance lead (CFO, fractional CFO, or controller), and the functional owners of the largest variance lines. Larger meetings dilute accountability; smaller meetings miss the operational context. Department heads attend for the portion of the review that covers their area.
Sources
- SMAART Company. Small Business Variance Analysis Frameworks, 2026.
- American Institute of Certified Public Accountants. Budget Variance Analysis for Small Business, 2026. aicpa.org
- Harvard Business Review. Making the Monthly Close Useful.
- Corporate Finance Institute. Variance Analysis Fundamentals. corporatefinanceinstitute.com
Discussion
Comments
Leave a comment