The most common failure mode in board reporting is not inaccurate numbers. It is accurate numbers without explanation. A board that sees revenue came in $180K below forecast has one question: why? If the answer is in the board pack, the meeting is productive. If the answer has to be assembled in real time, the meeting is the explanation.

Variance analysis is the practice of answering that question in writing, in advance, every month.

What variance analysis is not

It is not a list of excuses. It is not a document that explains why the forecast was right and the business was wrong. It is not an exercise in finding the number that was closest to actual.

Variance analysis is an accounting of causation. Revenue was $180K below forecast. $120K of that was a single large deal that slipped to next month — it's in the pipeline, contract is in legal. $60K was a churn event in the SMB segment that was not predicted by the leading indicators we track. The second item changes our churn model; the first does not.

That is variance analysis. Specific. Causal. Actionable.

The structure that works

A monthly variance commentary should cover three periods and two comparisons:

Comparisons:

Both comparisons are necessary. A month where actuals miss budget but beat prior period tells a different story than a month where actuals beat budget but miss prior period. The direction of the trend and the direction against plan can diverge, and that divergence is the most important thing in the analysis.

Periods:

The full-year reforecast is the most important output of variance analysis and the one most often omitted. If January revenue misses plan by 10%, the February forecast should reflect what you now know — not what you planned in October. A board that sees January miss and a reforecast that holds the original full-year number is reading a document that doesn't believe itself.

Line-by-line commentary

Not every line needs commentary. Routine variances — expense timing, small volume differences — can be noted in a summary table. The commentary should focus on material variances and inflection points.

Material is a judgment call, but a useful threshold is: any variance that would change a decision. A $20K miss on software spend is not material. A $200K miss on gross margin because COGS classification was wrong is material even if revenue is on plan, because it changes what the board believes about the unit economics.

Inflection points are changes in trend. If customer acquisition cost was flat for six months and increased 25% this month, that is material regardless of the dollar amount. The commentary should explain what drove it and whether it represents a structural change or a timing artifact.

The reforecast discipline

At the end of each month, update the full-year forecast based on what you now know. This is not a revision of the plan — the plan stays fixed as the benchmark. The reforecast is the current best estimate of where the year ends.

Present three columns in every variance table: Budget, Actual, and Reforecast. The gap between Budget and Reforecast is the running score on whether the company is on track to the original commitment. The gap between Actual and Reforecast is the confidence interval on the remaining months.

A reforecast that holds despite misses is an unreliable forecast. A reforecast that moves with every month of actuals — incorporating what you've learned — is a sign of a management team that understands their business.

What boards actually read

Boards have finite attention. The financial section of a board pack gets about fifteen minutes of read time before the meeting. In fifteen minutes, a board member will read the summary table, look for the line items that moved most, and read the commentary for those lines.

Write the commentary for that reader. Put the headline first — "Revenue missed plan by $180K, driven by two items" — before the detail. Use the detail to support the headline, not to replace it. A paragraph that buries the conclusion in the fourth sentence will be read once. A paragraph that leads with the conclusion will be read every time.

The goal of variance analysis is not to explain the past. It is to prove to the board — and to yourself — that you understand the business well enough to be trusted with the next quarter's plan.