How to write variance commentary that explains why
Strong variance commentary follows a quantified pattern: result versus named benchmark, material driver and its amount, timing or persistence, business implication, corrective or reinforcing action, owner, and expected next-period effect. It should not merely say a line is above or below budget, and it should distinguish verified facts from management explanations and unresolved residuals.
From the field · Regional business manager
Explain a regional sales miss without inventing a cause
Revenue is below plan. The data quantifies lower unit volume and an adverse mix shift, while supply disruption and competitor discounting are plausible but not yet measured. Good commentary preserves that distinction.
Columns used in the example
region · period · budget · actual · units · price · product_mix · evidence_status
Follow it with your own file
- State the total variance in amount and percentage with its comparison basis.
- Quantify the largest reconciled drivers in descending order.
- Label supported interpretations, untested hypotheses, and data gaps distinctly.
- Finish with an owner, action, and evidence needed for the next update.
Check: All quantified drivers plus the remainder must reconcile to the total variance. Remove causal words for explanations that the current evidence does not establish.
What needs resolving first
Comments such as travel was 12% unfavorable or revenue exceeded plan provide no mechanism, persistence, consequence, or response. They consume report space while leaving leaders with the same question: why?
Use a repeatable sentence architecture
Start with result and benchmark. Add because with quantified drivers, resulting in with the implication, and management will with a specific action. This structure can fit in two or three sentences.
Use tables for detailed values and prose for mechanism, consequence, and decision.
Do not manufacture certainty
An owner explanation is not automatically verified. Label preliminary attribution, list evidence, and retain an unexplained residual when drivers do not fully reconcile.
A timing explanation should specify when reversal is expected and appear in the forecast.
Build the result step by step
- Anchor the result: State actual, benchmark, absolute and meaningful relative gap, period, and importance.
- Quantify driver evidence: Name the few mechanisms and amounts that explain the movement, leaving residual visible.
- Classify persistence: Label timing, recurring, one-off, structural, forecast change, or unresolved.
- Specify response: Name owner, action, due date, decision dependency, and expected financial or operational effect.
What changes the answer
- Repeat neither the table nor a percentage without context.
- Quantify drivers so they reconcile to the gap.
- State whether movement is timing, one-off, or structural.
- End with owner, action, and expected effect.
Weak and useful variance commentary
| Weak pattern | Useful replacement | Adds |
|---|---|---|
| Costs were 12% high | Costs were $84k above plan, led by... | Magnitude and driver |
| Timing difference | $40k shifts to September due to... | Expected reversal |
| Monitor spend | Procurement owner will renegotiate by... | Owner and action |
Warning signs in the file
- Commentary restates color and percentage.
- Driver values do not sum to the gap.
- Timing items return month after month.
- Actions use vague verbs such as monitor or optimize.
Doing these steps in Prepyr
Upload a CSV working copy in Prepyr. Describe the result you need: Write variance commentary that quantifies drivers and separates evidence from hypothesis. Check the columns selected, preview any proposed changes, and compare the output with the source before applying or exporting it.
Prepyr steps used here: Variance analysis · Report narratives · Waterfall charts · Recurring reports
Make the output easy to check
- Benchmark named
- Amounts reconcile
- Persistence classified
- Implication stated
- Action specific
- Residual honest
Commentary that enables a reader to understand the mechanism, decide whether it changes the outlook, and track the response.
Questions readers usually ask next
How long should one variance comment be?
Usually two to four concise sentences; supporting detail belongs in a table or appendix.
Should favorable variances be explained?
Yes when material, surprising, nonrecurring, or relevant to forecast and future decisions.
What if the driver is unknown?
State the verified gap, label attribution unresolved, assign investigation, and avoid invented explanations.