---
name: variance-analysis-writer
description: Write narrative variance analysis that explains what moved in the numbers, by how much, and why. Use this skill whenever a user has budget-vs-actual or period-over-period figures that need a written explanation, a flux analysis, or management commentary, or says 'explain the variances', 'write the flux analysis', or 'why did the numbers move'. Trigger whenever financial movements need a clear written story for management or a board.
---

# Variance Analysis Writer

## What this does and why it matters
Numbers on their own do not tell leadership what to do. This skill converts a set of variances (actual vs budget, or period vs prior period) into a clear narrative that explains what moved, by how much, and why, so a reader understands the story behind the statements and can act on it. Good variance commentary is what turns a report into a decision.

## Inputs to gather
1. The two figure sets being compared and the periods.
2. The drivers behind the major movements, if the user knows them.
3. The audience and the level of detail wanted.
4. Any materiality threshold for what is worth discussing.

## Method

### 1. Find the significant variances by dollars and percent
Screen movements on both absolute dollars and percentage. A small percentage on a large base can matter more than a large percentage on a tiny line, and vice versa. Discuss what is material, not everything.

### 2. State each variance both ways
For each significant line, give the change in dollars and in percent, so the reader gets scale and proportion together.

### 3. Explain the driver, do not invent it
Attribute the movement to the reason the user provided. Where no reason was given, describe the movement precisely and flag that the driver needs confirmation. Inventing a cause is the cardinal sin of variance analysis.

### 4. Separate one-time from recurring
Distinguish a timing difference or one-off from a genuine trend, since they carry very different implications for the forecast.

### 5. Lead with the story
Open with a short paragraph summarizing the overall picture before the line-by-line detail, so a busy reader gets the point first.

## Output format
ALWAYS use:

# Variance Analysis: [Periods Compared]
## Summary (the overall story)
## Significant variances (line | $ change | % change | driver)
## Trends vs one-time items
## Items needing driver confirmation

## Anti-patterns to avoid
- Inventing the cause of a variance.
- Listing every line instead of what is material.
- Giving only percent or only dollars, losing scale or proportion.
- Treating a one-time item as a trend.

## Guardrails
This explains figures the user supplies; it is not financial advice, a forecast, or assurance on the numbers. Never compute figures that were not provided, and never fabricate a driver. Where a driver is unknown, say so and list it to confirm.

## Example
Marketing spend is up 30 percent (18k) versus budget. The narrative states both figures, attributes it to the driver the user gave (an added campaign), flags it as a deliberate one-time push rather than a trend, and notes the expected payback to confirm.