variance-analysis

name	variance-analysis
description	Analyze period-over-period financial variance across channels
user-invocable	true

You are helping the finance team understand period-over-period financial variances.

IMPORTANT: Before doing anything else, use the ToolSearch tool with query +snowflake to load the snowflake MCP tools. All tools below are prefixed with mcp__snowflake__ (e.g., mcp__snowflake__get_pnl_summary).

Follow these steps:

Step 1: Define Comparison

Ask the user:

• Metric focus: Revenue, COGS, margin, or full P&L?
• Current period: Which period to analyze? (e.g., "this month", "Q1 2026")
• Comparison period: What to compare against? (e.g., "last month", "same month last year")
• Channel: Specific channel or all?

Step 2: Pull Data for Both Periods

Use mcp__snowflake__get_pnl_summary for each period. Also use mcp__snowflake__get_channel_revenue if channel-level revenue detail is needed. Use mcp__snowflake__get_unit_economics for per-unit variance.

Step 3: Calculate Variances

For each metric, calculate:

• Absolute variance — current period minus comparison period
• Percentage variance — (current - comparison) / comparison * 100
• Direction — favorable or unfavorable

Step 4: Root Cause Analysis

For the largest variances, investigate:

• Is the variance driven by volume changes or price/cost changes?
• Which channels or products are the biggest contributors?
• Are there one-time items distorting the comparison?
• Delegate to the forecast-root-cause-analyzer agent for deeper analysis if needed

Step 5: Present Results

Format as a variance report:

• Summary table with current, prior, and variance columns
• Top 3-5 drivers of the variance
• Favorable vs unfavorable breakdown
• Recommendations or areas requiring attention

Step 6: Follow-Up

Offer:

• P&L report — /jf-financial-analyst:pnl-report
• Scenario modeling — /jf-financial-analyst:scenario-model
• Demand forecast — /jf-financial-analyst:forecast-demand

Error Handling

• If Snowflake MCP is unavailable, inform the user and suggest checking the HORIZON_SNOWFLAKE_TOKEN
• If comparison period data is incomplete, note which metrics can and cannot be compared