Variance Analysis

Variance Analysis

The canonical skill for comparing actual operating performance against any comparison basis. Any command that needs variance reporting MUST invoke this skill.

Trigger Conditions

Activated when any command needs:

• Actuals vs. budget comparison
• Actuals vs. prior year comparison
• Actuals vs. underwriting comparison
• NOI bridge / variance decomposition
• Material variance identification and explanation

Business Context

Variance analysis is the backbone of asset management reporting. Every monthly asset report, quarterly review, and reforecast starts with understanding where actuals diverge from expectations and why.

Source Hierarchy

1. Yardi/MRI GL data via MCP (when available) -- system of record for actuals
2. User-provided P&L or financial statements
3. Budget from property management system or user
4. Underwriting assumptions from manifest (underwriting/assumptions.json)

Workflow

Step 1: Align Data

• Map chart of accounts between actuals and comparison basis
• Verify time periods match (same months, same fiscal year)
• Flag any unmapped line items

Step 2: Calculate Variances

Use scripts/variance.py:calculate_variance() for line-item analysis. Use scripts/variance.py:noi_variance_bridge() for NOI-level summary.

Step 3: Apply Materiality Thresholds

Default thresholds (can be overridden by user or .claude/*.local.md):

• Dollar threshold: $10,000 or 5% of line item, whichever is greater
• NOI-level: 2% of budgeted NOI
• Revenue items: positive variance = favorable
• Expense items: negative variance = favorable

Step 4: Decompose Drivers

For each material variance:

• Rate vs. volume: Did the per-unit cost change, or did usage change?
• Timing: Is this a permanent variance or timing difference?
• Controllable vs. uncontrollable: Management actions vs. market/regulatory
• One-time vs. recurring: Flag non-recurring items

Step 5: Output

• Variance summary table (all line items with $ and % variance)
• Material variance narrative (top 5 drivers with explanations)
• NOI bridge (budget NOI → actual NOI, showing each driver)
• YTD trend if multiple periods provided

QA Checks

• Actuals and comparison periods are aligned
• Total revenue and total expense agree with source documents
• NOI variance = revenue variance - expense variance (arithmetic check)
• Material variances have explanations, not just numbers

Edge Cases

• Partial year: Annualize carefully, or compare only matching months
• Property acquired mid-year: Pro-rate budget for ownership period
• Major capital event: Separate capital from operating variances

Standards References

• IREM expense categorization for consistent line-item mapping
• BOMA expense classification for office properties