ebitda-bridge

EBITDA Bridge

Walks reported EBITDA to adjusted/normalized EBITDA, classifying every add-back. The classifier flags add-backs that look one-time but recur, pro-forma adjustments that lack supporting documentation, and any management add-back that doesn't survive scrutiny.

Usage

/ebitda-bridge <deal_id>

Arguments

• deal_id (required) — The deal to analyze. Requires reported EBITDA + management's adjustment schedule + 2-3 years of historical financials for recurrence testing.

Execution

1. Call build_ebitda_bridge(deal_id) from the MCP server. If the user supplied explicit financial data, pass it as financial_data. Do not pass a primitives argument to analyze_forensic_qoe; that tool runs the full deal-scoped battery.
2. The engine pulls the management EBITDA bridge from the data room (or extracts from CIM/QofE workbook).
3. For each adjustment, the classifier evaluates:
   • Recurrence — does this adjustment appear in 2+ of the last 3 years? (If yes, it's not one-time.)
   • Documentation — is there supporting evidence for the dollar amount?
   • Reasonableness — is the magnitude consistent with industry benchmarks?
4. Returns the bridge with each adjustment tagged and a normalized EBITDA that excludes questionable add-backs.

Output

Add-back	Mgmt $	Classification	Normalized $	Why
Owner comp normalization	1,200K	accepted	1,200K	sole-shareholder, replacement comp documented
"One-time" legal	450K	recurring	0K	also in FY23, FY22, FY21
COVID add-back	800K	questionable	400K	Q2'21 only, no supporting invoices for 50%

Plus a summary line: management EBITDA $X.XM → normalized EBITDA $Y.YM (delta $Z.ZM).

Why this matters

Inflated EBITDA is the #1 driver of overpaying in PE deals. The EBITDA bridge classifier is the single most important forensic test in QoE. Running it pre-LOI tells you whether the headline multiple you're solving for is even directionally right.