What is sum-of-the-parts valuation and when do analysts use it?

Why a single multiple fails for diversified companies

Most valuation methods apply a single multiple to a single financial metric. Comparable company analysis compares EV/EBITDA across peers. A DCF projects a single stream of future cash flows. These approaches work well for a focused business: a pure-play retailer, a single-product software company, an asset manager.

They break down for conglomerates and diversified businesses. Consider a company that owns a fast-growing software division (which should trade at 20x EBITDA), a mature industrials unit (which trades at 9x EBITDA), and a real estate portfolio (which is typically valued on yield or net asset value, not EBITDA at all). Applying a single blended EBITDA multiple to that company produces a number that accurately reflects none of its parts.

SOTP solves this by valuing each division on its own terms — using the multiple, method, or cash flow projection most appropriate for that specific business — then aggregating the results.

When analysts use SOTP

SOTP is the primary valuation framework in four situations:

• Conglomerates and diversified industrials: Companies like Berkshire Hathaway, Honeywell, or Siemens contain businesses with completely different characteristics. SOTP is the only sensible way to value them.
• Pre-spin-off or break-up analysis: When a company is considering selling a division or spinning it off, SOTP quantifies whether the parts are worth more separately than together — the core question activists often ask.
• Banks and financial institutions: Banks are often valued by summing their retail banking, investment banking, wealth management, and insurance arms — each with its own multiples and capital requirements.
• Holding companies: A listed holding company that owns stakes in several listed and unlisted businesses is most naturally valued by marking each stake to market and subtracting holding company costs and debt.

Step-by-step SOTP methodology

Building an SOTP model follows a consistent process:

Step 1 — Map the divisions. Identify each distinct business segment. Use the company’s own segment reporting as the starting point. Check whether any segments should be further subdivided (a “services” segment that contains both high-margin consulting and low-margin maintenance, for example, warrants splitting).

Step 2 — Pull segment financials. For each division, extract revenue, EBITDA, EBIT, and capital expenditure from segment disclosures. Where corporate costs are allocated across divisions, decide whether to include them or treat them as a central overhead deduction at the holding company level.

Step 3 — Choose the valuation method for each segment. This is the core judgment call. A software business gets comps-based EV/EBITDA or EV/Revenue. A real estate portfolio gets NAV (assets at market value less debt). A financial services arm might be valued on P/E or price-to-book. A stake in a listed entity is marked to market price.

Step 4 — Apply multiples and derive segment values. Run each segment through its chosen valuation method and derive an enterprise value range for each. Sensitivity-test the key multiples.

Step 5 — Aggregate and adjust. Sum the segment enterprise values. Then deduct: net debt (total group debt minus cash), pension deficits (if material), minority interests in subsidiaries you do not fully own, and any unallocated central costs capitalised at an appropriate multiple. The result is group equity value. Divide by shares outstanding to get implied share price.

The conglomerate discount: why parts sometimes exceed the whole

A well-observed empirical phenomenon: diversified companies often trade at a discount to their SOTP value. Estimates vary by study and period, but a 10–20% conglomerate discount is commonly cited.

Several explanations compete. Investors who want exposure to software can buy a pure-play software company — they do not need a conglomerate that bundles it with industrials and real estate. The complexity of a conglomerate makes analysis harder, which reduces analyst coverage and institutional ownership. Management attention is diluted across businesses. And conglomerate headquarters add costs that pure-plays do not have.

When the discount is large enough, it creates activist pressure — someone will argue, often correctly, that breaking up the company and selling the parts separately would return more value to shareholders. This is the logic behind most large corporate demergers and spin-offs. SOTP quantifies whether that argument has merit.

“SOTP is not just a valuation tool — it is a diagnostic. When the sum exceeds the market value by a wide margin, the market is telling you that it does not trust the company to allocate capital well across its businesses.”

SOTP compared to DCF and trading comps

The main limitations

Segment data quality varies widely. Companies disclose what they choose to disclose. If a company does not break out EBITDA by segment — only revenue — you are forced to estimate margins. Those estimates can be wrong, and errors compound when you are then multiplying by a multiple.

Central cost allocation is a judgment call. Headquarters costs — legal, finance, HR, executive pay — need to be allocated somewhere. Including them within each segment’s EBITDA reduces segment value; treating them as a holdco deduction changes the arithmetic. Different analysts make different choices, which can move implied value by 5–10%.

Inter-segment transactions distort the picture. If the software division sells licences to the industrials division at below-market rates, both divisions’ reported financials are distorted. Clean standalone financials are what you want — but they are often not what you get from group reporting.

What this means for you

SOTP analysis shows up in three places in practice. In equity research, it is the standard framework for publishing on diversified companies and conglomerates. In M&A advisory, it is used to identify break-up value — the ceiling on what a sum-of-parts buyer might pay if they planned to sell each division separately. And in activism, it is the analytical weapon of choice: prove the conglomerate discount, propose a split, push the board to act.

For anyone preparing for interviews in investment banking or equity research, SOTP is a go-to question for cases involving diversified businesses. The examiner wants to see that you understand not just the mechanics — sum the segments, subtract debt — but the economic logic: why different businesses deserve different multiples, what causes the conglomerate discount, and when breaking up is actually in shareholders’ interests.