When Alphabet (GOOGL) trades at one P/E and Amazon (AMZN) at a wildly different one, a single multiple is the wrong lens. Wall Street's name for the right lens is Sum-of-the-Parts — and it routinely surfaces 15–25% of "hidden" value in diversified companies.
What is Sum-of-the-Parts valuation?
Sum-of-the-Parts (SOTP), sometimes called "break-up analysis," values each operating segment of a company on the multiple a standalone competitor in that segment would earn — then sums the pieces to get an enterprise value. The single sharpest insight of SOTP is that mixed-business companies almost never deserve a single multiple, because their cash flows have very different durabilities and growth profiles.
A pure-play software company growing 20% might trade at 12x revenue. A pure-play insurance company might trade at 1.2x book value. If both businesses sit inside the same conglomerate, applying any "blended" multiple to the combined income statement is mathematically arbitrary — and almost always wrong.
The textbook formula is short:
SOTP Enterprise Value = Σ (Segment Value)
− Corporate Overhead
− Net Debt
± Conglomerate Discount/Premium
When should you use SOTP instead of a single multiple?
Three situations. Each one has a clear "tell" you can spot from a 10-K.
Situation one — diversified revenue mix. If the largest segment is less than 60% of revenue, single-multiple valuation is producing false signals. Examples: Alphabet (GOOGL) (Search, YouTube, Cloud, Waymo), Amazon (AMZN) (Retail, AWS, Advertising, Subscriptions), Disney (DIS) (Parks, Streaming, Studios, ESPN).
Situation two — different growth profiles inside one ticker. When you can show two segments are growing at very different rates — say JPMorgan (JPM) Consumer Banking at low single digits versus Investment Banking at high single digits — they require different discount rates and different multiples.
Situation three — there's a credible spin-off or break-up rumor. Once the market starts pricing in a separation, the only honest way to value the stock is to model the post-spin pieces directly. Recent examples: Honeywell (HON), Disney (DIS), General Electric (GE) (after its three-way split).
For a fundamental refresher on how to read segment data from financial statements, our DuPont analysis primer gives the relevant frame.
How do you actually build a SOTP model?
Four steps. The art is in choosing the comparables; the rest is bookkeeping.
Step one: identify each segment that operates as a distinct business. Read the segment-reporting note in the latest 10-K. The "operating segments" disclosure is your map — it tells you exactly what management considers a separately managed business.
Step two: pick a pure-play comparable for each segment. This is where most amateur SOTP analyses fall apart. The comparable has to be a true peer in growth, capital intensity, and margin profile. AWS comparables are Microsoft (MSFT) Azure (inside MSFT) and Oracle (ORCL) cloud — not Cisco (CSCO) data-networking.
Step three: apply the comparable's multiple to the segment's metric. EV/EBITDA is the safest default for mature cash-generating segments. EV/sales works for early-stage segments where margins haven't normalized.
Step four: subtract corporate overhead and net debt, and add cash. Then decide whether to apply a conglomerate discount or premium. The empirical default is a 10–15% discount unless management has demonstrated unusual capital-allocation skill.
| Step |
What it does |
Common mistake |
| 1. Identify segments |
Maps the company into business units |
Using marketing categories instead of GAAP segments |
| 2. Pick comparables |
Anchors valuation to real peers |
Choosing peers with different growth profiles |
| 3. Apply multiples |
Generates segment values |
Using EV/EBITDA on a money-losing segment |
| 4. Adjust for corp items |
Bridges to equity value |
Forgetting corporate overhead or unallocated SBC |
Worked example — what does SOTP say about Alphabet?
A simplified illustration. The point is not the precise number; it's how dramatically the picture changes versus a single multiple.
Imagine GOOGL reports four segments: Search, YouTube Ads, Google Cloud, and "Other Bets." Applying a single forward P/E of about ~22 to the total company gives roughly one number — call it baseline.
Now value each segment separately:
- Search (mature, high-margin, ~roughly 75% of operating income) → 18–20x peer multiple
- YouTube Ads (high-growth video advertising, comp set is META Reels + NFLX ad tier) → 22–25x peer multiple
- Google Cloud (growing roughly 30% YoY, comp set is Azure, AWS) → 8–10x forward revenue
- Other Bets / Waymo (option value, no current earnings) → strategic valuation
When you add these up — net of corporate costs, net cash, and a modest conglomerate discount — you typically get an SOTP that lands meaningfully above the blended-multiple baseline, with most of the gap coming from the Cloud and YouTube pieces. That gap is the "hidden value" SOTP is designed to surface, and it is exactly the dynamic that activist investors target when they push for spin-offs.
For real-world examples of how this kind of analysis powers contrarian investing, see our Carl Icahn profile or our broader hub on super investors.
What is the "conglomerate discount" — and why does it exist?
It's the gap between a diversified company's market cap and its summed standalone value. Empirically, that gap sits at roughly 10–25%.
Three honest reasons it exists:
Capital-allocation drag. When cash from a winning segment subsidizes a losing one, capital gets stuck. Markets discount this because the alternative — paying it out or buying back stock — is provably more accretive.
Reporting opacity. Segment financials are often less detailed than what pure-play competitors disclose. Analysts can't model what they can't see, so they default to conservative multiples.
Management bandwidth. Running four businesses competently is harder than running one. The discount partly captures the operational tax of complexity itself.
Critics argue the discount is overstated — that diversified businesses get cheaper financing and benefit from internal capital markets in ways pure-plays don't. The honest reading: the conglomerate discount is real, but it shrinks materially when management has a demonstrably elite capital-allocation engine. That's why Berkshire-style compounders or Markel-style operators sometimes trade at a premium to SOTP rather than a discount.
Where does SOTP fail?
In three predictable places. None of them invalidate the framework — they just narrow where it works.
When segments share resources you can't separate. Some companies have one engineering team, one sales force, and one data platform serving every segment. Splitting them on paper produces phantom synergies you can't actually capture. A meaningful share of META's "Reality Labs" cost overlaps with its ads business in non-obvious ways — break it up and you may not capture as much as the model suggests.
When the conglomerate's segments are intentionally complementary. Costco (COST) and its membership-fee dynamic is a classic case — the merchandise margin and the membership fee are inseparable parts of one business model. SOTP cannot value Costco well; the business is greater than the sum of its parts by design.
When tax leakage is large. A real break-up creates taxable events that the modeled SOTP ignores. For very large segments with significant accumulated gains, that tax bill can eat 10–20% of the theoretical unlock.
For investors who want to layer SOTP on top of a quality framework, our economic moats explainer gives the natural complement — SOTP identifies the value, moats decide whether it's durable.
What's the best way to use SOTP as a retail investor?
Three rules that turn a complex analyst tool into a usable filter.
Rule one: only do it for companies with clear segment reporting. If a company lumps everything into one or two segments, you're guessing — not modeling.
Rule two: focus on directional gaps, not precise targets. SOTP answers "is there meaningful hidden value here?" not "is this stock worth exactly $185." A 20%+ gap deserves your attention; a 5% gap is noise.
Rule three: pair the analysis with a catalyst hunt. A SOTP gap is only an opportunity if something might close it — activist pressure, a planned spin, a management change, a credit-rating event. Without a catalyst, gaps can persist for many years.
Ready to analyze these stocks yourself? Search any ticker on MainRatios to see valuations from 6 legendary investors - free.
