Sum-of-the-Parts Valuation Explained: How to Value a Conglomerate Segment by Segment — the Break-Up Arithmetic, the Discount the Market Applies, and Why the Gap Is Not Free Money
10 min read
- Sum-of-the-parts (SOTP) values a business one segment at a time, each on the multiple its own peers trade at, then sums the enterprise values and bridges to equity. Its job is to expose value that a single blended multiple hides — a group that is 60% low-growth industrials and 40% high-growth software will never be priced correctly on one number
- The build is disciplined, not clever: segment EBITDA (or EBIT) × segment-specific peer multiple = segment EV; sum the segments; then run the EV-to-equity bridge — subtract net debt, pension deficits and minority interests, add associates and any non-operating assets — to reach equity value and a per-share figure
- SOTP almost always prints above the market cap. That gap is the conglomerate discount — historically on the order of 10–20% on average (approximate; deep enough to exceed 30–40% on the worst-run diversified groups). The error students make is calling it mispricing. Most of it is structural
- The contrarian read: the discount is largely deserved. A break-up does not hand you the gap for free — it crystallises stranded corporate costs, forfeits debt capacity the diversified group enjoyed, triggers tax on separation, and severs whatever real synergies existed. The activist's SOTP is an argument dressed as a valuation, and the only question that matters is how much of the gap survives the cost of unlocking it
Why One Multiple Cannot Price a Business That Does Several Unrelated Things
Most valuation methods hand you a single number for the whole company — a DCF discounts one consolidated cash flow, a trading-comps analysis applies one peer multiple to one EBITDA line. That works when the business is one thing. It breaks the moment a company runs segments with genuinely different economics: a cash-generative industrial division that the market pays 7x EBITDA for, sitting alongside a software unit its pure-play peers trade at 20x. Blend them into one multiple and you value neither correctly — the market either under-pays for the software or over-pays for the industrials, and usually it under-pays for the whole.
Sum-of-the-parts exists to solve exactly that mismatch. Rather than forcing one multiple onto a mixed business, it values each segment against the peers that segment would trade with if it were public on its own, then adds the pieces back up. The method is standard IB and PE work for conglomerates, holding companies, and any group where a spin-off, break-up or divestiture is on the table — and it is a favourite interviewer probe precisely because it tests whether you understand that a multiple is a statement about a specific set of economics, not a number you can average.
The Build: Segment EBITDA, a Peer Multiple That Fits Each Segment, Summed to Enterprise Value
The mechanics are deliberately simple, because the judgement is in the inputs, not the arithmetic. For each operating segment you take a normalised profit measure — usually EBITDA, sometimes EBIT or revenue where a segment is loss-making or pre-profit — and apply the multiple derived from that segment's own comparable companies. Multiply, and you have the segment's standalone enterprise value. Do it for every segment, sum them, and you have the group's aggregate operating enterprise value before any of the corporate-level adjustments that decide whether the exercise is honest.
From Aggregate EV to Equity Value: The Bridge That Decides the Per-Share Number
Summing segment EVs gives an enterprise value, and enterprise value is not what a shareholder owns. The step that separates a competent SOTP from a naïve one is the EV-to-equity bridge run at the group level — the same bridge tested to death in IB interviews, applied here to the sum rather than to one company.
A Worked Sketch: Where the Blended Multiple Buries the Software Unit
Numbers make the mismatch concrete. Take a diversified group with three segments and a corporate centre, illustrative throughout.
| Segment | EBITDA | Peer multiple | Segment EV |
|---|---|---|---|
| Industrials | £500m | 7.0x | £3,500m |
| Consumer brands | £250m | 11.0x | £2,750m |
| Software | £120m | 18.0x | £2,160m |
| Unallocated corporate costs | (£70m) | ~10x | (£700m) |
| Aggregate operating EV | £800m | — | £7,710m |
Now the bridge. Assume £300m of associates and surplus assets, £2,000m of net debt, a £250m pension deficit, and £260m of minority interests, on 1,000m shares.
| Bridge item | Figure |
|---|---|
| Aggregate operating EV | £7,710m |
| Plus associates and surplus assets | £300m |
| Less net debt | (£2,000m) |
| Less pension deficit | (£250m) |
| Less minority interests | (£260m) |
| SOTP equity value | £5,500m |
| Per share (1,000m shares) | £5.50 |
Blend the same £800m of group EBITDA on one multiple and the market rarely gets near £7.7bn of EV — a diversified industrial-weighted group is more likely priced at 8–9x blended, roughly £6.4–7.2bn, dragging the implied share price below the £5.50 SOTP. That difference is not an error in one of the two numbers. It is the discount, and the rest of the piece is about whether it is a mispricing to arbitrage or a cost you would pay to unlock.
The Conglomerate Discount: Real, Persistent, and Roughly 10–20% on Average
Diversified groups trade below the sum of their parts as a rule, not an exception. The academic and practitioner work on the "diversification discount" has for decades put the average gap on the order of 10–20% (approximate; the seminal studies cluster nearer 13–15%, and the number is sensitive to method and period), with the worst-governed or most-sprawling groups discounted far more steeply. The discount is persistent enough that closing it is a recognised investment thesis in its own right.
The reasons the market applies it are structural, not sentimental. Capital allocation inside a conglomerate is opaque — investors cannot see whether cash thrown off by the industrials is being reinvested well or used to subsidise a weak division, so they discount the whole for the risk that it is the latter. A single share cannot be tuned to one segment, so specialist investors who would pay full multiple for the pure-play software or pure-play industrials avoid the mixed vehicle, thinning demand. Central overhead the segments do not carry on their own accounts sits at the centre as a drag. And complexity itself carries a governance and analysis premium — harder to model, harder to hold management to account, easier to hide underperformance. Each of those is a reason the discount is at least partly earned.
Why the Gap Is Not Free Money: Stranded Costs, Lost Leverage, Separation Tax, and Dis-Synergies
Here is where the naïve reading fails. A student who runs the build above concludes the group is worth £5.50 against a £4.50 share price and calls it a 22% upside waiting to be unlocked by a break-up. The market is not being stupid. Crystallising that gap costs real money, and four items eat into it before a single pound reaches shareholders.
Net those four against the headline gap and a 20% raw discount can shrink to a mid-single-digit real prize — sometimes to nothing, occasionally to negative, where the conglomerate genuinely was more valuable held together. This is the judgement SOTP is really testing: not whether you can add segment EVs, but whether you can tell a discount that reflects a fixable failure of capital allocation from one that is the fair price of costs a break-up would simply relocate.
The Activist's Weapon: SOTP as an Argument, Not a Valuation
SOTP is the standard opening move in an activist or break-up campaign, because it produces a headline number above the share price and a clean narrative — "the company is worth £X, management's strategy is costing you the gap, spin the divisions and capture it." It is a persuasive frame, and sometimes a correct one: where the discount is driven by lazy capital allocation, cross-subsidy of a failing unit, or a management team empire-building across unrelated sectors, a break-up can close most of the gap and the activist is right.
But the frame is an argument, and it is built from the analyst's own choices — which peer multiples, which segment margins, whether the corporate centre is charged at the multiple it deserves or waved through. Move the software multiple from 18x to 14x and the "unlock" halves. The tell of a strong analyst, on the sell-side defending management or the buy-side pressure-testing the activist, is refusing to take the SOTP as an objective valuation. It is a scenario — the value of the parts if they traded at chosen multiples, if stranded costs were contained, if the tax were sheltered, if the dis-synergies were smaller than feared. Every one of those is contestable, and the deal turns on which side models them honestly.
How SOTP Sits Next to the DCF and the Comps
SOTP is not a rival to the other valuation methods — it is a wrapper that applies them segment by segment. In practice each part can be valued on its own trading comps, its own precedent transactions, or its own discounted cash flow, and a rigorous SOTP often triangulates: comps for the segment where good pure-play peers exist, a DCF for the one where they do not. The consolidation into a group figure, and the bridge to equity, is what makes it SOTP rather than three unrelated valuations. That is also why it belongs in the same toolkit as the standard methods rather than apart from them — it is the technique you reach for the moment a single blended multiple stops telling the truth.
Careers: The Analyst Who Owns the SOTP Owns the Break-Up Story
On a live sell-side defence or an activist situation, the SOTP model is often the most scrutinised page in the deck, and the analyst who builds it is defending every input under partner and client questioning. The work is not the arithmetic — it is sourcing defensible peer sets for each segment, deciding how to charge the corporate centre, estimating the stranded costs and separation tax that the raw build omits, and running the sensitivity that shows how much of the gap survives realistic assumptions. A junior who can hand a partner "the raw SOTP is £5.50, but net of stranded cost and separation friction the realisable number is closer to £4.90 — the market is not as wrong as the activist claims" is doing the analysis the mandate turns on.
The instinct compounds across the job. It is the same discipline that separates a real quality-of-earnings review from a checklist and a genuine synergy case from a press release — taking a headline number apart, finding the costs the headline omitted, and pricing what is actually left. Do it well on an SOTP and you are demonstrating exactly the judgement that a diligent buy-side or advisory desk is hiring for.
Take Your Preparation Further
Sum-of-the-parts is the method that stitches the other valuation techniques together, so read it alongside the pieces that build them. Start with trading comps and precedent transactions for how each segment's peer multiple is derived and defended, and the EV-to-equity bridge for the group-level adjustments that turn summed segment EVs into a per-share number. For the intrinsic method you will fold into segments that lack clean peers, see the DCF and terminal value, and for the wider frame — why a multiple is a statement about specific economics rather than a number to average — read how to think about valuation.
For every valuation method on one page — when to use each, the multiples that matter, and the adjustments that catch candidates out — download our free Valuation Methods Cheat Sheet, and to build a segment-level intrinsic valuation yourself, use the Professional DCF Model Template.
Ready for personalised feedback? Book a 1-on-1 mentoring session with an experienced IB/PE professional.