Sector-Specific Valuation

Every methodology in this guide rests on assumptions about how a business works: that EBITDA measures operating performance, that debt is financing rather than operations, that the asset base is maintained through reinvestment, and that current earnings are a fair proxy for sustainable earnings. In several major sectors, at least one of those assumptions breaks. Banks fund their core business with debt. Mines and oil fields consume their own asset base through production. High-growth software companies run negative EBITDA by design. Regulated utilities have their earnings set by a government body rather than by the market.

Wherever an assumption breaks, the sector has developed its own metrics and frameworks that capture what actually drives value there. Knowing which framework applies, and more importantly why the standard one fails, is a core interview test: the fastest way to signal weak fundamentals is to apply EV/EBITDA to a bank. This survey works through the major sector frameworks a generalist is expected to recognize, the economics that justify each one, and the decision logic for choosing among them.

Technology and SaaS: Revenue Multiples and the Quality Behind Them

High-growth software breaks the standard toolkit in two places at once. Many SaaS companies generate negative EBITDA even while growing revenue 30 to 50 percent a year, because they deliberately pour cash into sales, marketing, and R&D; an EV/EBITDA multiple on a negative denominator is meaningless. And a DCF is technically buildable but ends up hostage to its terminal assumptions, since near-term cash flows are minimal or negative. The sector's answer is to value the revenue itself: EV/Revenue, or the more precise EV/ARR.

Annual recurring revenue (ARR) is monthly recurring revenue multiplied by 12. It deliberately excludes one-time revenue such as implementation fees and consulting, isolating the contractual subscription base that actually recurs. Revenue multiples dominate for three connected reasons:

• •EBITDA is often negative, so earnings-based multiples are undefined exactly where the sector's growth is fastest.
• •SaaS revenue is unusually high quality: recurring, contractual, predictable, and high margin, so a dollar of it is worth more than a dollar of one-time product revenue.
• •The market expects operating leverage to convert that revenue into substantial profit at maturity; the revenue multiple implicitly prices the future margin structure.

The Rule of 40

The standard framework for judging whether a SaaS company's growth-versus-profitability trade-off is acceptable:

A company growing 30 percent with a 15 percent EBITDA margin scores 45 and is healthy; one growing 10 percent with a 20 percent margin scores 30 and is underperforming. The logic is that growth and profitability trade off against each other: fast growth justifies thin or negative margins because the company is investing in its future, while slow growth demands strong margins because the business should be harvesting cash. Companies clearing 40 consistently trade at premium EV/Revenue multiples, and the multiple ladder rises steeply with the score.

The Metrics Behind the Multiple

Two SaaS companies with identical revenue can deserve very different multiples, and the difference lives in a handful of quality metrics.

Net revenue retention (NRR) measures the recurring revenue kept from the existing customer cohort over 12 months, including expansion (upgrades, added seats, cross-sells) and net of churn and downgrades. NRR of 120 percent means last year's customers alone generate 20 percent more revenue this year, before any new logo is signed. Above roughly 130 percent is elite and signals a land-and-expand model that compounds without proportional sales cost; below 100 percent means the existing base is shrinking faster than it expands, a red flag regardless of new customer wins.

Two further quality markers frame the multiple:

• •Gross margin: typically 70 to 85 percent. Pure-software companies at the top of that range earn premium multiples because each incremental revenue dollar falls almost entirely to contribution margin, while heavy professional-services or hardware content drags the multiple down.
• •Free cash flow margin: the market's focus for later-stage companies. Mature SaaS businesses demonstrate 25 to 35 percent FCF margins, and companies with improving 5 to 15 percent margins get credit for the trajectory, while those whose growth slows without cash conversion are penalized hardest.

One trap deserves attention here. Many SaaS companies show positive FCF only because stock-based compensation, a non-cash expense, is excluded from operating cash flow. If the company simultaneously spends cash on buybacks to offset SBC dilution, true cash generation is far lower than reported. Sophisticated investors increasingly look at FCF minus SBC, or FCF minus buybacks, as the honest measure for equity-compensation-heavy software companies.

The SaaS DCF and the Deal Market

A DCF is still built as a cross-check, with four adaptations: a longer projection period (7 to 10 years rather than 5) to carry the company from loss-making growth to profitable maturity; the margin trajectory as the central assumption (when and how fast negative EBITDA margins reach 25 to 30 percent at scale); a terminal value that typically represents 80 to 90 percent of total value, above the usual 60 to 80 percent, because near-term cash flows contribute so little; and a discount rate matched to maturity, roughly 12 to 15 percent for unproven early-stage models against 9 to 11 percent for mature profitable platforms. The outsized terminal weight is exactly why revenue multiples, which embed the market's view of long-run profitability, remain the primary anchor.

In M&A, strategic platform acquirers pay up for product gap fills, distribution leverage through their own sales forces, and competitive preemption, while software-focused sponsors are drawn by recurring revenue and apply operational playbooks on pricing and cost. Be careful with precedent transactions: SaaS multiples are violently cyclical. Median public EV/Revenue exceeded 15x in 2021 and compressed to roughly 5 to 6x by late 2022, so a precedent from the wrong vintage can misstate value by a factor of two or more. Thematic premiums (AI capability being the current one) deserve the same scrutiny: test whether the capability is genuinely differentiated before crediting it in the multiple.

Financial Institutions: When Debt Is the Business

Banks, insurers, and broker-dealers are the most important exception to everything built on enterprise value. For these businesses, debt is not financing; it is the raw material of the operating model. A bank takes deposits and wholesale borrowings and lends them out at a spread, so interest expense is a core operating cost, not a financing charge below the line. Two consequences follow: EBITDA is not a meaningful profitability measure, and enterprise value is not a meaningful concept (summing a money-center bank's funding liabilities into EV produces an absurdly large number that measures nothing). The balance sheet is the business: the quality and composition of the loan book and investment portfolio determine earning power and risk, so valuation anchors on equity and book value rather than on EV and EBITDA.

P/TBV and the ROE Engine

The primary bank multiple is price to tangible book value:

Tangible book value is total equity minus goodwill and intangibles, approximating the net asset value that would remain if the bank were wound down and all liabilities repaid. Where the multiple lands is governed by the relationship between return on equity and the cost of equity:

• •ROE above COE: the bank creates value on every dollar of equity and should trade above 1.0x TBV.
• •ROE equal to COE: the bank earns exactly its required return and belongs at 1.0x.
• •ROE below COE: the bank destroys value and should trade below 1.0x, whatever its size.

The relationship can be formalized as $P/TBV \approx \frac{ROE - g}{COE - g}$ where g is long-term growth. A bank earning a 15 percent ROE against a 10 percent COE with 3 percent growth justifies roughly (15 - 3) / (10 - 3) = 1.7x tangible book. In practice the profitability metric watched most closely is ROTCE, return on tangible common equity: net income to common shareholders over average tangible common equity, which matches the numerator of the return to the denominator of the multiple. P/E is used alongside P/TBV to capture the market's view of earnings power, including the mix and sustainability of net interest income, fees, and trading revenue.

The Dividend Discount Model

The intrinsic method for banks is the dividend discount model, which replaces the unlevered DCF. Instead of discounting unlevered free cash flow at WACC, the DDM discounts the dividends a bank can actually pay, at the cost of equity:

It fits banks for structural reasons. Regulatory capital requirements cap distributions, so the dividend stream is a genuine measure of distributable cash rather than an arbitrary payout choice; retained earnings grow book value and future earning power; and the terminal value is typically expressed as a terminal P/TBV multiple on terminal-year book value. Note what kind of valuation this is: a levered one. It produces equity value directly from equity cash flows and the cost of equity, with no EV bridge, because enterprise value does not exist as a concept here.

Insurance, Embedded Value, and Asset Managers

Insurers share the core FIG feature (debt-like obligations are the operating engine) and are valued on P/B and P/E, with book value including the investment portfolio. Property and casualty insurers hinge on the combined ratio and reserve adequacy; life insurers on investment spreads and mortality experience.

European and Asian life insurers add a framework of their own, embedded value: the present value of future profits from the existing book of policies plus adjusted net assets.

Embedded value is rarely used in the US, so cross-border life insurance deals require bridging between the US framework (P/B, P/E, DDM) and the European one (price to embedded value), and the two can imply materially different values for the same business. Asset managers round out the FIG family: they are valued on P/E and price to assets under management, with intrinsic value from a DCF on fee income, since AUM growth and fee rates drive the economics. One M&A-specific constraint worth carrying: in bank deals, pro forma regulatory capital ratios must clear minimums after closing, and goodwill created in the purchase price allocation reduces tangible capital, which directly constrains how much a buyer can pay.

Real Estate and REITs: Valuing Around Depreciation

The distortion at the heart of REIT valuation is depreciation. US GAAP requires real estate to be depreciated over 27.5 to 39 years, producing large non-cash charges, while well-maintained property typically holds or appreciates in value. Reported EPS is therefore depressed to the point of uselessness; EV/EBITDA partially corrects but misses property-level economics; and a standard cash flow DCF never directly values the real estate itself, which may be worth more or less than the cash flows it currently generates. The sector's toolkit replaces earnings with three purpose-built measures.

NAV and Cap Rates

Net asset value marks the property portfolio to fair value and nets off liabilities:

Fair value at the property level comes from the capitalization rate: a property's annual net operating income divided by its value, functioning like a yield.

A property generating $5 million of NOI at a 5 percent cap rate is worth $100 million. Lower cap rates mean higher values and lower expected returns; prime urban assets trade at tight cap rates while secondary suburban assets trade wide. A REIT trading above its NAV carries a premium for franchise value, management, or growth; one below NAV can be bought for less than the estimated value of its buildings, which is exactly the arithmetic activists and acquirers run.

Cap rate selection is the whole game, and it is subjective. Moving a single property's cap rate from 5.0 percent to 5.5 percent on $10 million of NOI cuts its value from $200 million to $182 million, a 9 percent decline from a 50 basis point tweak, and a portfolio compounds that sensitivity across dozens of assets. Analysts triangulate cap rates from comparable property transactions, broker estimates, and published market surveys.

FFO and AFFO

Funds from operations is the sector's replacement for net income:

Adding back the depreciation and stripping one-time disposition gains isolates recurring earning power from the real estate operations; P/FFO is the REIT equivalent of P/E and the primary trading multiple, with the level varying widely by property type according to growth prospects. Adjusted funds from operations goes one step further toward true cash:

Deducting the capex genuinely required to keep the buildings competitive makes AFFO the closest proxy for sustainable, distributable cash flow, the measure that matters for dividend capacity; P/AFFO is the analogue of price to free cash flow.

Rates and REIT M&A

REITs are yield instruments, and interest rates hit them through three channels simultaneously: rising rates push cap rates up (property values and NAV down), raise the cost of debt (FFO down), and make the dividend yield less competitive against bonds (the stock down). Entire REIT indices can fall 25 to 30 percent in a hiking cycle even while occupancy and rent growth stay healthy. REIT acquisitions are priced on premium or discount to NAV, implied cap rate, and price per square foot rather than on EV/EBITDA, and merger contribution analysis runs on NAV, FFO, and portfolio quality.

Natural Resources: Valuing Depleting Assets

Oil and gas producers and mining companies share two features that overturn standard valuation. First, earnings are dominated by commodity prices that management does not control: an E&P company generating $1 billion of EBITDA at $80/barrel may generate only $400 million at $50/barrel. Second, the asset base is consumed by the act of operating: every barrel produced and every ounce milled is gone. There is no perpetuity, so there is no conventional terminal value, and the valuation must be built from the reserves themselves.

Oil and Gas: Reserve-Based NAV and PV-10

The foundational E&P methodology values the present value of cash flows from proven reserves:

The build projects production field by field, applies a commodity price assumption, deducts operating costs and capex, and discounts the net cash flows. Because the price assumption is the single most sensitive input, NAV is presented under several decks:

• •Current strip pricing: futures prices year by year, the market's own forward view.
• •The analyst price deck: the bank's house assumptions, often near-term strip blended into a long-term normalized price.
• •Flat pricing: spot held constant, a simple baseline.
• •Stress cases: low-price scenarios that test viability.

The standardized version of this calculation is PV-10: the present value of future net revenues from proven reserves discounted at a fixed 10 percent, mandated by SEC rules and disclosed in every E&P company's 10-K. The fixed rate is not the company's cost of capital; it exists to make reserve values comparable across companies.

EBITDAX and E&P Multiples

The sector's operating earnings metric adds exploration expense back to EBITDA:

Under the successful efforts accounting method, dry holes and seismic work are expensed, creating EBITDA swings unrelated to production economics; under the full cost method they are capitalized. Adding exploration back makes companies on the two methods comparable, and EV/EBITDAX replaces EV/EBITDA as the standard multiple. Supporting metrics value the asset base directly: EV per BOE of proven reserves, EV per flowing barrel (daily production), and EV per net acre for undeveloped land positions. Large E&P acquisitions are, at bottom, reserve purchases underwritten on NAV with EBITDAX multiples as the market cross-check; ExxonMobil's acquisition of Pioneer Natural Resources was priced exactly this way. Midstream businesses (pipelines, gathering, processing) shift back toward EV/EBITDA and distributable cash flow because their revenues are fee-based rather than commodity-linked, and integrated majors get a sum-of-the-parts treatment with each segment on its own metrics.

Mining: NAV, Mine Life, and Jurisdiction

Mining formalizes the same logic asset by asset:

Each mine's projection comes from its life of mine plan, an engineering document specifying annual volumes, ore grades, recovery rates, costs, and capex over the remaining productive life, disclosed in NI 43-101 (Canada) or JORC (Australia) compliant technical reports. This grounding in verified geological data makes mining cash flow projections unusually hard-edged compared with a generalist DCF. Discount rate conventions differ from WACC logic: 5 percent is the industry standard for gold (the gold price itself is treated as carrying much of the risk), 8 to 10 percent for base metals like copper and zinc, and 10 to 15 percent for development-stage projects carrying execution and permitting risk; pure exploration assets are valued on EV per resource ounce or as options rather than on NPV.

P/NAV is the sector's trading multiple, the analogue of P/TBV for banks: above 1.0x the market pays a premium for management, growth pipeline, or exploration upside; below 1.0x it discounts for jurisdictional risk, capital allocation doubts, or commodity pessimism. Senior gold producers cluster around 0.8 to 1.3x while juniors trade in far wider ranges. EV per ounce provides the quick cross-check, but only alongside reserve quality: an undeveloped resource ounce might be worth $50 to $100 while a reserve ounce in an operating mine fetches $200 to $400. Two companies with identical 10 million ounce endowments can deserve very different per-ounce values; one with a 15-year mine life, low all-in sustaining costs, and tier-1 jurisdictions (Canada, Australia) might command around $350/oz while a peer with an 8-year life, higher costs, and a riskier jurisdiction trades near $180/oz. Because deposits cannot be relocated, jurisdictional risk (expropriation, resource nationalism, permitting uncertainty, infrastructure gaps) is priced directly, with tier-1 assets commanding premiums on the order of 20 to 30 percent over equivalent assets in higher-risk locations.

Two cross-cutting points for both industries. The reserve replacement ratio (reserves added over reserves produced) is the depletion health check: above 100 percent the resource base grows, below it the company is liquidating itself in slow motion, and a producer running 300,000 ounces a year against 3 million ounces of reserves has exactly ten years of mine life to fix that. And because EBITDA in both sectors swings with the commodity cycle, multiples should be applied to mid-cycle normalized earnings, the technique covered in the cyclicals section below.

Healthcare and Pharma: Probability-Weighted Value

Healthcare spans two business types that demand different mathematics. A commercial-stage pharmaceutical company with marketed drugs is a large operating business valued on EV/EBITDA, DCF, and comps, with sector adjustments. A clinical-stage biotech with no revenue and a lead candidate in trials cannot be valued on any of those; its value is the probability-weighted expected value of the pipeline. Most large pharma companies contain both at once, which is why the sector defaults to sum-of-the-parts.

Commercial Pharma and the Patent Cliff

The defining adjustment for marketed portfolios is the patent cliff: when a blockbuster loses patent protection or regulatory exclusivity, generic or biosimilar entry typically erodes 70 to 90 percent of its revenue within 2 to 3 years. Pfizer's Lipitor, a $13 billion a year drug at peak, lost more than $10 billion of annual revenue within two years of expiry. This makes a standard terminal value assumption dangerous: perpetuity growth on the current earnings base silently assumes today's blockbusters never expire. The remedy is to model each major product individually, with revenue trajectories that build in expiration dates and generic erosion, rather than applying a blanket growth rate. EV/EBITDA for major pharma runs a wide range, with the low end for companies staring at near-term expirations and the high end for those with long exclusivity runways and deep pipelines.

rNPV: Risk-Adjusted NPV

For development-stage assets, the standard is risk-adjusted NPV, which handles the binary nature of drug development (a candidate either reaches the market or generates nothing) by probability-weighting the cash flows themselves:

P(success to year t) is the cumulative probability of surviving every clinical and regulatory gate up to that year, so weighted cash flows shrink the further out they sit. The probabilities are anchored in historical phase transition rates:

These averages mask wide variation by therapeutic area: oncology runs below them, particularly at Phase II, while rare disease and gene therapy programs run above in later stages thanks to targeted patient populations. A credible rNPV uses indication-specific rates, not the blended table. The remaining inputs are the peak sales estimate (patient population, share, pricing, competition; the most subjective and most impactful number), the revenue ramp to peak (roughly 3 to 5 years for specialty drugs, 5 to 8 for primary care), the exclusivity runway, and the cost structure.

The classic error is double-counting risk. The discount rate in an rNPV should be moderate, roughly 8 to 12 percent, because clinical risk is already carried in the probability weights; layering a 15 to 20 percent discount rate on top counts the same risk twice and systematically undervalues pipelines. Very high discount rates (some practitioners have used up to 40 percent for early-stage work) belong to the older approach that loaded all risk into the rate instead of modeling it explicitly.

Sum-of-the-Parts and Healthcare Services

Diversified pharma is valued by adding the pieces, each on its own method:

Platform value captures the manufacturing, distribution, and regulatory infrastructure that supports future launches, often estimated off corporate overhead or revenue. A compact example: a blockbuster with 6 years of exclusivity worth $60 billion on a product DCF, a marketed portfolio of eight other drugs worth $85 billion, a Phase III candidate worth $12 billion on rNPV, three Phase II candidates worth a combined $5 billion, less $15 billion of net debt, gives an SOTP value of $147 billion. The decomposition is the point: it shows where value sits and where the cliffs are, which a single consolidated multiple hides, and it explains why pharma M&A is chronically active (acquisitions refill the pipeline ahead of expirations).

Healthcare services businesses (physician practices, surgery centers, behavioral health, home health) are valued on standard EV/EBITDA with three sector overlays: reimbursement risk, since revenue depends on Medicare, Medicaid, and commercial payer rates and a richer commercial mix earns a premium; regulatory and licensure requirements, which are simultaneously a barrier to entry and a compliance risk; and provider supply constraints, since clinician shortages cap growth and inflate labor costs.

Industrials and Cyclicals: Mid-Cycle Normalization

Cyclical businesses (steel, chemicals, transportation, construction, machinery) are chronically mispriced by point-in-time multiples because the denominator moves with the economy. At the peak, EBITDA is inflated by unsustainable pricing and utilization, so trailing EV/EBITDA looks cheap exactly when the company is most expensive; buy on that multiple and the cycle turns, earnings halve, and the "fair multiple on peak earnings" reveals itself as a large overpayment. At the trough the mirror image holds: earnings are depressed, the multiple looks stretched, and the company looks expensive precisely when the assets are cheapest. Hence the counter-intuitive rule for cyclicals: the P/E screen inverts, looking richest at the bottom of the cycle and cheapest at the top. Mid-cycle normalization cuts through both distortions by estimating what the business earns under normal conditions.

Three Ways to Normalize

1. 1.Historical average: average EBITDA over a full cycle, typically 5 to 7 years spanning expansion and contraction. A chemical company posting $300M, $400M, $500M, $600M, $450M, $350M, $280M over seven years has mid-cycle EBITDA of roughly $411M, wherever it sits in the cycle today. The limitation: the average assumes earning power has not structurally changed, so acquisitions, divestitures, or cost transformation break it.
2. 2.Through-cycle returns on capital: average the company's ROIC over a full cycle and apply it to the current invested capital base, which accommodates a business that has grown or shrunk.

1. 3.Cycle position adjustment: judge where the company currently sits (early recovery, mid-cycle, late peak, downturn) and haircut or uplift current EBITDA toward the normalized level. More flexible, more judgment-dependent.

Applying the Through-Cycle Multiple

Normalized EBITDA must be paired with a through-cycle multiple: the peer group's average multiple across a full cycle, not the multiple at today's cycle position, otherwise one distortion is swapped for another. A worked contrast shows the stakes. A steel producer earns $800 million of EBITDA in a construction boom; at 8x that implies $6.4 billion of enterprise value, and the trailing multiple looks reasonable. Two years later EBITDA falls to $300 million and the company trades at $3.5 billion, an apparently expensive 11.7x. But mid-cycle EBITDA, the five-year average, is about $550 million, worth $4.4 billion at 8x: the boom-time buyer overpaid by roughly $2 billion against through-cycle value by anchoring on peak earnings.

The same discipline reaches into the DCF. Projections for a cyclical should not extrapolate the current year; from a peak they should mean-revert toward mid-cycle over the first 2 to 3 forecast years, from a trough they should recover to it, and the terminal value must be built on normalized cash flow, never on the peak or the trough. Within the industrial universe, cyclicality itself varies: aerospace and defense carries a stabilizing baseline of government revenue (with the commercial aerospace side still tied to airline capex cycles), commodity chemicals sit at the volatile extreme, and asset-heavy transport (shipping, rail) warrants an eye on fleet and rolling-stock values alongside earnings.

Retail and Consumer: EBITDAR and Unit Economics

Retail valuation runs on three layers that barely exist elsewhere: a real estate decision that distorts the standard multiples, store-level economics that determine whether growth creates or destroys value, and brand equity that sets pricing power.

EBITDAR and the Own-Versus-Lease Distortion

Two identical retailers can show very different EBITDA purely because one owns its stores (no rent, higher D&A, more debt) and the other leases (rent inside EBITDA, lighter balance sheet). EV/EBITDAR, earnings before interest, taxes, depreciation, amortization, and rent, neutralizes the difference by adding rent back. The matching principle applies: when the denominator adds back rent, the numerator must add operating lease liabilities to enterprise value, or the multiple mixes a lease-inclusive value measure with a lease-exclusive earnings measure. The same logic serves any lease-heavy industry, airlines and hotels included. Within the sector the multiple ladder runs from luxury and premium brands at the top (durable pricing power), through specialty and grocery in the middle, to undifferentiated general merchandise and value formats lower down.

Unit Economics

Aggregate financials are only meaningful against the store-level engine underneath them, and four metrics summarize it.

• •Average unit volume (AUV): revenue per location, whose growth signals demand and pricing power.
• •Four-wall margin: store-level profit after rent, labor, COGS, and utilities, divided by store revenue. Around 20 to 25 percent is healthy, below 15 percent is fragile, above 30 percent exceptional.
• •Same-store sales (SSS) growth: year-over-year revenue growth at stores open at least 12 months, isolating organic demand from unit expansion. It decomposes into traffic (visits) and ticket (spend per visit); traffic-led growth is worth more than ticket-led growth, which may just be inflation pass-through with a ceiling. Two or more consecutive quarters of negative SSS is a red flag that overrides other metrics, signaling brand erosion or competitive leakage, and it makes every new store opening riskier.
• •Payback period: years for a new store to return its buildout capital. Two years is excellent, four to five the outer limit.

The strategic payoff of these numbers is scalability. A 500-store chain with strong unit economics and headroom to 2,000 stores deserves a materially higher multiple than a 500-store chain that has saturated its market, because each incremental opening in the first case creates value and in the second destroys it.

Brand, Franchise Economics, and DTC

Brand equity is the intangible that makes consumer revenue predictable and margins defensible; strong brands command premium multiples and weak ones trade at discounts vulnerable to promotion wars and channel shift. Franchise-heavy models carry a structural premium over company-operated peers: royalty streams of roughly 4 to 6 percent of franchisee sales are high-margin, capital-light, and recurring, and because franchisees fund the buildouts, FCF conversion runs 80 to 90 percent of EBITDA against 40 to 60 percent for company-operated fleets, justifying a premium of several turns of EBITDA. Private equity roll-ups exploit multiple arbitrage in fragmented retail and restaurant categories: acquire a platform around 10 to 12x, bolt on smaller targets at 6 to 8x, exit the scaled platform at 12 to 14x. And high-growth direct-to-consumer brands with limited profitability are valued like technology companies, on EV/Revenue, because their unit economics are still maturing toward an expected at-scale margin.

Media and Entertainment: Subscribers, ARPU, and Content

Streaming rebuilt media economics around subscription relationships, and the valuation framework followed. The metrics divide into the subscriber economics of platforms, the sum-of-the-parts problem of diversified conglomerates, and the managed decline of linear advertising businesses.

Subscriber Economics

The headline platform metric prices the customer relationships directly:

Netflix at roughly 300 million subscribers and around $400 billion of enterprise value implies about $1,300 per subscriber; platforms with weaker monetization command a fraction of that. What separates them is the quality of each relationship, captured by two numbers. ARPU, average revenue per user, is streaming revenue over average subscribers, reflecting pricing power, tier mix, and advertising contribution; rising ARPU means extracting more value from the same base. Churn is the share of subscribers canceling per period; low churn signals content engagement and switching costs, and it converts directly into subscriber lifetime value:

At $15 monthly ARPU, 2 percent monthly churn implies an LTV of $750 per subscriber while 5 percent churn implies $300: a three-point churn difference destroys 60 percent of subscriber value, which is why two platforms with identical subscriber counts can deserve wildly different EV/subscriber multiples. High churn also forces heavy content and marketing spend just to hold the base flat, compressing free cash flow.

The framework itself is migrating. In the land-grab years platforms were valued on subscriber growth and market share, with losses tolerated; as the leaders matured, the market shifted decisively to operating margin and free cash flow, and mature platforms increasingly get valued like standard operating companies on EV/EBITDA. Netflix's decision to stop reporting quarterly subscriber counts formalized the transition: subscriber count is a scale-era metric, and profitability is the mature-era one.

Conglomerates, Content Libraries, and Linear Decline

Diversified media companies are sum-of-the-parts stories because their segments obey different economics: streaming on EV/Revenue or EV/subscriber migrating toward EV/EBITDA, theme parks on EV/EBITDA driven by attendance and per-capita spend, film studios on a content library DCF plus the value of the annual slate, and linear TV networks on EV/EBITDA applied to structurally declining earnings. The conglomerate wrapper itself has historically cost around 13 to 15 percent, a discount reflecting investor difficulty valuing the pieces and skepticism about capital allocation across them, which is precisely what fuels spin-offs and activist campaigns.

Valuing the content library means projecting licensing and distribution cash flows from the existing catalog (adjusted for the strategic trend of withholding content for one's own platform), distinguishing franchise properties with effectively perpetual value from ordinary titles that fade, and crediting the retention value of catalog that keeps subscribers from churning even when it earns no direct revenue. Content spend should be read as investment rather than cost: the analytical question is the return on each content dollar in retention, acquisition, and ARPU, and platforms that demonstrably earn high returns on content deserve higher multiples. For advertising-driven traditional media, valuation is EV/EBITDA with a trajectory adjustment: linear advertising is in secular decline from cord-cutting, so digital-revenue mix earns a premium, and audience metrics (ratings, reach, demographics) drive the CPMs that set revenue quality.

Infrastructure and Utilities: The Regulated Asset Base

A regulated utility is the one business whose earnings are set by a regulator rather than by competition, and the entire valuation framework flows from that fact. The regulator determines what the utility may invest, what return it may earn on that investment, and how the costs are recovered through customer tariffs:

The regulated asset base (RAB), called the rate base in the US, is the value of the infrastructure (wires, pipes, plants, treatment facilities) on which the utility earns its allowed return; it grows with capex and shrinks with depreciation. The allowed return is set in periodic rate cases, typically every 2 to 5 years, from a cost-of-capital assessment; US allowed ROEs for electric utilities usually land around 9.0 to 10.5 percent. The arithmetic makes rate cases the highest-stakes events in the sector: on a $20 billion rate base, a 50 basis point cut in the allowed return removes $100 million of annual earnings. Strictly, the allowed ROE applies to the equity-funded share of the rate base, so on a 50 percent equity layer the impact halves to $50 million; the full-RAB version is the simplified convention for quick math, and the distinction is worth stating when you use it.

Growth follows mechanically: the only durable way a utility grows earnings is to grow the RAB through capital investment, since every incremental dollar of asset base earns the allowed return. That is why utility equity stories are capex stories, and why the sector runs a self-reinforcing loop of investment, earnings growth, dividend growth, and fresh capital raising.

How Utilities Are Valued

• •Dividend discount model: utilities are the sector where the DDM works best, because regulated earnings make dividends predictable, payout ratios run high at 60 to 75 percent, and growth tracks a disclosed, pre-approved capex plan. The build is typically two or three stages: explicit dividend projections off the capex plan, then a sustainable terminal growth rate.
• •EV/EBITDA: standard but structurally low-growth, with electric utilities typically around 8 to 12x and water utilities at a premium (roughly 12 to 17x) for scarcity value and essentiality.
• •P/E: unusually prominent here because the investor base is equity income buyers who think in EPS and dividend yield.
• •EV/RAB: the sector-specific multiple, enterprise value per dollar of regulated assets. Around 1.0 to 1.2x is fair value for a utility earning its allowed return; above roughly 1.5x prices superior allowed returns or growth; below 1.0x prices regulatory risk or management concerns.

Two risk lenses complete the picture. Utilities are bond proxies: their stable yields compete with fixed income, so a 100 basis point move in long Treasury yields can move utility valuations 10 to 15 percent with no change in fundamentals, rallying when rates fall and derating when they rise. And regulatory risk is the sector-specific exposure: a regulator cutting the allowed ROE reduces earnings across the entire rate base at a stroke, so the quality of the regulatory relationship in each jurisdiction matters as much as the financial metrics, especially in M&A.

Choosing the Right Approach

The synthesis is one sentence: the industry determines the methodology, never the other way around. The first analytical step in any valuation is identifying what type of business you are looking at, because that answer selects the metrics, the multiples, and the intrinsic model all at once.

Three principles govern the choice. First, match the metric to the value driver: recurring revenue for SaaS, tangible equity for a bank, reserves for a mine; the denominator must capture what actually creates value in that model. Second, triangulate even when the sector method is well suited: check a mining NAV against EV/EBITDAX comps, a biotech rNPV against EV/Revenue comps for similar clinical-stage names, a utility DDM against P/E and EV/RAB. No single method is sufficient anywhere. Third, understand why the standard approach fails before reaching for the alternative; knowing that banks use P/TBV is trivia, while knowing that deposits are operating raw material is the understanding that lets you apply and defend the framework.

The sum-of-the-parts thread that ran through this survey deserves its own mention: whenever one company spans several of these economic models (an integrated oil major, a diversified pharma, a media conglomerate), each segment gets valued on its own appropriate method and the pieces are summed. And three mismatch errors account for most sector-related valuation failures in junior work and interviews: applying EV/EBITDA to a financial institution, applying a normal multiple to a cyclical's peak-year EBITDA, and valuing a pre-revenue biotech on earnings multiples it does not have. Each produces output that looks precise and is fundamentally wrong, which is worse than no output at all.