Skip to main content

Which Multiples Work Best for Each Sector?

Not all multiples are created equal. A metric that reveals value in one sector can obscure it in another. Price-to-earnings tells you something useful about a bank's valuation but nearly nothing about an oil exploration company that spent $500 million on a discovery with no current earnings. Return on invested capital dominates the valuation of a utility but says little about a biotech company that may never generate positive ROIC on early development costs. Understanding which multiples best capture the economic reality of each sector—and what adjustments each requires—separates sophisticated investors from those blindly averaging P/E ratios across sectors.

Quick Definition​

Sector-specific multiples are valuation benchmarks tailored to an industry's business model, capital requirements, and cash flow characteristics. Technology companies are best valued on EV/Revenue or forward earnings multiples reflecting anticipated growth. Utilities are best valued on dividend yield, ROIC, and regulated return assumptions. Banks are best valued on price-to-book and forward P/E. Oil companies are best valued on EV per unit of production. Each sector has metrics that illuminate value and others that distort it.

Key Takeaways​

  • Technology and high-growth sectors: EV/Revenue and forward P/E weight growth; EV/EBITDA may miss reinvestment needs
  • Utilities and infrastructure: Dividend yield, regulated return on equity, and price-to-book reveal fair value; P/E is less useful
  • Financial services: Price-to-book, loan-to-deposit ratios, and stressed capital metrics matter; P/E is backward-looking and cyclically distorted
  • Energy and cyclicals: EV/EBITDA (through-cycle normalized) and EV/production reveal value; P/E is unreliable at peaks and troughs
  • Healthcare and pharmaceuticals: EV/Revenue, forward P/E (weighted for pipeline), and royalty rates; early-stage companies require probability weighting
  • Real estate: NOI yield, cap rate, and FFO multiples; P/E is nearly meaningless due to leverage and depreciation accounting
  • Retail and consumer: Same-store sales growth, EV/sq ft, and forward multiples matter; P/E distorted by depreciation and leverage

Technology and Software​

Best Metrics​

  • EV/Revenue: For unprofitable or early-stage SaaS; typical range 5-10x for 30%+ growers
  • Forward P/E: For profitable, mature SaaS; typical range 20-35x for 15%+ growers
  • PEG Ratio: Forward P/E divided by expected growth rate; <1.5 suggests reasonable valuation
  • EV/EBITDA: Less useful for asset-light software but works for telecom and infrastructure software

Why These Work​

Software companies have three characteristics that drive metric choice:

  1. Capital-light: No factories or mining operations; minimal capex relative to revenue
  2. High margins: Mature SaaS typically runs 30-50% EBITDA margins; unprofitable startups may run at 5-10% negative
  3. Fast growth: 20-40% annual growth is common for successful SaaS; this justifies high multiples

EV/Revenue works for unprofitable companies because it sidesteps the question of whether they'll ever be profitable. A SaaS company generating $100M in revenue at 50% gross margin is economically valuable even if it's reinvesting all gross profit into R&D. EV/Revenue at 5-8x ($500-800M valuation) may be fair if the company is moving toward 30%+ operating margins.

Forward P/E works for profitable companies because it captures the full picture: size, growth, and profitability. A company earning $1 per share growing 25% annually deserves a 30-35x multiple (PEG ~1.2-1.4).

Common Pitfalls​

1. Using Trailing P/E (GAAP earnings) on unprofitable or early-stage companies

Amazon was unprofitable for years despite massive revenue growth. Comparing its P/E to General Electric was nonsensical. Use EV/Revenue instead.

2. Assuming SaaS margins are normal

Mature SaaS (Adobe, Salesforce) runs 30-40% operating margins. Early-stage SaaS (Datadog at IPO) was 20% negative margins. Don't assume they're equivalent even if both trade at 8x revenue.

3. Ignoring free cash flow conversion

High-revenue growth doesn't matter if the company is burning cash faster than ever. Compare EV to free cash flow, not just revenue.

Utilities and Infrastructure​

Best Metrics​

  • Dividend Yield: Most important; compare to historical and peer range
  • Regulated Return on Equity (ROE): What regulators allow; typically 8-11%
  • Price-to-Book: If ROE is above cost of equity, P/B >1x is justified
  • EV/EBITDA: Works but less intuitive than yield-based metrics

Why These Work​

Utilities are capital-intensive but generate predictable cash flows with regulatory protection. A utility investing $100M to earn a regulated 9% ROE ($9M/year) has transparent economics. The value is the present worth of that perpetual 9% return stream.

If a utility trades at 8% yield (meaning investors demand 8% return) and regulators allow 10% ROE, the stock is undervalued. Investors should push for the spread. If it trades at 4% yield (investors content with 4%) and ROE is 10%, it's expensive.

Price-to-book reveals the same story economically. If a utility's assets are worth $100M in book value and it can earn 10% on those assets, annual earnings are $10M. If required return is 8%, the asset base is worth $125M in present value terms. P/B should be 1.25x. If it trades at 0.9x, it's undervalued.

Common Pitfalls​

1. Ignoring regulatory risk

Utilities are only as good as their regulators. If a state regulator slashes allowed ROE from 10% to 8%, equity value collapses. Political risk is real.

2. Assuming perpetual dividend growth

Many utilities have raised dividends for 50+ years. But dividend growth is constrained by ROE and payout ratios. A utility with 10% ROE and 70% payout ratio can sustain ~3% long-term dividend growth. Don't assume 5-6%.

3. Confusing leverage with value

A utility financed 60% with debt and 40% with equity appears to have higher equity returns. But leverage increases equity risk. Compare utilities on an unlevered, after-tax basis.

Financial Services (Banks, Insurance, Asset Managers)​

Best Metrics​

  • Price-to-Book: Most important for banks; 1.0-1.5x is typical for healthy banks
  • Tangible Book Value Per Share: Book value minus intangibles; more conservative
  • Forward P/E: On normalized, through-cycle earnings
  • ROE and ROIC: What matters more than size
  • Loan-to-Deposit Ratio: Funding stability; should be <100% for safety

Why These Work​

Banks earn spreads (lending at 5%, borrowing at 2%, capturing 3% margin). Their capital is equity, which they deploy into loans and securities. Return on equity captures how profitable that deployment is. If ROE is 12% and required return is 10%, P/B should be 1.2x.

P/B works because it relates directly to ROE via the return equation: Value = Book Value × (ROE / Cost of Equity).

Trailing P/E is less useful because bank earnings are cyclical. In 2007, major banks earned high returns and traded at 12-15x. In 2008, earnings collapsed and P/E became meaningless. Forward P/E on normalized earnings is better.

Common Pitfalls​

1. Valuing banks at cycle peaks or troughs

In boom times, bank earnings are inflated by low loan losses and high deal activity. In recessions, earnings are crushed. Always use through-cycle earnings, not trailing.

2. Ignoring capital ratios and regulatory constraints

A bank earning high ROE but maintaining minimal capital ratios faces higher failure risk. Compare Tier 1 capital ratios and stress-test results.

3. Confusing deposit size with profitability

Large deposit bases are valuable only if they're profitable. Many banks have taken on deposits at low margins to gain scale, eroding ROE.

Energy (Oil, Gas, Mining)​

Best Metrics​

  • EV/EBITDA (normalized/through-cycle): Most important; use average commodity prices over 5-10 years
  • EV/Production: For upstream oil/gas; value per barrel of oil equivalent (BOE)
  • Price per Reserve: For exploration companies; value per barrel of reserves
  • FCF Yield: On normalized free cash flow
  • P/E (forward, normalized): Only on normalized earnings

Why These Work​

Energy is cyclical. Oil at $120/barrel generates vastly different cash flows than oil at $50. Valuing an oil company at $120 oil prices vastly overvalues it. Through-cycle analysis uses normalized commodity prices (typically $70-90 for oil in recent years) to calculate sustainable cash flows.

EV/EBITDA on normalized EBITDA avoids the peak/trough distortion. A producer with 200 million barrels of reserves, producing 50 million barrels annually at $70/barrel oil, generates ~$3.5B EBITDA at reasonable margins. If it trades at $14B enterprise value, that's 4x through-cycle EBITDA—a reasonable multiple.

EV/Production for upstream companies captures the value of replacement cost. If a company produces 50 million BOE annually, EV/50M = $ per BOE. Compare to peer producers and replacement cost of reserves.

Common Pitfalls​

1. Using current commodity prices to value energy companies

Oil at $150 is not normal and will not persist. Using peak pricing destroys valuation discipline.

2. Ignoring reserve replacement and depletion

An energy company is a depleting asset. If it's not replacing reserves through drilling or acquisition, it's a shrinking business. Check reserve replacement ratios.

3. Conflating cash flow with free cash flow

Energy companies often report high operating cash flow but face enormous capex (offshore platforms, pipelines). Calculate true FCF = operating CF - capex.

Flowchart​

Healthcare and Pharmaceuticals​

Best Metrics​

  • EV/Revenue: For early-stage biotech with pipeline
  • Forward P/E: For mature pharma with established products
  • P/Sales (risk-adjusted): Weight by probability of pipeline success
  • Royalty Rates: As % of peak sales for drugs in trials
  • Cost of Capital: Risk-adjusted for clinical trial risk

Why These Work​

Pharma earnings come from approved drugs; biotech earnings come from... potentially nothing. Early-stage biotech has no earnings, only prospects. EV/Revenue captures the bet on whether the pipeline converts to revenue.

More sophisticated is probability-weighted valuation: a drug in Phase I trials has 10% success probability; Phase III has 70%. Weight expected revenues accordingly. If a Phase III drug can generate $500M peak revenue at 40% net margin ($200M EBITDA), and it's 5 years to approval, the probability-adjusted value is $200M × 0.7 × 1/(1.10^5) = ~$124M.

Mature pharma is easier: established drugs with known revenues use standard P/E and EV/EBITDA. The key is separating cash generated by existing products from reinvestment needs.

Common Pitfalls​

1. Ignoring patent cliffs

A drug generating $2B in annual revenue with a patent cliff in 2 years has only 2 years of value, not perpetual value. Biotech and pharma earnings quality depends entirely on product durability.

2. Using trailing earnings on clinical-stage companies

Early-stage biotech doesn't have earnings. Don't use P/E. Use EV/Revenue, risk-adjusted DCF, or sum-of-the-parts (valuing each drug separately).

3. Assuming all biotech is high-growth

Failed trials, adverse events, and competitive pressure kill value quickly. Many biotech companies are gambling businesses, not growth businesses. Price accordingly (lower multiples for execution risk).

Real Estate (REITs)​

Best Metrics​

  • Capitalization Rate (Cap Rate): NOI / Property Value; 4-7% typical depending on property type and location
  • Funds From Operations (FFO) Multiple: EV/FFO; 12-20x typical
  • Price-to-Book: For asset-heavy REITs; 0.8-1.2x typical
  • Dividend Yield: Most REITs pay 70-90% of FFO as dividends

Why These Work​

REITs own real estate and must distribute 90% of taxable income to shareholders. Their value comes from the rental income stream. Cap rate directly reflects this: a property generating $10M NOI (net operating income) valued at $125M has a 8% cap rate.

If investors demand 6% return, the property is fairly valued. If they demand 8%, it's undervalued. Cap rates embed required returns directly.

FFO (net income plus depreciation and amortization) is more meaningful than GAAP earnings for REITs because depreciation is a non-cash charge. A REIT earning $10 in GAAP earnings but $15 in FFO is economically stronger.

Common Pitfalls​

1. Using P/E instead of FFO multiples

Depreciation distorts GAAP earnings; use FFO to understand true profitability.

2. Assuming cap rate spread over bonds perpetually

A REIT offering 4% yield (cap rate) minus 2% bond yield (180 bp spread) looks cheap. But the 180 bp spread covers property risk, tenant credit risk, and management quality. Don't assume it tightens unless fundamentals improve.

3. Ignoring leverage and refinancing risk

REITs often use 40-50% leverage. If refinancing comes at higher rates, cap rates must widen and values fall. Check debt maturity schedules.

Retail and Consumer Discretionary​

Best Metrics​

  • Same-Store Sales (SSS) Growth: Year-over-year sales growth in comparable stores; most important
  • EV/Square Foot: For physical retailers; value per sq ft of sales space
  • Forward P/E: On normalized earnings
  • Return on Invested Capital: On store base; reveals productivity
  • EV/EBITDA: Less useful than comps-based multiples but works for cross-sector comparison

Why These Work​

Retail is about same-store sales growth. A retailer with negative SSS is shrinking even if opening new stores. Positive, accelerating SSS indicates a strong business. These metrics should flow through to forward earnings.

EV/Square footage reveals real estate economics. If a retailer trades at $500 per square foot and generates $2000 annual revenue per square foot with 10% margins, that's $200 profit per square foot. That's an unusually high return.

Common Pitfalls​

1. Confusing total sales growth with like-for-like growth

Opening stores is easy; getting sales to grow in existing stores is hard. A retailer growing total revenue 20% but with negative SSS is in trouble. Adjust multiples accordingly.

2. Ignoring e-commerce disruption

Traditional retail has lost share to e-commerce. Compare traditional retailers' growth to e-commerce penetration and their own e-commerce sales. Omnichannel retailers deserve premium multiples if SSS is growing.

3. Using historical margins

Retail margins are under structural pressure. Don't assume 15% operating margins if the company has run 12% for 5 years.

Real-World Examples​

Valuing Tesla vs. Traditional Auto​

Tesla grew at 50% annually and trades at 50-60x P/E. Ford grows at 0-5% and trades at 5-7x P/E. The multiple difference is justified—Tesla is a growth company, Ford is cyclical. But Tesla uses EV/Revenue (8-10x) as well, acknowledging that profitability is not guaranteed. Traditional auto uses P/E, relying on established profits.

Key insight: Tech multiples for Tesla, traditional multiples for Ford, because their economic models are fundamentally different.

Valuing Exxon vs. Permian Producers​

In 2022, Exxon traded at 8x EV/EBITDA (normalized) with $60-65 oil assumptions. Smaller Permian producers traded at 4-5x the same metrics. The discount for smaller producers reflected execution risk and reserve depletion concerns. Using EV/EBITDA normalized (not spot prices) made the comparison meaningful.

Valuing JPMorgan Chase vs. Regional Banks​

JPMorgan traded at 1.3x price-to-book with 15% ROE. Regional banks traded at 0.7-0.9x with 10-12% ROE. Multiples reflected earnings power and durability. JPMorgan's diversification and scale justified higher multiples despite lower absolute ROE (capital constraints, regulatory capital ratios).

Common Mistakes​

1. Applying the Same Metric Across Sectors

Using P/E for software, utilities, energy, and banks is like using temperature to measure distance—technically possible but nonsensical. Each sector has metrics that reveal value and others that distort it.

2. Forgetting Cyclicality

Almost every sector has cyclical elements. Energy is obviously cyclical; so are banks, homebuilders, and materials. Normalize through-cycle multiples or face severe mispricings.

3. Comparing Multiples Across Countries Without Adjusting

A utility in Germany trading at 12x forward P/E and one in the U.S. at 14x appear similar but may not be. Adjust for risk-free rates, growth expectations, and regulatory frameworks.

4. Ignoring Reinvestment Needs

A tech company with 40% EBITDA margins but requiring 15% revenue growth in capex is not as profitable as one requiring 5%. Compare FCF/revenue, not just EBITDA margins.

5. Using Trailing Multiples in Fast-Growing Sectors

A SaaS company growing 40% annually will have vastly different forward and trailing multiples. Always use forward multiples for high-growth companies.

FAQ​

Q: Which multiple is most important: P/E, EV/EBITDA, or EV/Revenue?

It depends on sector and company stage. For mature, profitable companies in most sectors: P/E. For cyclicals, leveraged, or capital-intensive: EV/EBITDA. For unprofitable or early-stage: EV/Revenue. No single metric works everywhere.

Q: How do I choose between forward and trailing multiples?

Use forward for high-growth companies (growth embedded in future earnings, not yet realized). Use trailing for stable companies (current earnings best reflect normalized profitability). For cyclicals, use normalized earnings from neither trailing nor forward, but adjusted for cycle.

Q: Can I use one multiple for the entire market?

No. The S&P 500 in 2024 trades at ~20x forward P/E, but that masks huge variation: tech at 30x, financials at 12x, energy at 9x. Use sector-appropriate multiples and compare within sectors.

Q: What's the most dangerous metric mistake?

Comparing companies at different cycle points. Valuing an energy company at peak cycle multiples or a bank at trough cycle earnings destroys valuation discipline. Always normalize or use appropriate multiples.

Q: How do I know if a metric is appropriate for a sector?

Ask: (1) Does it capture the economic driver of value? (2) Does it work for both profitable and unprofitable companies in the sector? (3) Can I explain why this multiple is high or low? If you can't answer these, the metric is wrong for that sector.

  • How to Choose Comparable Companies — Build disciplined comps universes using sector-appropriate peers and multiples.
  • Valuation Multiples Explained — Deep dive into P/E, EV/EBITDA, EV/Revenue mechanics and limitations.
  • Cyclical Stock Valuation Traps — Master the specific mistakes that plague valuation of cyclical and commodity-linked stocks.
  • The Problem with Multiples — Understand when multiple-based valuation fails and why DCF provides crucial validation.

Summary​

Different sectors require different valuation metrics because they have different economic characteristics. Technology companies, capital-light and fast-growing, are best valued on revenue and forward earnings multiples. Utilities, capital-intensive with regulated returns, are best valued on dividend yield and ROE. Banks, dependent on leverage and net interest margins, are best valued on price-to-book and forward earnings. Energy, cyclical and commodity-dependent, is best valued on normalized EBITDA. Healthcare, uncertain and pipeline-dependent, requires probability weighting and forward metrics. Real estate, asset-rich and income-generating, is best valued on cap rates and FFO. Retail, growth-dependent, relies on same-store sales momentum and forward multiples. Mastery of sector-appropriate metrics and the adjustments they require separates disciplined analysts from those who mechanically average P/E ratios across incomparable businesses.

Next​

Continue to Traps in Valuing Cyclical Stocks to learn the specific mistakes that plague analysis of companies with commodity exposure, economic sensitivity, or cyclical earnings patterns.