Comps vs DCF: when each wins
Two fundamental investors can face the same company with identical historical data and arrive at completely different valuations depending on which tool they reach for first. One deploys a peer-set comparable analysis and concludes the company is trading at a 15% discount to fair value. The other builds a discounted cash flow model and argues it is worth 40% less than current price. Who is right? The answer lies in understanding what each tool actually measures and when that measurement matters most.
Quick definition: Comps measure what the market currently pays for similar companies; DCF estimates intrinsic value based on a company's standalone cash flow generation and perpetual earning power. Comps work best when the market is pricing reasonably; DCF works best when the market has mispriced growth, margins, or terminal value.
Key takeaways
- Comps are fastest and least assumptive; they reveal market consensus and are useful for sanity checks and relative positioning within a sector
- DCF forces explicit assumptions about cash flow, growth deceleration, and terminal value, making all hidden beliefs visible and testable
- Comps fail when the peer set is all wrong (wrong business model, wrong growth profile); DCF fails when your assumptions about the future are systematically off
- The best approach layers both: use comps to set a base case, then DCF to test whether the market's implied growth and margin expectations are rational
- For value traps and high-growth stories, DCF and comps can diverge wildly; understanding why matters more than choosing one tool
Two Roads to Valuation
What comps actually measure
Comparable multiples are a market-derived valuation baseline. When you calculate that a peer set trades at an average of 18x P/E or 6x EV/Revenue, you are answering a single question: "What are investors willing to pay today for a dollar of earnings or revenue from companies like this?" This is a powerful input, but it is fundamentally a measurement of consensus, not intrinsic value.
Comps embed all of the market's collective expectations: the growth rate everyone assumes, the margin path everyone projects, the duration of competitive advantage everyone believes exists. When the market is relatively efficient—when peers are truly comparable and expectations are anchored in fundamental reality—the market's consensus multiple is a reliable guide. You can be reasonably confident that buying a company at median-peer multiples will not lead to permanent capital impairment, because you are buying in line with what others are willing to pay.
But comps also hide all of that embedded information inside a single number. A 20x P/E ratio on a growth stock does not tell you whether the market expects 30% growth or 50% growth. A 4x EV/Revenue on a SaaS company does not reveal what profitability the market is pricing in. You see the price paid but not the assumptions animating that price.
The speed advantage. You can run comps in 30 minutes: screen peers, gather trailing multiples, calculate median or mean, apply to target company, done. A rigorous DCF takes hours of work building revenue and margin projections, estimating a WACC, stress-testing terminal-value assumptions. When you are allocating capital under time pressure—screening a universe of 50 stocks for the 5 to investigate deeply—comps are the practical choice. They let you eliminate obvious mispricings fast.
The consensus check. In a long-time-horizon portfolio, comps are your reality check. If you believe a company is worth 40% more than the market suggests, but comps show it is trading at a slight premium to peers, you should be uncomfortable. Maybe your DCF is overoptimistic on growth or margin trajectory. Maybe the market knows something about competitive risk you have not yet internalized. Comps force humility.
What DCF actually measures
A DCF builds a cash flow projection for the company and discounts it back to present value. Unlike comps, which ask "what is the market paying?", DCF asks "what would this company be worth if its cash flows played out exactly as I project?" It is a first-principles calculation: no peer set, no market consensus, just your explicit assumptions about growth, margins, capex, and the cost of capital.
The power of DCF is that it surfaces every assumption. You must commit to a view: What is next year's revenue? The year after? When do margins inflect? How fast do margins improve? When do they stabilize? What is the terminal growth rate? What is the cost of capital? Every assumption is visible, challengeable, and stress-testable.
The weakness of DCF is the same: if any assumption is materially wrong, the entire valuation breaks. If you assume 35% long-term growth but the company actually decelerates to 15%, your DCF will be wildly optimistic. If you assume a cost of capital of 8% but it should be 12%, your valuation balloons 30–40%. Unlike comps, which are anchored to real market prices of comparable companies, DCF is only as good as your forecast.
The leverage to key variables. In a typical DCF, 40–60% of value derives from the terminal value—the assumption about what the company is worth at the end of your explicit forecast period (usually 5 or 10 years). This is both a feature and a bug. It is a feature because it forces you to think about the long-term durability of the business and the steady-state growth rate. It is a bug because that terminal value is your least certain assumption.
Consider a high-growth software company. In Year 1–5 of your DCF, you forecast 30% CAGR and operating margins expanding from 5% to 20%. This is granular and testable against current trends. But your terminal value assumes the company grows at 2.5% perpetually and maintains 25% operating margins forever. This is a guess. If the terminal growth is actually 1%, the valuation drops 15–20%. If the company does not reach 25% margins and instead maxes out at 18%, the valuation drops another 10–15%.
Comps work best when the market is reasonably priced
Comps are most reliable in mature, efficient markets where the peer set is truly comparable and expectations are well-anchored. Consider a large-cap financial institution like JPMorgan or Bank of America. The peer set is large and clear (other large US banks). The business model is standardized (commercial and investment banking, asset management, consumer finance). Earnings are relatively visible and non-volatile. Historical P/E ratios give good guidance.
In such cases, comps provide a robust sanity check. If JPMorgan is trading at 12x P/E and peers are at 11–13x P/E, and the company's growth, margin profile, and risk are truly in-line with the peer set, then 12x is a reasonable valuation. A DCF on JPMorgan, in contrast, requires explicitly forecasting interest-rate paths, credit cycles, regulatory changes, and competitive dynamics—all highly uncertain. The comps approach is wiser: buy when it is cheap relative to peers, sell when it is expensive.
Similarly, large consumer-staples companies like Coca-Cola or Procter & Gamble operate in mature, cyclical markets. Growth is low and visible (2–4% annually). Margins are stable (operating margins 15–20%). Competitive moats are durable and tested through decades. Peers are easily identified. Comps will give you a reasonable sense of whether the stock is expensive or cheap relative to other stable, low-growth consumer franchises. A DCF adds little value because the company's future is largely visible in its past.
DCF wins when growth is mispriced or margins are at inflection
DCF shines when the market is systematically mispricing a key variable—usually growth or margin trajectory. This happens in three main scenarios:
The ignored groomer. A company is growing much faster than the market realizes, and multiples are anchored to legacy, lower-growth comps. Consider a software company that was treated as a mature, single-digit-growth business for years, but has accelerated to 30% growth as a new product gains traction. Comps would still be drawn from other mature software companies trading at 8–10x revenue. But this company has the growth profile of a much more valuable asset. A DCF that explicitly models the growth acceleration will identify the mismatch.
Example: Adobe in the early years of Creative Cloud transition. When Adobe began shifting from perpetual licensing to cloud subscription (2012–2015), investors initially valued it as a legacy software company trading at 15–18x P/E. But the subscription model was far more valuable: recurring revenue, higher margins over time, lower customer churn. Comps to perpetual-license software undervalued it. A DCF recognizing the margin expansion path from subscription recurring revenue would have captured the value the market was missing.
The value trap. A company is trading at a cheap multiple relative to peers because the market senses structural margin or growth decay. But the decay is transient, and the company will snap back. Comps anchor on the depressed current state; a DCF with assumptions about margin recovery will identify the opportunity.
Example: Costco in the late 1990s. It traded at 25–30x P/E while most retailers traded at 10–15x. Comps suggested Costco was absurdly expensive. But a DCF recognizing that Costco's business model (membership model, high inventory turnover, high member loyalty) supported both higher growth and superior margins would have justified the premium. The market was right, but comps would have misled you into thinking it was overpriced.
The broken business recovering. A company faced a temporary shock (loss of major customer, competitive threat, operational misstep) that depressed multiples. If the company will recover, comps will dramatically undervalue it because peers are not facing the same shock.
Example: Amazon Web Services was a small, unprofitable division in Amazon's 2008–2010 financials. If you ran comps on Amazon as a whole, it traded at modest multiples relative to growth—because the core retail business was low-margin and slow-growing. But a DCF that explicitly modeled AWS as a high-margin, fast-growing business separate from retail would have identified that AWS was the hidden gem justifying a much higher valuation. Comps were hiding AWS's value because they could not separate it.
DCF fails when your forecast is systematically wrong
The kryptonite for DCF is forecast error. If you systematically overestimate growth or underestimate competition, your DCF will be wrong by orders of magnitude. This is especially true for:
Disruptive technologies. How much will AI revenue matter to a cloud infrastructure company in 2030? Will autonomous vehicles eliminate a major automotive manufacturer's cost structure or destroy it? How will streaming cannibalize traditional media? These are genuinely unknowable. A DCF that makes confident assumptions about disruption can be spectacularly wrong. In such cases, comps to similar companies (or historical precedent of other tech disruptions) may be safer than building a DCF on a scenario you cannot validate.
Cyclical peak or trough. In a cyclical industry (autos, semis, construction), if you build a DCF at the peak of a cycle, you will assume high margins that will inevitably compress. If you build it at a trough, you will assume margins that will recover. A comps check—looking at what the company traded for at similar points in the cycle historically—can guard against this.
Competitive breakdowns. Your DCF assumes the company maintains its current market share and competitive position. But if a new competitor emerges or technology shifts, market share can evaporate. This is why comps matter as a reality check: if the entire peer set is collapsing in multiples, maybe your single-company DCF is missing something about the competitive environment.
The hybrid approach: layer them together
The best investors use comps and DCF not as either-or, but in tandem, allowing each to inform and challenge the other.
Step 1: Comps as your starting point. Build a peer-set analysis and identify the median and mean multiples. This is your first answer: "What does the market think this company is worth?" Calculate the company's implied valuation using these multiples.
Step 2: Reverse-engineer the comps assumptions. Take the peer-set multiple and reverse-engineer what growth rate and margin path the market is implicitly pricing in. If the peer set trades at 18x P/E on average, and earnings growth is 12%, the market is assuming P/E / Growth ratio of 1.5x. Is that assumption reasonable for the growth trajectory and risk profile? Or is it too optimistic? This exercise exposes whether the market's consensus is rational.
Step 3: Build a DCF. Now build your DCF with explicit assumptions about revenue growth, margins, and terminal value. Arrive at an intrinsic value. Compare it to the comps-implied value.
Step 4: Reconcile the difference. If comps and DCF agree, you have high conviction. If they diverge, dig into why. Is your DCF too optimistic about growth or margins? Or is the market missing something—is the comps multiple anchored to legacy expectations? What would have to be true for each valuation to be correct?
Example: Consider a $50 billion market-cap SaaS company with $2 billion revenue and 40% growth.
-
Comps: Peer set (ServiceNow, Workday, CrowdStrike) trades at 9x EV/Revenue. Implied value: $18 billion EV (not including net debt). Company trades at $50B market cap. Conclusion: 2.8x overvalued relative to peers.
-
DCF: Build explicit models for revenue (40% growth for 3 years, decelerating to 15% by year 7, 3% terminal). Margins expand from 15% EBITDA to 30%. WACC of 8%. Implied enterprise value: $55 billion. Conclusion: market is fairly priced.
Which is right? The divergence suggests either (a) the peer set has mispriced the company's growth relative to comps, or (b) your DCF is too optimistic on margin expansion or growth durability. To resolve, stress-test your DCF: what happens if growth decelerates faster? What if margins only reach 25%? What if WACC is 10% instead of 8%? If your DCF is robust, it is probably right. If it is fragile (small changes swing the valuation 20%), you should weight comps more heavily.
Real-world examples
Netflix early growth (2007–2012). Netflix was growing revenue 30–50% annually. Traditional media comps (Time Warner, News Corp) traded at 8–12x P/E. Netflix was unprofitable, so P/E was not meaningful. Comps suggested massive overvaluation. But a DCF recognizing that Netflix was building a durable, high-margin streaming business with structural growth would have justified much higher valuations. In this case, DCF was right; comps were wrong because they were not comparable.
Tesla (2020–2021). Tesla traded at an astronomical multiple—80x P/E or more—in 2020–2021 despite being a small player in automotive. Comps to traditional automakers (Ford, GM at 5–6x P/E) suggested Tesla was insanely overpriced. A DCF that assumed 40%+ revenue growth and eventual 15%+ net margins would have justified a premium, but 80x was optimistic even under bullish assumptions. In this case, comps undervalued the growth story, but Tesla was trading at a multiple that neither comps nor DCF could fully justify.
AT&T dividend (2015–2020). AT&T paid a high dividend yield (4–5%) and traded at 12–14x P/E, in line with other telecom peers. Comps said it was fairly valued. But a DCF recognizing that AT&T faced structural decline in legacy telecom business, high debt, and low growth would have highlighted that the yield was not attractive—investors were being paid for stability, but not adequately compensated for decay. In this case, comps were optimistic; DCF was more realistic.
Common mistakes
Picking the comps set to match your conclusion. You want a company to look cheap, so you build a peer set of faster-growing, higher-margin companies. Of course it looks cheap against those peers—but you have chosen the wrong comps. Always build the peer set first, before you care what the answer is.
Assuming terminal value in a DCF is destiny. A terminal value of 50% of total DCF value does not mean you are 50% uncertain. It means the bulk of value depends on long-term assumptions you cannot validate. Stress-test terminal value hard. Build scenarios.
Ignoring cycle and timing in comps. Comparing a cyclical company's P/E at the peak of the cycle to the five-year average peer P/E is misleading. You should compare to peer multiples at similar points in the cycle.
Treating DCF as gospel. Your DCF is your best guess given current information, not the truth. It is wrong until proven otherwise. Use comps as a reality check.
FAQ
When should I use comps and when DCF? Use comps first as a fast screening tool. Use DCF when you think the market has mispriced growth, margins, or when comps do not exist or are not comparable. Use both when the stakes are high.
Can I average a comps valuation and DCF valuation? Only if both are well-founded. If one is relying on a peer set you do not trust or a DCF with fragile assumptions, weighting both equally is lazy. Dig deeper.
What if comps and DCF differ by 20–30%? That is the margin of safety zone—explore why. Do not assume one is right and one is wrong. Stress-test both.
How do I adjust comps for a company that is growing much faster than peers? Use growth-adjusted multiples: EV/Revenue divided by growth rate, or a PEG ratio. Comps can work even for high-growth companies if you adjust for growth differences.
Is comps valuation the market price or fair value? It is the market price reflected in multiples. Fair value requires a judgment call: is the market being rational? Use comps as one input to that judgment.
Related concepts
- Relative valuation methodology — Understanding how multiples are calculated and when they are most reliable
- DCF for beginners — Deep dive into how terminal value and discount-rate assumptions drive DCF outputs
- Valuation ratios — P/E, EV/Revenue, and other multiples in historical and forward-looking context
- Business model analysis — Why companies with different business models cannot always be compared via simple multiples
Summary
Comps and DCF are complementary tools, each with distinct strengths. Comps are faster, grounded in real market prices, and useful for identifying relative mispricings within a sector. DCF forces explicit assumptions, allowing you to spot growth or margin inflections the market may be missing. Neither is universally superior. Mature, stable businesses are best valued via comps; high-growth or restructuring stories are best unlocked via DCF. The best investors layer both: they use comps to set a base case and reality-check, and DCF to test whether the market's embedded assumptions about growth and profitability are rational. When the two methods converge, you have high-confidence that your valuation is sound. When they diverge, that divergence is your signal to dig deeper—one of your two methods is seeing something real that the other is missing.