Why Valuation Is Art, Not Science

Introduction

After working through the preceding sections of this guide, covering comparable companies, precedent transactions, DCF analysis, LBO models, M&A valuation, and sector-specific approaches, it would be tempting to conclude that valuation is a mechanical exercise: input the right data, apply the right formulas, and the "correct" answer emerges. This conclusion would be deeply wrong.

Valuation is fundamentally an exercise in informed judgment, not computation. The formulas provide a framework, but the output depends entirely on the assumptions fed into them. And those assumptions, from the revenue growth rate to the peer group selection to the terminal growth rate, involve judgment calls that reasonable analysts can (and do) disagree about. Experienced practitioners say valuation is "75% art and 25% science." The 25% science (the formulas, the frameworks) can be learned in months. The 75% art (the judgment) takes years to develop.

Each of the core methodologies has at least one critical input that is fundamentally subjective:

If the market always reflected "true" value, there would be no need for valuation. The market price is one data point, but it reflects sentiment, positioning, and short-term dynamics alongside fundamentals. The 2021 technology rally (when unprofitable companies traded at 30-50x revenue) and the subsequent 2022 correction (when the same companies lost 60-80% of their value while their businesses barely changed) demonstrate that market prices can diverge dramatically from intrinsic value for extended periods.

This divergence is precisely why DCF analysis exists: to provide a fundamental anchor independent of market sentiment. But the DCF itself depends on assumptions that the analyst must choose, which brings the judgment question full circle.

1. Assumption Selection

The most obvious form of judgment: choosing the growth rate, the WACC, the terminal value, the peer group. Each choice is defensible within a range, and the analyst's job is to select assumptions that are reasonable, documented, and internally consistent, not "correct" in some absolute sense.

2. Methodology Weighting

Which methodology deserves the most weight? In an M&A sell-side process, precedent transactions may carry the most weight because they directly reflect acquisition pricing. In a standalone valuation, the DCF may carry more weight because it is independent of market distortions. In a bank valuation, P/TBV and the DDM dominate because standard enterprise value metrics do not apply. The weighting decision is itself a judgment call that depends on the context.

3. Interpretation of Divergence

When methodologies produce different answers (which they almost always do), the analyst must interpret the divergence. If the DCF produces a higher value than comps, is the market undervaluing the company, or are the DCF projections too optimistic? If precedent transactions show a much higher value than comps, is the control premium justified, or were the precedent deals driven by unique circumstances? These interpretive questions require commercial judgment that no formula can provide.

4. Communication and Persuasion

Valuation is not just an analytical exercise; it is a communication exercise. The analyst must present the range in a way that supports the strategic recommendation, responds to counterarguments, and helps the client make an informed decision.

Valuation Range

The span between the lowest and highest defensible values produced by the combined valuation methodologies. The range, not a single point, is the true output of a valuation analysis. It reflects the inherent uncertainty in every assumption and the legitimate differences in perspectives across methodologies. The football field chart visually presents this range, showing where methodologies converge (high confidence) and where they diverge (areas requiring judgment). Presenting a single-point valuation without a range implies false precision and is considered poor practice.

In valuation, the standard is not correctness (which is unknowable for forward-looking assumptions) but defensibility: can the assumption be explained, justified, and withstood scrutiny from a senior banker, a client, a counterparty, or in litigation?

Defensible Assumption

A valuation input that can be supported with evidence, reasoning, and consistency. A defensible revenue growth assumption is one that is grounded in historical trends, management guidance, industry data, and/or comparable company growth rates. A non-defensible assumption is one that has no evidentiary support or contradicts available data. For example, projecting 20% revenue growth for a company that has grown 3-5% historically, without citing a specific catalyst (new product launch, market expansion, acquisition), is not defensible. The same 20% growth rate supported by a signed contract, a product launch timeline, and comparable company precedent IS defensible. In a fairness opinion context, every assumption may be examined in litigation, where "defensible" becomes a legal standard: could a reasonable financial professional, using the same data, arrive at this assumption?

The defensibility standard has practical implications for how valuations are built. Analysts do not simply choose the assumption that produces the "right" answer. They choose assumptions within a defensible range and then present the sensitivity analysis showing how the output changes across that range. The base case reflects the analyst's best judgment; the sensitivity table shows the full range of defensible outcomes.

This is why documentation matters so much in investment banking models. Every WACC input is sourced ("Beta: 1.15, Bloomberg adjusted beta as of 03/18/2026"). Every peer group inclusion and exclusion is justified. Every projection assumption has a stated basis (historical trend, management guidance, industry forecast). An undocumented assumption is an indefensible assumption, regardless of how reasonable the number might be.

The fairness opinion crystallizes the judgment nature of valuation into a formal legal document. The bank opines that the consideration in a proposed transaction is "fair, from a financial point of view." This opinion does not say the price is "correct" or "optimal" or "the best possible outcome." It says the price falls within the range of outcomes that a reasonable analysis would support.

This standard, "within a defensible range," is the practical definition of what valuation judgment means. A deal at the 25th percentile of the football field may be fair (within the range) even though it is at the lower end. A deal at the 90th percentile may also be fair but raises questions about whether the buyer is overpaying. The fairness opinion committee exercises judgment about whether the specific price, in the specific context of this deal, is reasonable.

Courts that review fairness opinions evaluate the process (was the analysis thorough? were multiple methodologies used? were the assumptions documented and reasonable?) more than the output (was the specific number correct?). This judicial focus on process over outcome reinforces the fundamental point: valuation is about the quality of the judgment, not the precision of the answer.

The Sell-Side Advisor's Judgment

On a sell-side M&A engagement, the banker must set the valuation range that frames the entire sale process. Set it too high and the process attracts no bids. Set it too low and the seller leaves money on the table. The range is derived from the analytical methodologies, but the final recommendation requires judgment about the buyer universe, the current M&A environment, the target's competitive positioning, and how the auction process will unfold.

The most experienced MDs calibrate the range based on factors that do not appear in any model: their personal relationships with potential buyers, their read on the credit markets (will financing be available at the leverage levels assumed in the LBO?), and their assessment of the CEO's credibility with buyers. These qualitative inputs affect the valuation as much as the quantitative analysis.

The Buy-Side Advisor's Judgment

On a buy-side engagement, the judgment is about the maximum defensible price. The analytical framework produces an implied value range, but the decision about how much to offer involves weighing the synergy value (how much to share with the target through the premium), the competitive dynamics (is another bidder likely to top the offer?), the strategic urgency (how much does the acquirer need this target?), and the financial capacity (can the acquirer fund the price without unacceptable credit deterioration?).

The Valuation Gap Problem

One of the most common dynamics in M&A is the valuation gap: the difference between what sellers expect and what buyers offer. In the 2022-2023 period, this gap widened dramatically as rates rose: sellers anchored on 2021-era valuations while buyers repriced to reflect the new cost of capital. Global M&A volume dropped approximately 35-40% as deals could not close because the two sides could not agree on price. By 2024-2025, the gap narrowed as both sides adjusted expectations, and deal volume recovered (up 36% in 2025 versus 2024). The valuation gap is not resolved by better models; it is resolved by time, market convergence, and the banker's skill at managing expectations on both sides.

Investment bankers are advisors, not judges. They represent a client and their analysis serves the client's objectives. This creates an inherent tension with objectivity.

On a sell-side engagement, the banker presents the valuation in the most favorable defensible light: selecting peer groups that produce higher multiples, emphasizing DCF scenarios with optimistic assumptions, and pointing to precedent transactions with the highest premiums. This is advocacy, not dishonesty. Every number must be supportable, but the framing serves the client's goal.

On a buy-side engagement, the banker does the reverse: selecting lower-multiple peers, using conservative DCF assumptions, and highlighting precedent transactions with lower premiums.

Valuation judgment is susceptible to cognitive biases that experienced practitioners learn to recognize:

Anchoring. The first number mentioned in a negotiation tends to anchor all subsequent discussion. If the sell-side banker presents a football field with a midpoint of $52, subsequent negotiation gravitates toward that number, even if the buyer's analysis suggests $45. This is why the initial valuation presentation is so consequential.

Confirmation bias. Analysts tend to seek evidence that supports their initial view. A banker who believes the company is worth $50 per share may unconsciously select assumptions that produce $50 rather than objectively evaluating the full range. The requirement to present multiple methodologies and sensitivities is the primary defense against confirmation bias.

Overconfidence in projections. Management teams systematically overestimate future performance. A study of sell-side analyst forecasts found that consensus revenue estimates overstate actual results by an average of 5-8% for high-growth companies and 2-3% for mature companies, with the error increasing for longer forecast horizons. The analyst building the DCF from management projections (which are typically even more optimistic than consensus) must calibrate for this bias by comparing management forecasts to historical accuracy and to consensus estimates. If management projected 12% growth last year and delivered 8%, their current projection of 15% should be viewed skeptically.

Sunk cost fallacy in deal advisory. After months of work on a sell-side engagement, there is psychological pressure to find a way to make the deal work, even if the bids are below the valuation range. The banker may be tempted to revise the analysis downward to make a low bid appear more acceptable. Maintaining the original analytical framework as an independent benchmark, separate from the advocacy presentation, helps resist this pressure.

Valuation is not a one-time exercise. It is a living analysis that evolves as new information becomes available and the deal process unfolds.

In the pitch phase, the valuation is built from public data and broad assumptions. It is directional, showing the board the likely range. In the engagement phase, the bank gains access to management projections and internal data, refining the analysis significantly. During the process phase, as bids arrive, the valuation becomes a benchmarking tool: each bid is compared to the range. In the closing phase, the valuation supports specific pricing arguments in the final negotiation.

The formulas do not change during this evolution, but the purpose of the analysis, the audience, and the context all shift, requiring the banker to adapt the same toolkit to different objectives at each stage.

This evolution also means the valuation is never "done." A model built in January for the pitch may be updated in March when management provides internal projections, refreshed in May when a peer company reports earnings that shift the comps, and revised again in July when market conditions change the WACC. Each update requires the analyst to re-exercise judgment about whether the new information warrants changing the base case or simply widening the sensitivity range. The model is a living document, and the judgment embedded in it is continuously tested against new data.

Valuation judgment is not taught; it is developed through accumulated experience across many deals and market cycles. A first-year analyst learns the mechanics. A second-year analyst learns to spot errors and calibrate assumptions. An associate learns to interpret divergence and present persuasively. A VP develops commercial awareness of buyer behavior and deal dynamics. An MD synthesizes decades of deal experience into pattern recognition that allows them to assess a deal's feasibility in minutes.

This progression explains the "75% art" assessment. The science (formulas, frameworks, calculations) can be mastered quickly. The art (judgment about assumptions, interpretation of divergence, contextual weighting, commercial awareness) takes years and is the scarce skill that makes experienced investment bankers valuable. The best valuation analysts are not the ones who build the most sophisticated models. They are the ones who ask the right questions about the assumptions inside those models.