# Comparable Company Analysis: Excel Guide

*Alex Tapio · 2026-06-11 · 13 min · Model Deep-Dives*

Canonical: https://finamodel.com/blog/comparable-company-analysis

Learn comparable company analysis (trading comps) step by step: select a defensible peer set, build EV/EBITDA and EV/Revenue multiples, handle outliers, and turn median peer multiples into an implied valuation range, with a full worked example in Excel.

**Comparable company analysis (trading comps) values a business by comparing it to similar public companies. Instead of forecasting cash flows like a DCF, you observe what the market actually pays for peers - expressed as multiples such as EV/EBITDA and EV/Revenue - and apply those multiples to your target's financials. This guide walks through the full trading comps process in Excel: selecting a defensible peer set, building the multiples table, handling outliers, and converting median peer multiples into a credible implied valuation range, with a fully worked example.**

Comparable company analysis is the workhorse of relative valuation. Where a discounted cash flow model builds value from the bottom up, a comps analysis reads value off the market: if investors pay 30x EBITDA for a basket of similar businesses today, that tells you something concrete about what your target is worth right now. It is fast, market-grounded, and - crucially - easy for a buyer, board, or banker to sanity-check.

It is also easy to get wrong. The entire analysis lives or dies on peer selection, on calculating enterprise value consistently, and on resisting the temptation to cherry-pick the multiples that flatter your number. Done well, trading comps give you a defensible valuation range in an afternoon. Done badly, they give you false precision dressed up as market evidence.

```mermaid
flowchart TD
    A["Select Peer Set: screen by industry, size, growth, geography"] --> B["Gather Data: market cap, net debt, revenue, EBITDA, EBIT, EPS"]
    B --> C["Build Enterprise Value: market cap plus net debt"]
    C --> D["Compute Multiples: EV/Revenue, EV/EBITDA, EV/EBIT, P/E"]
    D --> E["Summary Statistics: min, 25th, median, 75th, max"]
    E --> F["Apply to Target: median multiple times target metric"]
    F --> G["Implied Valuation: equity value and per-share range"]
```

*The trading comps workflow: from peer selection to an implied valuation range.*

---

## What Is Comparable Company Analysis?

Comparable company analysis - "trading comps" or simply "comps" - estimates a company's value by comparing it to publicly traded peers. The logic is straightforward: similar companies should trade at similar multiples of their financial metrics. If a set of comparable software businesses trades at a median of 30x EV/EBITDA, and your target generates $77M of EBITDA, then the market is implying a valuation in the neighbourhood of $2.3bn for an equivalent business.

The multiples bridge the gap between companies of different sizes. A peer doing $600M of revenue and a target doing $350M are not directly comparable in absolute dollars, but their *valuation per dollar of revenue* (EV/Revenue) is. That normalisation is the entire trick: convert each peer's value into a ratio, take the central tendency of those ratios, and apply it to your target.

Trading comps are one of the three core valuation pillars, alongside **precedent transactions** (what acquirers actually paid for similar companies) and the **DCF** (intrinsic value from projected cash flows). Bankers almost never rely on a single method - they triangulate all three into the "football field" valuation range that anchors a fairness opinion or pitch. For the income- and asset-based methods that sit alongside comps, see our guide to [business valuation methods](/blog/business-valuation-methods).

---

## Why and When to Use Trading Comps

Comps shine when you have a clean set of public peers and a snapshot of current market sentiment is what you need:

- **IPO pricing and equity research** - what should this company trade at relative to its sector?
- **M&A** - a quick, market-grounded read on a target before committing to a full DCF.
- **Fairness opinions** - board-level evidence that a deal price is reasonable versus the market.
- **Private company valuation** - apply public peer multiples (often with a discount for illiquidity) to a private target.

They are weaker when peers are scarce, when the sector is mispriced (a bubble inflates every multiple equally), or when the target is structurally different from anything public. In those cases lean harder on the DCF. The two methods are complements, not rivals - comps tell you what the market pays *today*, a DCF tells you what the cash flows are *worth*. For a side-by-side of the major model types, see [DCF vs LBO vs 3-statement](/blog/dcf-vs-lbo-vs-3-statement).

---

## The Trading Comps Workflow

A clean comps model in Excel has a predictable structure:

1. **Peer Set** - the universe of comparable companies and the screening rationale.
2. **Raw Data** - share price, share count, debt, cash, and the financial metrics (revenue, EBITDA, EBIT, net income) for each peer.
3. **Enterprise Value Bridge** - market cap plus net debt for each peer.
4. **Multiples** - EV/Revenue, EV/EBITDA, EV/EBIT, and P/E.
5. **Summary Statistics** - min, 25th percentile, median, mean, 75th percentile, max.
6. **Implied Valuation** - apply the chosen statistic to the target and bridge to equity value.

The discipline is the same as any good model: inputs in one place, calculations referencing them, no hardcoded multiples buried in a formula. That is what makes the analysis auditable when a counterparty pushes back on your peer set - and they always do.

---

## Step 1 - Select a Defensible Peer Set

Peer selection is the single most important decision in the entire analysis. A weak peer set produces a precise-looking but meaningless number. Screen on the dimensions that actually drive valuation:

- **Industry and business model** - same sector, same revenue model (e.g. subscription SaaS, not licence-and-services).
- **Size** - companies within a reasonable band of revenue or market cap. A $50bn mega-cap and a $500m mid-cap rarely trade alike.
- **Growth profile** - high-growth companies command higher multiples. Pairing a 30%-grower with a 5%-grower distorts the median.
- **Profitability and margins** - EBITDA margins and capital intensity should be broadly comparable.
- **Geography** - different markets carry different risk premia and tax regimes.

Growth is the dimension analysts most often eyeball wrong. Before you finalise the set, compute each candidate's revenue CAGR over the last few years and make sure the cohort is genuinely similar - a peer growing twice as fast as the rest will drag the median multiple up and inflate your valuation. Use the calculator below to pin down each peer's growth rate before you let it into the set:

<!-- tool:cagr-calculator -->

Document *why* each company is in the set. When a buyer's advisor challenges your comps - and they will - "I included it because it screened on SIC code" is not a defence. "Same end market, same subscription model, revenue within 0.5x–2x of the target, similar growth" is.

---

## Step 2 - Build the Enterprise Value Bridge

Multiples come in two flavours, and mixing them is the most common comps error. **Enterprise value (EV)** multiples (EV/Revenue, EV/EBITDA, EV/EBIT) are capital-structure neutral - they value the whole firm. **Equity value** multiples (P/E) value only the shareholders' stake. The numerator and denominator must always match: EV pairs with pre-interest metrics (EBITDA, EBIT, revenue); equity value pairs with post-interest metrics (net income, EPS).

Enterprise value starts from market capitalisation and adds net debt:

```
Market Cap     = Share Price x Fully Diluted Shares
Net Debt       = Total Debt - Cash and Equivalents
Enterprise Value = Market Cap + Net Debt
```

```excel
// Market capitalisation
= SharePrice * DilutedShares

// Net debt (negative when the company is net cash)
= TotalDebt - CashAndEquivalents

// Enterprise value
= MarketCap + NetDebt
```

A net-cash company (more cash than debt) has an EV *below* its market cap, because an acquirer effectively gets that cash back. Watch the sign: in the worked example below, two peers carry negative net debt and their EV is lower than their equity value.

For a complete walkthrough of why EV and equity value differ and how to move between them, our [DCF model tutorial](/blog/dcf-model-excel-tutorial) covers the same bridge from the intrinsic-value side.

---

## Step 3 - Calculate the Valuation Multiples

With EV in hand, each multiple is a single division. The four workhorses:

```
EV / Revenue   = Enterprise Value / Revenue
EV / EBITDA    = Enterprise Value / EBITDA
EV / EBIT      = Enterprise Value / EBIT
P / E          = Market Cap / Net Income   (equivalently, Share Price / EPS)
```

```excel
// Enterprise value multiples
= EV / Revenue
= EV / EBITDA
= EV / EBIT

// Equity multiple (price / earnings)
= MarketCap / NetIncome
```

Which multiple matters most depends on the business. **EV/EBITDA** is the default for most mature, profitable companies because it strips out capital structure, tax, and depreciation policy. **EV/Revenue** is the go-to for high-growth or pre-profit companies where EBITDA is thin or negative. **P/E** is common for banks and stable, mature businesses, but it is sensitive to leverage and one-off items. Best practice is to show several and lead with the one most relevant to the sector.

---

## Worked Example: A Six-Company SaaS Comp Set

Suppose we are valuing **TargetCo**, a private B2B SaaS business with $350M of revenue, $77M of EBITDA (a 22% margin), $56M of EBIT, and $42M of net income. We assemble six comparable public software companies and pull their data. All figures are in $M.

| Peer | Market Cap | Net Debt | Enterprise Value | Revenue | EBITDA | EBIT | Net Income |
| :--- | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
| NimbusSoft | 5,400 | (200) | 5,200 | 600 | 168 | 138 | 120 |
| DataForge | 3,300 | 300 | 3,600 | 480 | 120 | 96 | 72 |
| GridLogic | 2,200 | 150 | 2,350 | 380 | 84 | 62 | 48 |
| Streamline | 1,500 | 250 | 1,750 | 300 | 60 | 42 | 30 |
| ApexAnalytics | 6,200 | (100) | 6,100 | 620 | 198 | 168 | 150 |
| VectorWorks | 2,600 | 400 | 3,000 | 420 | 92 | 68 | 52 |

Dividing EV by each metric (and market cap by net income for P/E) gives the multiples table:

| Peer | EV/Revenue | EV/EBITDA | EV/EBIT | P/E |
| :--- | ---: | ---: | ---: | ---: |
| NimbusSoft | 8.7x | 31.0x | 37.7x | 45.0x |
| DataForge | 7.5x | 30.0x | 37.5x | 45.8x |
| GridLogic | 6.2x | 28.0x | 37.9x | 45.8x |
| Streamline | 5.8x | 29.2x | 41.7x | 50.0x |
| ApexAnalytics | 9.8x | 30.8x | 36.3x | 41.3x |
| VectorWorks | 7.1x | 32.6x | 44.1x | 50.0x |

A quick read: EV/EBITDA is tightly clustered between 28x and 33x, which is exactly what you want - it signals a coherent peer set. EV/EBIT and EV/Revenue are wider, reflecting different depreciation policies and margin profiles across the group.

---

## Step 4 - Handle Outliers and Choose the Summary Statistic

Now collapse each column into summary statistics. **Use the median, not the mean.** Multiples are routinely skewed by one or two extreme observations, and the mean is dragged toward them; the median is robust to outliers and is the market standard for comps.

```excel
// Central tendency (prefer median)
= MEDIAN(EV_EBITDA_range)
= AVERAGE(EV_EBITDA_range)

// Dispersion for the valuation range
= QUARTILE.INC(EV_EBITDA_range, 1)   // 25th percentile
= QUARTILE.INC(EV_EBITDA_range, 3)   // 75th percentile
= MIN(EV_EBITDA_range)
= MAX(EV_EBITDA_range)
```

Applied to our six peers:

| Statistic | EV/Revenue | EV/EBITDA | EV/EBIT | P/E |
| :--- | ---: | ---: | ---: | ---: |
| Minimum | 5.8x | 28.0x | 36.3x | 41.3x |
| 25th percentile | 6.1x | 28.9x | 37.2x | 44.1x |
| **Median** | **7.3x** | **30.4x** | **37.8x** | **45.8x** |
| Mean | 7.5x | 30.3x | 39.2x | 46.3x |
| 75th percentile | 9.0x | 31.4x | 42.3x | 50.0x |
| Maximum | 9.8x | 32.6x | 44.1x | 50.0x |

**Outlier discipline:** before applying these, scan for any peer that looks structurally different - a company mid-restructuring, one with a depressed or negative EBITDA, or one trading on takeover speculation. If a peer's multiple sits far outside the cluster for a *reason*, exclude it and note why; do not silently delete it. Here the group is clean enough to keep all six, and the 25th–75th percentile band gives us a defensible range rather than a single point.

---

## Step 5 - Apply the Multiples to the Target

The final step converts peer multiples into an implied value for TargetCo. For each EV multiple, multiply the median by the target's corresponding metric to get implied enterprise value, then subtract the target's net debt of $120M to reach equity value. For P/E, multiply the median by net income to get equity value directly.

```
Implied EV            = Median Multiple x Target Metric
Implied Equity Value  = Implied EV - Target Net Debt
Implied Share Price   = Implied Equity Value / Diluted Shares
```

Running each median multiple through TargetCo's financials:

| Multiple | Median | Target Metric | Implied EV | Less Net Debt | Implied Equity |
| :--- | ---: | ---: | ---: | ---: | ---: |
| EV/Revenue | 7.3x | $350M | $2,555M | ($120M) | $2,435M |
| EV/EBITDA | 30.4x | $77M | $2,341M | ($120M) | $2,221M |
| EV/EBIT | 37.8x | $56M | $2,117M | ($120M) | $1,997M |
| P/E | 45.8x | $42M (NI) | - | - | $1,924M |

The four methods cluster between roughly **$1.9bn and $2.4bn of equity value**. EV/EBITDA - our lead multiple for a profitable SaaS business - lands at **$2,221M**. With 25M fully diluted shares, that implies about **$88.83 per share**.

<!-- template:comps -->

### Framing the Range

Rather than quoting a single number, present a range driven by the EV/EBITDA quartiles - this is the comps equivalent of a DCF sensitivity table:

| Scenario | EV/EBITDA | Implied EV | Implied Equity | Per Share |
| :--- | ---: | ---: | ---: | ---: |
| Conservative (25th) | 28.9x | $2,225M | $2,105M | $84.21 |
| Base case (median) | 30.4x | $2,341M | $2,221M | $88.83 |
| Optimistic (75th) | 31.4x | $2,418M | $2,298M | $91.91 |

That **$84–$92 per share** band is the honest output of the analysis: a market-implied range, not false precision. You would then place it on a football field next to your DCF and precedent-transaction ranges and look for where the methods overlap.

---

## Trading Comps vs. Precedent Transactions vs. DCF

| Method | What it measures | Strengths | Weaknesses |
| :--- | :--- | :--- | :--- |
| **Trading comps** | What the market pays for similar *public* companies today | Fast, market-grounded, easy to sanity-check | No control premium; only as good as the peer set; reflects current sentiment |
| **Precedent transactions** | What acquirers *actually paid* for similar companies | Includes control premium; real deal evidence | Stale data; deal-specific synergies distort multiples; disclosure gaps |
| **DCF** | Intrinsic value from projected cash flows | Fundamentals-driven; not hostage to market mood | Highly assumption-sensitive; terminal value dominates |

Precedent transaction multiples are typically *higher* than trading comps because acquirers pay a control premium to take over a company. A common rookie error is comparing the two directly without adjusting for that premium. Use trading comps for "where would this trade as a public company" and precedent transactions for "what would someone pay to buy it outright."

---

## Common Mistakes to Avoid

1. **A sloppy peer set.** Including companies that merely share a sector code but differ wildly in size, growth, or model. This is the number-one way to produce a confident, wrong answer. Screen rigorously and document the rationale.
2. **Mismatched numerator and denominator.** Dividing enterprise value by net income, or market cap by EBITDA. EV pairs with pre-interest metrics; equity value pairs with post-interest metrics. Always.
3. **Forgetting the net-debt sign.** Treating net cash as if it added to enterprise value. A net-cash peer has EV *below* its market cap - get the sign wrong and every multiple is off.
4. **Using the mean instead of the median.** One outlier peer drags the mean and inflates your valuation. The median is the market standard precisely because it resists outliers.
5. **Mixing trailing and forward metrics.** Comparing one peer on last-twelve-months EBITDA and another on next-year EBITDA. Pick one basis (LTM or NTM) and apply it consistently across the whole set and the target.
6. **Ignoring non-recurring items.** Failing to normalise EBITDA for one-off legal settlements, restructuring charges, or stock-based compensation. Clean the numbers before you build the multiples.
7. **False precision.** Quoting "$88.83 per share" as *the* answer. Comps produce a range. Present the quartile band and triangulate against other methods - never a single decimal-point figure.

---

## Key Takeaways

- **Peer selection is everything.** The quality of a comps analysis is capped by the quality of its peer set. Screen on industry, size, growth, margins, and geography - and document why each company is in or out.
- **Keep multiples consistent.** Enterprise value multiples (EV/Revenue, EV/EBITDA, EV/EBIT) value the whole firm; P/E values equity only. Match pre-interest metrics with EV and post-interest metrics with equity value, every time.
- **Build EV correctly.** Enterprise value is market cap plus net debt. Net-cash companies have an EV below their market cap - mind the sign.
- **Lead with EV/EBITDA, but show several.** EV/EBITDA is the default for profitable businesses; EV/Revenue suits high-growth or pre-profit companies; P/E suits mature, stable ones. Present a spread and lead with the most relevant.
- **Use the median and quartiles.** The median resists outliers; the 25th–75th percentile band gives you a defensible valuation range instead of a single point.
- **Triangulate, don't isolate.** Comps tell you what the market pays today; a DCF tells you what the cash flows are worth; precedent transactions tell you what buyers actually paid. The credible answer is where the three overlap.

To put this into practice, download the [comparable company analysis template](/templates/comps) and pair it with our [DCF model tutorial](/blog/dcf-model-excel-tutorial) and [business valuation methods](/blog/business-valuation-methods) guide. For the cost-of-capital inputs that feed the DCF side of your football field, the [WACC calculator](/tools/wacc-calculator) and [CAGR calculator](/tools/cagr-calculator) are good starting points.


## Frequently asked questions

### What is comparable company analysis?

Comparable company analysis (trading comps) is a relative valuation method that values a business by comparing it to similar publicly traded companies. You calculate valuation multiples such as EV/EBITDA and EV/Revenue for a set of peers, take the median, and apply it to your target's financials to derive an implied valuation range. Unlike a DCF, which builds value from projected cash flows, comps read value directly off what the market currently pays for comparable businesses.

### What multiples are used in trading comps?

The four most common are EV/Revenue, EV/EBITDA, EV/EBIT, and P/E. Enterprise value multiples (EV/Revenue, EV/EBITDA, EV/EBIT) are capital-structure neutral and value the whole firm, so they pair with pre-interest metrics. P/E is an equity multiple and pairs with net income or EPS. EV/EBITDA is the default for mature, profitable companies; EV/Revenue suits high-growth or pre-profit businesses; P/E is common for banks and stable, mature companies.

### Why use the median instead of the mean for peer multiples?

Valuation multiples are frequently skewed by one or two extreme peers, and the mean is dragged toward those outliers. The median is robust to outliers and is the market standard for comps. Best practice is to report the median as the base case and use the 25th and 75th percentiles to frame a valuation range, rather than relying on a single average that an outlier can distort.

### How do you calculate enterprise value for a comp?

Enterprise value equals market capitalisation plus net debt, where market cap is share price times fully diluted shares and net debt is total debt minus cash and equivalents. A company with more cash than debt has negative net debt, so its enterprise value is below its market cap because an acquirer effectively receives that cash. Getting the net-debt sign right is essential, as every EV multiple depends on it.

### What is the difference between trading comps and precedent transactions?

Trading comps measure what the market pays for similar public companies today, with no control premium. Precedent transactions measure what acquirers actually paid to buy similar companies outright, which includes a control premium and often deal-specific synergies. Precedent transaction multiples are therefore usually higher than trading comps. Use trading comps for a public-market valuation and precedent transactions for an acquisition value.

### What are the most common mistakes in a comps analysis?

The biggest mistakes are: (1) a sloppy peer set with companies that differ in size, growth, or business model; (2) mismatching numerator and denominator, such as dividing enterprise value by net income; (3) getting the net-debt sign wrong; (4) using the mean instead of the median; (5) mixing trailing and forward metrics across peers; (6) failing to normalise EBITDA for one-off items; and (7) quoting a single precise number instead of a defensible range.