How to Walk Through a Comps Analysis in an Interview
"Walk me through a comps analysis" is one of the three questions, alongside the DCF and the LBO, that almost every investment banking interviewer reaches for at some point in the process. It tests something slightly different than the other two, though. A DCF rewards careful modeling discipline. Comps reward judgment: can you look at a messy set of public companies and decide, quickly and defensibly, which ones actually belong in the comparison. That judgment call is exactly what interviewers probe for, and it's where most candidates lose points even when they know the mechanics cold.
This guide covers the framework, a full worked example, and the peer-selection follow-ups that separate a candidate who memorized the steps from one who understands why the steps exist.
The Three-Step Comps Framework
A comparable company analysis, or "comps," values a company by looking at how the market is currently pricing similar public companies, then applying that pricing to the target. It breaks into three steps.
Step 1: Build the Peer Group
Start with companies in the same industry, then narrow the list using business model, growth rate, margin profile, and size. The goal is a tight group, usually somewhere between 5 and 10 companies, where every name earns its place. A peer group padded with loosely related companies produces a wider, less meaningful range of multiples.
Step 2: Calculate the Valuation Multiples
For each peer, calculate the relevant multiples. EV/EBITDA is the default for most industries because it's neutral to capital structure, so a company with a lot of debt and a company with almost none can still be compared on equal footing. P/E and EV/Sales get used in specific situations, which the next section covers.
Step 3: Apply the Multiples to the Target
Take the median (or sometimes the mean) multiple from the peer group and apply it to the target's own financials to get an implied enterprise value. From there, bridge to enterprise value and equity value and, if needed, an implied share price the same way you would in a DCF.
A Worked Example
Say you're valuing MidCo, a company with $600M of revenue and $120M of EBITDA. You've built a peer group of five public companies in the same sector, and their EV/EBITDA multiples come out to 7.6x, 8.1x, 8.5x, 8.9x, and 9.4x.
The median multiple is 8.5x. Apply that to MidCo's EBITDA: $120M times 8.5x gives an implied enterprise value of $1.02B.
From there, bridge to equity value the same way you would with any enterprise value figure. MidCo carries $150M of net debt, so subtracting that leaves an implied equity value of $870M. Divide by MidCo's 40M diluted shares outstanding and you get an implied share price of roughly $21.75.
Worth noting: the output here is a range, not a single number. A careful answer would also show what the implied value looks like at the low end (7.6x) and high end (9.4x) of the peer set, since that range, not just the median, is what tells you how much confidence to put in the result.
How to Defend Your Peer Group Under Pushback
Peer selection is the step interviewers probe hardest, because it's the one place in the whole exercise that requires judgment rather than arithmetic. If you select the wrong peers, every multiple downstream is built on a bad foundation, no matter how clean your math is.
A few dimensions matter more than people expect going in:
Business model, not just industry. Two companies can sit in the same sector and still not belong in the same peer group if one sells subscriptions and the other sells one-time hardware. Their margin structures and growth profiles will diverge for reasons that have nothing to do with how the market should value them relative to each other.
Growth rate and margin profile. A high-growth, high-margin company will trade at a premium multiple to a slower-growing, lower-margin one in the same industry, and that premium is deserved. Lumping both into one peer group and taking a median muddies the comparison rather than clarifying it.
Size. A $200M-revenue company and a $20B-revenue company in the same industry often trade at different multiples because of liquidity, access to capital, and scale advantages that are real, not noise.
The strongest version of this answer isn't a candidate who lists more criteria. It's one who can explain, out loud, why they excluded a company that looked like an obvious fit on the surface: "I left Company X out of the peer group despite similar revenue, because its margin profile is meaningfully different given its distribution-heavy business model." That kind of specific, defensible exclusion is what interviewers are actually listening for.
Which Multiple to Use
EV/EBITDA is the workhorse multiple for most industries, but it isn't universal. Financial institutions, where the balance sheet itself is the business, get valued on P/E or price-to-book instead, since EV/EBITDA doesn't translate cleanly when debt is a raw material rather than a financing choice. High-growth companies that aren't yet profitable, especially in software, often get valued on EV/Sales because EBITDA is negative or too small to be a meaningful denominator.
There's also a trailing-versus-forward decision buried in every comps analysis: do you use the last twelve months of financials or the next twelve months of consensus estimates? The two can produce noticeably different multiples for the same company, especially when growth is accelerating or decelerating, and LTM vs. NTM is worth knowing cold if you want the full picture on how that choice gets made.
Comps vs. the Other Valuation Methods
Interviewers like to ask how comps fit alongside DCF and precedent transactions, since the three methods are usually triangulated together rather than used in isolation.
A DCF is intrinsic. It values a company based on its own projected cash flows, independent of what the market happens to think today, which makes it useful when you suspect the market is mispricing a sector but also means the output is only as good as the assumptions feeding it. Comps, by contrast, are entirely market-based: they tell you what the market is paying for similar businesses right now, for better or worse. If the whole sector is overheated or depressed, comps will reflect that distortion rather than correct for it.
Precedent transactions are also market-based, but they capture what acquirers have actually paid to take control of similar companies, which typically runs higher than where those same companies trade as minority stakes on the public market. A strong answer names that distinction without being asked: comps tell you where a stock trades, precedents tell you what a buyer would pay for the whole thing.
Common Follow-Ups and Pitfalls
"What do you do if there aren't good public comps?" Widen the criteria along the dimension that matters least for the specific question you're answering, look at adjacent sub-sectors, or lean more heavily on DCF and precedent transactions to fill the gap. Say this directly rather than forcing a bad peer group to work.
"Your comps and your DCF disagree. Which do you trust?" Neither automatically. A strong answer explains why they might diverge, maybe the market is currently discounting the sector, or the DCF's growth assumptions are too optimistic, rather than picking one number and ignoring the other.
"What's a weakness of comps?" The output is a snapshot of current market sentiment, not a company's intrinsic worth, so it can be just as wrong as the market is at any given moment. A thin peer group also makes the result noisy: three or four companies is a much shakier median than ten.
The pitfall that costs candidates the most points isn't a conceptual one. It's forgetting to make comparability adjustments before calculating multiples in the first place, the same kind of EBITDA add-backs and normalization that go into cleaning up EBITDA for any valuation purpose. A peer's reported EBITDA that still includes a one-time litigation charge or a non-recurring gain will distort its multiple, and skipping that cleanup step is one of the fastest ways to undercut an otherwise solid answer.
Practicing the Full Walkthrough
The comps question rewards the same preparation as the DCF and LBO questions: know the three-step framework cold, but spend more of your practice time on peer selection and multiple choice, since that's where the real judgment gets tested.
Offcycle has comps mechanics built into structured flashcards alongside the DCF, LBO, and accounting concepts that tend to come up in the same conversation, plus custom practice sets you can build around whatever you're weakest on and a readiness score that tracks your progress across valuation, accounting, and M&A topics so you know exactly where you stand before you walk into the room.
The trial is free for 7 days and doesn't require a card to start.