Why do valuation conventions differ by sector? Match the right multiple to each of: a pre-profit SaaS company, a mature industrial, a commercial bank, an airline, and an E&P energy company. Explain why EV/EBITDA is wrong for the bank.

What interviewers are actually testing

Sector questions probe whether you can leave the generic valuation script behind and think inside an industry's own economics. They come as "what makes a good [sector] business?", "how would you value a [bank/airline/software company]?", "what metrics would you look at?", or "pitch me a company in an industry you follow." The interviewer is testing three things.

First, do you match the numerator and multiple to the economics? The single fastest tell of genuine fluency vs memorization is knowing you value a bank on P/B and P/E (not EV/EBITDA), a pre-profit SaaS company on EV/Revenue (not P/E), an airline on EV/EBITDAR (rent-adjusted). Explain why: for a bank, debt is raw material so enterprise value is meaningless; for early software there are no earnings yet so you value the top line and imply margins.

Second, do you know the KPIs that actually run the business? Not a laundry list: the two or three numbers management obsesses over. Software: net revenue retention and CAC payback. Banks: NIM, ROE, efficiency ratio, CET1. Retail: same-store sales and inventory turns. Airlines: RASM, CASM, load factor. Energy: production, reserves, lifting costs. Naming NRR and explaining that >100% means the base expands even with zero new customers signals real understanding.

Third, can you connect model → driver → risk → valuation? "Airlines have high fixed costs and a commodity fuel input, so tiny changes in load factor or fuel swing margins violently: that operating leverage plus cyclicality is why they trade at low multiples and why the balance sheet matters as much as the P&L." That chain (economics to risk to why the multiple is what it is) is what sounds like an insider.

The meta-signal: pick ONE sector you can genuinely go deep on (ideally tied to the group you're interviewing with) and be able to name a real company, its KPIs, and roughly how it's valued. Depth in one sector beats a shallow tour of five. And when you don't know a sector, apply the five-question universal framework out loud: that shows you have a system, not just memorized facts.

Common traps

Trap 1: Using EV/EBITDA on a bank or insurer. For financials, debt is the raw material of the business, not a financing choice, so enterprise value is meaningless. Value the equity directly with P/E and P/B.

Say it out loud: "You can't use EV/EBITDA on a bank: its debt is raw material, not capital structure, so enterprise value has no meaning. You value the equity directly: P/B and P/E, where P/B only exceeds 1x when ROE beats the cost of equity."

Trap 2: Putting a P/E on a pre-profit growth company. Early software and biotech have no meaningful earnings, so P/E is undefined or absurd. Value the top line on EV/Revenue and imply future margins.

Say it out loud: "A pre-profit SaaS company has no earnings to put a P/E on, so you value it on EV/Revenue or EV/ARR and key the multiple off growth, net retention, and the margin the business will reach at scale."

Trap 3: Reciting a KPI list without knowing what each means. Naming 'net revenue retention' is worthless if you can't say >100% means the existing base expands faster than it churns. Interviewers immediately follow up on any metric you name.

Say it out loud: "Net revenue retention above 100% means existing customers spend more each year than churn takes away, so revenue grows before you sell a single new logo: that's why it's the metric that separates a great SaaS business from a mediocre one."

Trap 4: Ignoring rent/leases for asset-light-by-lease businesses. Airlines, retailers, and restaurants lease heavily, so EBITDA is distorted by whether they lease or own. Use EV/EBITDAR to compare like-for-like.

Say it out loud: "Two airlines with identical operations can have very different EBITDA if one leases its fleet and one owns it, so you add rent back and use EV/EBITDAR, otherwise you're penalizing the one that leases and the comparison is apples to oranges."

Trap 5: Treating EBITDA as cash flow in capital-heavy sectors. In energy, telecom, and industrials, maintenance capex is enormous, so EBITDA massively overstates cash generation. Use EBITDA less capex, or the sector's specific numerator (EV/EBITDAX, per-flowing-barrel).

Say it out loud: "In capital-heavy sectors EBITDA overstates cash because you have to spend huge maintenance capex just to stand still, so I'd look at EBITDA minus capex or free cash flow, and for E&P use EV/EBITDAX or EV per flowing barrel and a NAV off the reserves."

Trap 6: Giving a generic 'DCF and comps' answer to a sector question. When asked what makes a good software business, answering 'strong cash flows and a good management team' shows zero fluency. Lead with the model, the defining KPI, and the value driver.

Say it out loud: "What makes software special isn't just cash flow: it's recurring revenue with net retention above 100% and near-zero marginal cost, which gives huge operating leverage as you scale, so incremental revenue drops almost entirely to margin. That's the specific reason the model is so valuable."