Private Equity Masterclass

Structure

Four divisions, two transaction types

Investment banks contain four main divisions: asset management, equity research, sales & trading, and investment banking itself. "Investment banking" as a job title refers specifically to the last of these — the advisory business.

Banks earn an advisory fee (a small percentage of transaction volume — on a $1bn deal, sub-1%, but that's still tens of millions) on two kinds of transaction:

Capital raises

Equity deals: IPOs (Initial Public Offerings), follow-on offerings (a public company issuing more shares), convertible bonds
Debt deals: bank debt, high-yield bonds, investment-grade bonds

M&A Mergers & Acquisitions

Sell-side: banks always prefer this — you're paid when the deal announces and closes
Buy-side: paid only if your client wins; often just a retainer if they don't, since one of several bidders always loses

Structure

Product groups vs. coverage groups

Banks organise their bankers one of two ways — sometimes both at once.

Product groups

One transaction type (e.g. M&A, ECM — Equity Capital Markets, DCM — Debt Capital Markets) across many industries
Deep expertise in a transaction, broad industry exposure
Drawback: pigeonholed exit options, especially in ECM/DCM. M&A product groups are the exception — exits are wide open

Coverage groups

One industry (e.g. healthcare, TMT), all transaction types
Deep view of an industry's landscape, exposure to varied deal types
Drawback: you won't enjoy every transaction type equally — heavier on presentation and pitch work when a deal doesn't materialise

Goldman Sachs runs almost entirely on coverage groups; JPMorgan keeps a separate M&A product group; elite boutiques like Evercore and Centerview are M&A-only by design. The split is bank-specific — check each one individually.

Why clients pay

Five ways banks add value

Maximise financial outcomes — the right price for equity, the right amount of debt, the highest price on a sale.
Process management — running due diligence, data rooms, and legal coordination that a company can't run on its own expertise.
Understanding market climate — knowing when to launch and when to shelve. An IPO can die in the week markets tank after months of prep.
Connecting companies — senior bankers hold the relationships (Michael Grimes at Morgan Stanley knowing Zuckerberg, Cook) that a $500m company can't get on its own.
Industry & product knowledge — pattern-matching a tricky situation against every similar deal the bank has seen before.

Reason 1, illustrated — competitive tension in a sell-side process

Banks contact buyers the client wouldn't have found or approached alone — bidding blind against each other pushes the price past what a direct sale would achieve, easily clearing the ~1–2% fee.

Where the jobs are

Bank tiers, and the ladder inside one

Tier	Focus	Typical deal size	Examples
Bulge bracket	All products, all industries, global	up to $50–100bn	Goldman Sachs, Morgan Stanley, JPMorgan, BofA
Elite boutique	Mostly M&A, non-capital-intensive	$1bn – 20bn+	Evercore, Centerview, Lazard
Middle market	All products, smaller-revenue clients	$0.5bn – a few bn	Jefferies, Piper Sandler, Raymond James
Boutique	One industry or function, small deals	sub-$500m	varies widely

The hierarchy — five rungs, and what actually changes at each one

Managing Director — reviews pitch books, brings the client relationships that generate the deals; the rainmaker role, worth 80%+ of the value chain.
Director / SVP / Principal — starts forming client relationships directly; barely touches models or decks anymore.
Vice President — present at client meetings, gives direction to analysts and associates, but not yet owning the relationship.
Associate — double-checks the analyst's work cell by cell, builds some pages and models directly.
Analyst — the workhorse; builds the model from scratch, 60–90% of the underlying work.

Value concentrates at the top, where an MD's client relationships are what actually bring deals in the door. ~$175k analyst-to-associate base, rising toward $600k–$1m total comp by VP — figures as stated in the source video (2025); comp varies by bank and moves year to year, so treat as directional, not a quote.

Statement 1 of 3 · over a period of time

The income statement

Definition

Reports a company's revenue, costs, expenses and profit or loss over a period of time.

The line-by-line walk a banker expects in an interview — memorise the sequence, not just the words:

Top-line to bottom-line — a $100 revenue base, walked down to net income

Every real income statement collapses to this shape — Apple splits COGS (Cost Of Goods Sold) by product vs. services, Coca-Cola keeps one SG&A (Selling, General & Administrative) line, but revenue → COGS → gross profit → OpEx → EBIT → interest & tax → net income never changes.

Three rules for what belongs on it

Must fall 100% within the period. Rent spanning 2025–2027 on a 2025 income statement only shows the 2025 slice.
Must affect the company's taxes (and so net income). Debt repayment and dividends flow straight to the balance sheet — they never touch the income statement.
Must be a revenue, expense, gain, or loss. Gains/losses are non-core: asset sales, lawsuit settlements, FX moves. Never debt repayment, dividends, or CapEx (capital expenditure) — none of those are gains or losses.

↓ Jump to a full worked income statement

The rule underneath everything

Accrual accounting

Definition

Measures performance based on when things happen, not when cash moves — the single most-tested concept in banking accounting interviews.

Two rules do all the work:

Revenue recognition — two mirror-image examples

Matching principle is the other half: COGS is recorded when the sale happens, not when the inventory was bought — $100k spent on water bottles sits off the income statement until the bottles actually sell.

Statement 2 of 3 · over a period of time

The cash flow statement

Cash is king

Revenue isn't cash. EBIT isn't cash. EBITDA (EBIT plus depreciation and amortization added back) isn't cash. Net income isn't cash. Only cash is cash — and cash generation is the true measure of a company's value (ask BlackBerry).

If you could only keep one statement, it's this one — it reconciles non-cash accounting items (AR, inventory) back to real cash, and captures flows the income statement never sees at all (CapEx, debt raised).

The three sections — and exactly what moves each one

Operating always starts from net income and adjusts for non-cash items and working-capital changes; sum all three sections and you get the net change in cash for the period. PP&E (Property, Plant & Equipment) is what CapEx typically buys, sitting in the investing section above.

What counts as a non-cash add-back?

Any expense that reduced net income without any actual cash leaving the business gets reversed here. The two you'll see in almost every example in this document are D&A (Depreciation & Amortization — spreading an asset's cost over its life) and stock-based compensation (paying employees in shares rather than cash) — both covered in full back in the income statement section above.

Real-world cash flow statements usually carry a few more. Impairment / write-downs — marking an asset's recorded value down on paper (say, goodwill from an acquisition that isn't performing), with no cash changing hands. Deferred taxes — the gap between the tax expense shown for accounting purposes and the actual cash tax paid this year, caused by timing differences between tax rules and accounting rules. Unrealized gains or losses — a paper gain on an investment still being held, or an FX translation swing — the same kind of item that sits in AOCI on the balance sheet.

The test is always the one from the income statement rules: did this actually cost, or bring in, real cash this period? If not, it gets added back (or subtracted, for a non-cash gain) here.

The add-back, made concrete — Harlow Fabrication buys a machine

The $500,000 never touches the income statement in one lump sum — Year 0's cash outlay shows up once, in investing activities. From Year 1 onward, depreciation quietly reduces net income every year with no cash actually moving, which is exactly why it has to be added back to get from "profit" to "cash."

↓ Jump to a full worked cash flow statement

Statement 3 of 3 · a snapshot in time

The balance sheet

The golden rule

Assets = Liabilities + Shareholders' Equity

Assets are what a company owns that generates future cash. Liabilities are what it owes. Shareholders' equity is what the owners keep once everything else is paid — it is both a source of funding and a running scorecard of value created.

Current (≤ 1 yr)		Non-current (> 1 yr)
Cash & equivalents, marketable securities, accounts receivable, inventory, prepaid expenses	Assets	Long-term investments, PP&E, intangibles, goodwill, deferred tax assets, equity-method investments
Revolver, current portion of LTD, accounts payable, accrued expenses, deferred revenue	Liabilities	Long-term debt, lease liabilities, deferred tax liabilities, pension liabilities

↓ Jump to a full worked balance sheet

The trickiest section of the balance sheet

Shareholders' equity, line by line

Built from four kinds of line item

Common stock is booked at par value (a nominal figure, e.g. $0.0001/share) and never changes with the market — the cash actually raised above that goes to additional paid-in capital instead.

Expand each line item for what the abbreviation stands for, the plain-language definition, and a quick worked number.

Common Stock at par value

The legal "stated" value of every share issued and outstanding. Par value is a deliberately tiny, arbitrary number — often $0.0001 or $0.01 a share — and it never moves, no matter what the stock actually trades for.

100 shares × $0.01 par = $1 common stock, regardless of what those shares were actually sold for.

APIC Additional Paid-In Capital

Also called surplus capital. The cash shareholders paid above par value when the shares were first issued — effectively "total cash raised, minus the par value sliver." Like common stock, it's frozen at the moment of issuance and doesn't move with the share price afterward.

100 shares sold at $10 each = $1,000 raised. Common stock = $1 (above). APIC = $1,000 − $1 = $999.

Preferred Stock

A separate class of ownership carrying a fixed dividend and priority over common shareholders if the company is liquidated. It gets its own par value and APIC lines when a company has it — most don't. The video's own estimate: roughly 1–5% of companies carry preferred stock at all.

Retained Earnings RE

The single most-tested line item in this whole section, and the direct thread connecting the income statement to the balance sheet: prior retained earnings + net income − dividends. A brand-new company starts at $0.

Year 1: $0 + $100 net income − $0 dividends = $100. Year 2: $100 + $200 net income = $300.

AOCI Accumulated Other Comprehensive Income

A catch-all for gains and losses the company hasn't actually "realized" through a sale yet — unrealized gains or losses on investments still being held, and currency-translation adjustments from converting an overseas subsidiary's books back into the parent's reporting currency.

Say a company holds $500k of bonds it bought at par, and rates move so they're now worth $520k on paper. It hasn't sold them — no cash has changed hands — so that +$20k can't go through net income. It sits in AOCI instead. Same logic if a UK subsidiary's £10m of assets are suddenly worth $12.6m instead of $12.4m purely because GBP/USD moved: the $200k swing lands in AOCI, not the income statement.

Treasury Stock

The cost of shares the company has bought back on the open market but not formally retired. It's a contra-equity account — it always shows up as a negative number, reducing total shareholders' equity.

NCI Non-Controlling Interest

The slice of a majority-owned subsidiary (>50% owned, but not 100%) that belongs to someone else. Own 75% of a company and you still have to consolidate 100% of its financials — the other 25% is isolated here as NCI. It comes back in a bigger way in Part III.

Seeing it in practice

One company, three statements, fully tied together

Synthetic figures — built to teach the connections, not a real filing

Everything above is easier to hold onto against one real-looking set of books. Meet Meridian Systems, Inc., a mid-size software company, FY2025 — every number below was chosen so the three statements actually tie to one another, the way the video keeps insisting they do.

Meridian Systems, Inc.

Income Statement · Year Ended December 31, 2025 · $ in millions

Revenue	1,200
Cost of goods sold	(360)
Gross profit	840
Research & development	(210)
Selling, general & administrative	(290)
Operating income (EBIT)	340
Other expense, net	(10)
Interest expense	(25)
Earnings before tax	305
Income tax (21%)	(64)
Net income	241

→

D&A ($85m) and stock comp ($45m) never get their own line on the income statement — they're mixed invisibly into the COGS, R&D, and SG&A totals above (which do have their own lines). That invisibility is exactly why the cash flow statement below has to pull them back out and add them back separately — it's undoing a cost that never actually cost the business any cash.

Meridian Systems, Inc.

Cash Flow Statement · Year Ended December 31, 2025 · $ in millions

Operating

Net income	241
+ Depreciation & amortization	85
+ Stock-based compensation	45
− Increase in accounts receivable	(25)
− Increase in inventory	(8)
− Increase in prepaid expenses	(5)
+ Increase in accounts payable	12
+ Increase in deferred revenue	18
Cash flow from operations	363

Investing

Purchases of marketable securities	(60)
Proceeds from maturities of securities	40
Capital expenditures	(95)
Cash flow from investing	(115)

Financing

Proceeds from issuance of debt	50
Repayment of debt	(35)
Dividends paid	(60)
Repurchase of common stock	(25)
Cash flow from financing	(70)

Net change in cash	178
Cash, beginning of year (plug: 420 − 178)	242
Cash, end of year	420

→

The full walk, in one line: $241m net income + $130m non-cash add-backs − $8m working-capital moves = $363m cash from operations → −$115m investing → −$70m financing = $178m net change in cash → + $242m already in the bank = $420m. That $420m — not the $241m — is what lands on the Balance Sheet's cash line below. Net income is where this statement starts, not the answer it produces.

Meridian Systems, Inc.

Balance Sheet · As of December 31, 2025 · $ in millions

Assets

Cash & equivalents	420
Marketable securities	150
Accounts receivable	180
Inventory	60
Prepaid expenses	25
Total current assets	835
PP&E, net	310
Goodwill	220
Intangible assets, net	95
Deferred tax assets	15
Total non-current assets	640
Total assets	1,475

Liabilities

Accounts payable	95
Accrued expenses	70
Deferred revenue	110
Current portion of LT debt	40
Total current liabilities	315
Long-term debt	380
Lease liabilities	35
Deferred tax liabilities	20
Total non-current liabilities	435
Total liabilities	750

Shareholders' Equity

Common stock & APIC	340
Retained earnings	405
AOCI	(5)
Treasury stock	(15)
Total shareholders' equity	725
Total liabilities + equity	1,475

→

Retained earnings of $405m = $224m brought forward + $241m net income − $60m dividends paid. Total assets ($1,475m) matches total liabilities + equity ($1,475m) exactly — the golden rule, holding.

The core distinction

The house analogy

A $500,000 house, funded half by down payment, half by mortgage

Selling the house means paying off the mortgage regardless of the price agreed — so the price quoted to a buyer is always the enterprise value, never just the equity slice.

Value to shareholders only

Equity value, properly

Definition

Value to only equity shareholders after all debt and other obligations are satisfied. In theory: market value of total assets − market value of total liabilities = market value of common shareholders' equity.

Three ways to calculate it — but for public companies, only one matters in an interview:

1 · Market value of net assets — theoretical 2 · Share price × diluted share count — public companies, use this 3 · Last funding round valuation — private companies

Worked example — Apple

$213/share × 14.84bn diluted shares = $3.16 trillion — i.e. the market cap.

The video quotes $3.17tn; the inputs as stated multiply to 3.161, so treat the final digit as rounding in the share price or share count, not a different method.

Equity value ≠ shareholders' equity

Shareholders' equity (book value) includes preferred stock and NCI; equity value doesn't. Book value is historical cost — what was actually paid, frozen on the day it happened. Equity value is forward-looking market opinion — what investors believe the company can generate in cash from here.

Worked example — book value vs. equity value

$500m buildings + $200m inventory + $100m patents − $200m debt = $600m book value. The company trades at $1,200m equity value.

Equity value ÷ book value = 2.0× — the market sees roughly double the value the accounting record shows, priced on growth and earning power the balance sheet can't capture.

Who gets paid first

Investor hierarchy

Different capital comes with different liquidation priority — the order investors are paid if a company sells everything and winds down.

$50m in cash, distributed against a $75m debt claim — nothing left for anyone else

Value to every capital provider

Enterprise value

Definition

The value of core business operations to all investors — capital-structure-neutral, and often described as a company's takeover price.

EV = Equity Value + Debt + Preferred Stock + Non-Controlling Interest − Cash

The bridge from equity value to enterprise value — worked example

All figures $m. $1,000 equity value + $100 preferred + $300 debt + $50 NCI − $150 cash = $1,300 enterprise value. Run it in reverse (given EV, solve for equity value) and the signs simply flip.

Reverse worked example

EV $7,000m, preferred $200m, debt $2,000m, NCI $300m, cash $3,000m.

Equity value = 7,000 − 200 − 2,000 − 300 + 3,000 = $7,500m

Non-operating items get stripped out because EV is meant to reflect only core operations: excess cash, marketable securities, minority stakes in unrelated businesses, assets held for sale, excess PP&E not used in operations.

Making multiples apples-to-apples

Non-controlling interest

Own >50% of a company and accounting rules force you to consolidate 100% of its financials into your own — even the sliver you don't own. That unowned sliver isn't cash raised or debt owed; it's a placeholder for value your own financials now include but don't actually own. NCI is how enterprise value accounts for it — added onto the bridge alongside debt and preferred stock, so that a multiple like EV/EBITDA compares like with like.

Purpose of EV/EBITDA

Its whole purpose is to let you compare how the market values two companies' actual operations, without either company's financing choices or accounting policies distorting the comparison. It does this because both halves of the ratio are deliberately built to be capital-structure-neutral: EV values the whole business regardless of how it's financed, and EBITDA strips out interest and tax the same way.

LargeCo owns 80% of SmallCo — building an apples-to-apples EV/EBITDA

EV/EBITDA = 620 / 60 = ~10.3×. Skip the NCI add-back and the numerator would only reflect 80% of SmallCo while the denominator counts 100% of it — the multiple would be quietly wrong.

▶ Study this in more depth — watch it in the source video (from 4:23:55)

Two families of method

Relative vs. intrinsic

Relative

Trading comps — what similar public companies trade at
Transaction comps — what similar companies sold for

Intrinsic

Discounted cash flow — what the business itself can generate
LBO (Leveraged Buyout), dividend discount model (the latter mostly for financial institutions)

There's no single correct valuation — only a more or less defensible one. Bankers triangulate across all of these into a football field: a range built from the 52-week trading price, analyst targets, trading comps, transaction comps, and the DCF stacked side by side.

Relative valuation, present tense

Trading comps

Interview answer

A relative valuation methodology comparing financial metrics across a set of similar public companies to determine a target's valuation.

Build the peer set
4–8 companies, similar on business model, industry, geography, size, and financial profile. More than ~15–20 stops helping — you're no longer comparing like with like.
Select, collect, calculate
Pick metrics that fit the company's stage (revenue multiples for growth, EBITDA/P·E for mature), pull the data, calculate the multiples.
Analyse & value
Choose a defensible multiple range from the peer set, apply it to the target's own metric.

Each of the three is worked through below. The weight is not evenly spread: steps 1 and 2 are where nearly all the effort and nearly all the judgement sit, and step 3 — the part that actually produces a valuation — is arithmetic.

Step 1

Build the peer set

Who counts as a comparable company is the first judgement call, and it constrains everything downstream — get the set wrong and the rest of the process will produce a precise answer to the wrong question.

Peer-set construction — Netflix, three plausible peer groups

Peer group	Companies	Why they qualify
Streaming	Disney, Comcast, Warner Bros. Discovery, Paramount	Direct streaming competitors, despite very different broader businesses (theme parks, cable)
Traditional media	AMC Networks, Lionsgate, Fox, Sony	Studios producing the same kind of content
Broader tech	Alphabet, Meta, Amazon, Spotify	Netflix is fundamentally a tech company at its core

Excluded deliberately: Apple (streaming is a rounding error of its business), Tesla and Microsoft (unrelated products), Tencent (similar model, wrong geography).

Step 2

Select, collect, calculate

Three separate jobs wearing one label. Only the first requires thought.

Select

Which metrics

Driven by the company's performance and its stage in the business life cycle
Is it profitable? If not, margin and profit multiples carry no information
This choice determines which multiples are even available to you

Collect

From where

Inside a bank: FactSet or Bloomberg, and the filings themselves for historicals
Outside one: SEC filings direct, BamSEC, Yahoo Finance — genuinely fine
Databases contain errors. Figures get checked back to the filing, not trusted on sight

Calculate

Then the easy part

Growth rates, margins, equity value, enterprise value, multiples
Mechanical once the data is in — arithmetic, not judgement
Drop any multiple that has gone meaningless (see below)

Note where the work actually is. Collecting accurate figures is the slow, manual part; calculating is trivial; and selecting is the part that needs a framework. That framework is the business life cycle.

The business life cycle — the frame for choosing metrics

Profit is negative through introduction and only crosses zero once the company scales — which is exactly why a pre-profit company cannot be valued on a profit multiple. Two honest caveats: real companies sit between stages rather than inside one, and plenty never reach decline at all because they reinvent themselves into a new growth curve.

The four textbook stages map loosely onto the three or four buckets you actually use when picking metrics. The table below is the working version — what you collect, and what you can price it on.

Stage → financials you collect → multiples that work

Stage	Revenue growth	Financials you collect	Multiples that work
Early stage startups, pre-profit	50–200%+	Revenue growth · gross margin, as a read on unit economics · EBITDA/operating margin and its trend · customer metrics (monthly and daily active users, churn, customer acquisition cost, lifetime value) · market size / total addressable market	EV/Revenue · EV/Monthly recurring revenue · EV/Gross profit · EV per business metric (e.g. EV/Monthly active user)
Mid-stage scaling, turning profitable	20–50%	Revenue growth · EBITDA margin moving from around breakeven into the mid-teens, or already there · net margin around breakeven to single digits	EV/EBITDA and EV/EBIT become the norm · EV/Gross profit still used · P/E once profit is substantial enough to mean something
Mature established, cash-generative	Low single digits – 10%	EBITDA and net margins, which are expected to be high · cash flow · capacity to return cash through dividends and buybacks	EV/EBITDA and P/E · EV/(EBITDA − capex) and EV/Free cash flow are defensible but rarer, being slower to build and more assumption-laden
Declining	Flat to negative	As mature — margins and cash generation	As mature — EV/EBITDA and P/E

Three things worth pulling out. A negative margin is still worth collecting if it is improving — −100%, then −50%, then −10% is one of the strongest signals an early-stage company can give, and venture funding is what buys it the time to get there. Total addressable market rarely appears in the spreadsheet itself; it sits behind the decision about which peers belong in the set at all. And mature does not mean declining — Apple grows a few percentage points a year with very high EBITDA margins, and by then investors are there for the cash returned to them rather than for growth.

↓ Drill this table — drag the metrics into their stage ↓ Or run the whole decision end to end (Case 5)

Whatever the stage, one block of data is always collected, because enterprise value cannot be built without it:

Share priceShares outstandingCashDebtPreferred stockNon-controlling interest

A reality check on the last two, after a whole module spent on them in Part III: across hundreds of real companies, only something like 1–5% actually carry preferred stock or non-controlling interest at all. You need to know what they are and how they enter the bridge, but most of the time the line is simply zero.

The rule that kills a multiple

A multiple stops being usable when its denominator collapses. A company near breakeven can show a 2,000× EV/EBITDA — not because anyone believes it is worth 2,000 years of earnings, but because you have divided a large enterprise value by an EBITDA of almost nothing. Negative denominators fail the same way.

Those cells get dashed out rather than included. The hard cut in the table below is 100×; the practical view is that anything much above roughly 50× has stopped carrying information.

What the actual comps table looks like — streaming peer group

Illustrative — synthetic figures for teaching purposes

Company	Share price	Mkt cap	EV	Rev growth '26E	EBITDA margin	EV/EBITDA	P/E
Disney	$105	$190bn	$235bn	3%	18%	9.2×	19.4×
Comcast	$38	$145bn	$260bn	1%	27%	6.8×	10.2×
Warner Bros. Discovery	$10	$24bn	$55bn	−3%	15%	6.1×	—
Paramount	$12	$9bn	$22bn	0%	11%	7.4×	—
Netflix (target)	$1,150	$495bn	$505bn	9%	30%	15.8×	32.1×

Netflix trades well above every streaming peer on both multiples — the fundamentals rule from earlier holds: higher growth and margins earn a premium multiple. P/E is dashed out for Warner Bros. Discovery and Paramount because their net income is too thin for the multiple to mean anything, exactly as flagged in the multiples section.

Reading that table back through step 2 — why these columns and not others

Netflix runs a 30% EBITDA margin and mid-twenties net margins. It is profitable, comfortably, so EV/EBITDA and P/E both mean something and both earn a column — with EV/EBITDA the better of the two, for the capital-structure reason set out in the P/E disclosure below.

Revenue multiples are deliberately absent. Netflix still grows quickly for its size, but it is a mature, high-margin business, and pricing it on revenue would throw away the very thing that makes it valuable. That is the stage framework doing its job: the table's shape was decided before a single figure was collected.

Also worth noting — the colour coding is a teaching device. Real bank comps tables are not colour-coded.

Expand any column above you're not sure of.

Share price

The market's current price for one share of the company's stock — the raw number everything else in the table is built from.

Mkt cap Market Capitalization

Share price × total shares outstanding. This is the exact same thing as equity value from Part III, just wearing its more common trading-desk name — value to common shareholders only, nothing else.

EV Enterprise Value

The value of the whole business to every capital provider — equity value plus debt, preferred stock and NCI, minus cash. Fully covered back in Part III; it's repeated here because it's the numerator for the multiple two columns to the right.

Rev growth '26E

How much more revenue the company is projected to generate in 2026 versus the prior year. The "E" means estimated — nobody knows this for certain yet; it's the market or analyst consensus forecast.

EBITDA margin

EBITDA divided by revenue, as a percentage. Shows how much of every dollar of sales survives as operating profit before interest, tax, depreciation and amortization — a read on how efficiently the core business converts revenue into earnings.

EV/EBITDA

The whole-business price tag, measured in years of pre-interest, pre-tax operating earnings. See the "Purpose of EV/EBITDA" box back in Part III's non-controlling interest section for the full explanation of why this is the standard comps multiple.

P/E Price / Earnings

Share price divided by earnings per share — equivalently, market cap divided by net income. It tells you how many years of the company's current per-share profit it would take, at today's earnings, to earn back what you paid for one share. A higher P/E means the market is paying more per dollar of current profit, usually because it expects that profit to grow faster or believes it's safer and more durable — the same underlying logic as EV/EBITDA, just applied to the equity slice alone.

One real difference worth knowing: P/E is built on net income, which already has interest expense baked in — so unlike EV/EBITDA, P/E is not capital-structure-neutral. Two companies with identical operations but different debt loads will show different P/E ratios purely because of financing choices, which is exactly the distortion EV/EBITDA exists to avoid.

Step 3

Analyse & value

Two different activities under one heading. The first is where you form a view; the second is multiplication.

Analyse

Read each peer's growth and margins against its own multiple
Trading above the set → the market expects higher growth and profitability
Trading below it → the market expects less of both
Not a law. Plenty else moves a multiple — but it holds most of the time, and a real exception is worth understanding rather than ignoring

Value

Choose a range of multiples from the peer set
Apply it to the target's own metric
That is the entire methodology, and it is genuinely surprising how little there is to it
All of the difficulty has already happened, in steps 1 and 2

Who actually picks the range

Not the analyst, usually. A more senior analyst might take a stab and put it to the associate, who takes their own stab and puts it to the VP or director — and the director, with decades of looking at these companies, sets the low and the high that go in the book. The arithmetic is delegated all the way down; the judgement is held right at the top. It is the clearest illustration of the hierarchy from Part I of how the work actually divides.

How the range gets set	What it is	How often you see it
Judgement low / high	A senior banker picks a low and a high from the peers they consider genuinely comparable	The norm
25th / 75th percentile	Mechanical quartiles of the peer set	Rare — too formulaic and restrictive to survive a banking presentation
Mean / median	Not a range at all — a single benchmark for whether the target looks rich or cheap	Common, as a sanity check rather than an output

Two habits make the output more defensible. Use more than one metric — EV/EBITDA and P/E, not one of them — and use more than one year, the current year and the next. Each additional pairing is another data point to triangulate against, and a range that survives all of them is a range you can defend in a room.

Worked example — applying a multiple range

Target's 2026E EBITDA = $100m. Peers trade at 10×–15× 2026E EBITDA.

Low end: 100 × 10 = $1.0bn · High end: 100 × 15 = $1.0bn – $1.5bn range

Working it on the streaming set — and what the answer exposes

Illustrative — derived from the synthetic figures in the comps table above

Streaming peer	EV/EBITDA	Statistic
Disney	9.2×	Low — high	6.1× – 9.2×
Comcast	6.8×	Median	7.1×
Warner Bros. Discovery	6.1×	Mean	7.4×
Paramount	7.4×

Netflix's own EBITDA is implied by its table row: EV of $505bn at 15.8× gives roughly $32.0bn. Apply the peer range to it:

Low: 32.0 × 6.1 = $195bn · High: 32.0 × 9.2 = $294bn · Actually trades at $505bn

The streaming comps say Netflix is worth roughly $195–294bn. The market says $505bn — the range tops out at well under two-thirds of the traded value. The honest reading is not that the market has mispriced it by half — it is that a peer set built from Disney, Comcast, Warner Bros. Discovery and Paramount is the wrong lens for a business the market prices as a high-growth technology company. Step 3's output has just indicted step 1's input, which is precisely what a director is paid to notice before the range reaches a client. Run the same exercise against the broader-tech peer group and you get a different answer — and choosing between the two is the whole job.

The other way comps get used

Not every comps exercise is trying to produce a valuation range. Often the question is simply relative: we trade at 10× revenue, the seven peers sit around 7× — we are expensive relative to our industry. No range, no football field, just a positioning read. It is a completely legitimate use of the same table, and a common one.

▶ Study steps 2 and 3 in more depth — watch them in the source video (from 4:45:33)

Relative valuation, past tense

Transaction comps

Interview answer

A relative valuation methodology comparing financial metrics across a set of previously completed M&A transactions similar to the one being analysed.

	Trading comps	Transaction comps
Time horizon	Present / future (NTM — Next Twelve Months)	Historical, at time of deal (LTM — Last Twelve Months / NTM)
Peer criteria	Strict	Looser — smaller universe of deals
Includes a premium?	No	Yes — inflates the multiple
Needs updating?	Constantly	Never, once built
Reliability	Good — same-day market conditions	Weaker — conditions vary deal to deal
Typically yields the	Middle valuation	Highest valuation of the three methods

LTM (Last Twelve Months) / NTM (Next Twelve Months), anchored to a deal's announcement date (example: 1 March 2025)

Build the transaction set
4–8 precedent deals, judged on the same criteria as the trading-comps peer set, plus one more: recency. A deal from five years ago carries less weight than one from last quarter, because multiples, rates and sentiment all drift.
Select, collect, calculate
The same three jobs as trading comps, run against a thinner, harder-to-reach dataset — plus one extra figure to calculate that trading comps never needs: the premium paid.
Analyse & value
Set a defensible multiple range from the deal set, apply it to the target's own metric.

The three steps keep the same names as trading comps' — deliberately. It's the same underlying process, pointed at a harder dataset.

Step 1

Build the transaction set

Same criteria as trading comps' peer set — business model, industry, geography, size, financial profile — plus recency, which trading comps doesn't need. A deal announced last quarter says more about what the market will pay today than one from five years ago.

The precedent transactions behind ByteTech Co.'s range, below

Illustrative — synthetic figures for teaching purposes

Acquirer	Target	Announced	EV	LTM EBITDA	NTM EBITDA	Premium
Argon Capital	Fleetbase	Aug 2022	$1,850m	16.0×	12.5×	28%
Vantage Holdings	Ridgeline SaaS	Nov 2022	$980m	14.2×	11.0×	24%
Union Point Partners	Metric Works	May 2023	$1,410m	11.5×	9.2×	21%
TechSphere	CloudNova	Mar 2024	$3,600m	12.0×	9.5×	32%
Solera Group	Northlight	Jan 2024	$2,240m	10.0×	8.0×	19%

Five deals spanning 2022–2024 — exactly the 10.0×–16.0× LTM and 8.0×–12.5× NTM range applied to ByteTech Co. in step 3, below. Premiums cluster 19–32%, consistent with the 10–30% rule of thumb covered in step 2.

Step 2

Select, collect, calculate

The same three jobs as trading comps' step 2 — each one harder against a transaction dataset, plus one figure trading comps never has to touch: the premium paid.

Select

Which metrics

Driven by the target's stage, same logic as trading comps
Constrained further by what the deal actually disclosed — private targets leave far less on the record than a listed peer
LTM and NTM EBITDA (or revenue, pre-profit) do almost all of the work

Collect

From where

Harder than trading comps — a smaller universe of deals, thinner disclosure
Old equity research and Bloomberg data as of the announcement date, deal proxies, press releases — sometimes reconstructing what an NTM projection looked like years after the fact
Once a deal has closed the numbers never move — reuse another team's file rather than rebuild it

Calculate

Multiples, plus one extra

EV/EBITDA (or revenue) LTM and NTM, exactly as before
Plus the premium paid — see below
Growth rates and margins are a bonus here, not a requirement, unlike trading comps

Step 2 in practice — five precedent deals, mean and median calculated across the set

Illustrative — synthetic figures for teaching purposes

Acquirer	Target	Ann. date	EV	EBITDA		EV/EBITDA		Premium (1-day)
Acquirer	Target	Ann. date	EV	LTM	NTM	LTM	NTM	Premium (1-day)
TechSphere Inc.	CloudNova	Jul 2025	$3,600m	$300m	$360m	12.0×	10.0×	32%
DataCore Systems	NetLogic Software	Feb 2024	$2,250m	$180m	$225m	12.5×	10.0×	28%
CyberEdge Group	SecureNet Holdings	Sep 2023	$1,800m	$150m	$200m	12.0×	9.0×	35%
NextGen Cloud	Streamify Media	Mar 2023	$4,000m	$250m	$320m	16.0×	12.5×	40%
OmniTech Corp.	AppLink Solutions	Feb 2022	$1,200m	$120m	$150m	10.0×	8.0×	25%
Mean						12.5×	9.9×	32%
Median						12.0×	10.0×	32%

Mean and median land close together here because the set is tight and none of the five deals is an outlier — that won't always hold, which is exactly why step 3 treats mean/median as a sanity check rather than the range itself. Data for transaction comps is sparser than for trading comps, so growth rates and margins aren't required to build this table — and once it's built accurately, it never needs doing again for these particular deals.

Acquisition premium

Definition

The percentage difference between the price per share the acquirer pays and the target's unaffected share price — the price before any rumours or leaks moved it.

Real example — Palo Alto Networks' $25bn acquisition of CyberArk, July 2025

$383 → $434 almost overnight — the jump on the 29th, a day before the formal announcement, is the market pricing in a leak. $25bn total deal value included a 26% premium to CyberArk's unaffected 10-day volume-weighted average price.

Four reasons a buyer pays above the unaffected price at all:

Control

Buying the whole company, not a minority stake — full control over its future decisions is worth paying for.

Synergy

Expected cost savings or revenue gains from combining the two businesses — covered properly in the accretion/dilution module. The more synergy the acquirer expects, the more it can afford to pay.

Competition

More than one bidder chasing the same target pushes the price — and the premium — higher.

Incentive

It removes execution risk for the seller. A stock that might double organically over a couple of years is a bet; a guaranteed premium today isn't — the premium is what buys the shareholders' certainty now.

Typical range

10–30%, with 30% the safe one-number answer if asked for a single figure. It can run well past that — strategic buyers chasing large synergies, a hot M&A market, and multiple competing bidders for a strong performer all push it higher; a financial buyer (private equity) optimising for leverage and efficiency tends to sit closer to the low end.

Worked example — ByteTech Co. (the source video's own illustrative figures)

Trades today at 9.5× LTM EBITDA / 8× NTM EBITDA. Five precedent deals in the sector closed at 10×–16× LTM and 8×–12.5× NTM (premium-inflated, so naturally above where ByteTech trades unowned).

LTM: $200m × [10×–16×] = $2.0bn – $3.2bn EV
NTM: $225m × [8×–12.5×] = $1.8bn – $2.8bn EV

Step 3

Analyse & value

Same mechanics as trading comps' step 3 — pick a range, multiply it by the target's own metric — with one twist on how the range gets picked.

How the range actually gets set

Min-to-max, judged by a senior banker, is the norm — same as trading comps. 25th/75th percentile shows up occasionally. Mean and median, calculated back in step 2, stay what they were there: a sanity check on whether a given deal looks rich or cheap against the set, not the range itself.

Applied to ByteTech's own metrics, that's exactly the worked example above — the step 1 deal set's 10×–16× LTM and 8×–12.5× NTM range, multiplied straight through step 2's calculated multiples.

▶ Study transaction comps in more depth — watch it in the source video (from 5:02:00)

Intrinsic valuation

The discounted cash flow

Time value of money

A dollar today is worth more than a dollar tomorrow — you could invest it and earn a return in the meantime.

Worked example — your rich uncle's offer

Option A: $100k in one year + $100k in two years. Option B: $180k today. Assume a reliable 10% annual return.

PV(yr 1) = 100,000 / 1.10 = $90,909
PV(yr 2) = 100,000 / 1.10² = $82,645
Total PV of Option A = $173,554 — less than the $180,000 on the table today. Take Option B.

Interview answer — what is a DCF?

An intrinsic valuation methodology based on a company's future free cash flows. Project cash flow 5–10 years, calculate terminal value (exit multiple or perpetuity growth), discount both back at the WACC (Weighted Average Cost of Capital) to reach enterprise value.

Five steps, in order

Step 1 — unlevered free cash flow

UFCF = EBIT × (1 − tax rate) + D&A − CapEx − Δ Net Working Capital

Available to all investors, debt and equity alike — which is why it starts from EBIT (before interest) rather than net income.

Explicit forecast period

How many years you actually project, before handing the rest to terminal value. Two standard answers, picked by company type rather than a fixed rule: 5 years for mature, stable companies, where growth and margins have already normalised and years 6–10 wouldn't move the answer much anyway; 10 years for early-stage, high-growth companies — long enough for growth to steady to a lower rate and margins to stabilise by the end of the window. 15–20+ years for long-ramp industries like biotech, which can run negative free cash flow for years before a drug approval (or equivalent) flips the switch — the presenter cites personally building 20–25 year DCFs for biotech clients at JPMorgan. The governing principle above all three: forecast as long as you can reasonably defend your assumptions, then rely on terminal value for the rest.

Step 2 — WACC

WACC = (%Equity × Cost of Equity) + (%Debt × Cost of Debt × (1 − tax rate))

Worked example — 80% equity / 20% debt, 20% tax rate

Step 3 — terminal value

Perpetuity growth

TV = FCF × (1+g) / (WACC − g)

Grounded in economic theory
g typically 2–3% — a company can't outgrow the economy forever
Very sensitive to small changes in g or WACC

Exit multiple

TV = Final-year EBITDA × exit multiple

Simple — it's just a comps analysis
Anchored to real market data
Reintroduces market noise into an otherwise intrinsic model

Worked example — exit multiple

Year-5 EBITDA $200m × 8× exit multiple = $1.6bn terminal value

Steps 4–5 — discount, then bridge to share price

From enterprise value to an implied share price

Compare the $8 implied price against the current market price — trading above it signals overvalued against these assumptions; trading below signals a buying case, assuming the model's inputs hold up.

For a private company there's no market share price to compare against — the DCF simply stops at enterprise value or equity value.

Quiz · Part II material

Accounting drills

Drag & drop

Sort the line items

The fastest way to stop second-guessing where something belongs. Drag an item into a bucket — or tap the item, then tap the bucket. It turns green or red the instant it lands, and you can move it again until it's right.

The last two are the life-cycle pair from step 2 of trading comps, deliberately split: what you collect at each stage, then what you can price it on. They are two different decisions and it is worth being able to make them separately.

Quiz · Part III material

Equity & enterprise value drills

Quiz · Part IV material

Valuation drills

Quiz · everything

Mixed drills

Full worked cases

Build it yourself, start to finish

Five synthetic companies with the raw data laid out in front of you, including items that don’t belong where you might first put them. Work down the steps in order — each one checks itself, and you can reveal any answer if you get stuck. The last one is different in kind: it hands you a company and no instructions, and the first thing you have to decide is what to measure it on at all.

Industry Fundamentals

Four divisions, two transaction types

Capital raises

M&A Mergers & Acquisitions

Product groups vs. coverage groups

Product groups

Coverage groups

Five ways banks add value

Bank tiers, and the ladder inside one

Accounting

The income statement

Three rules for what belongs on it

Accrual accounting

The cash flow statement

The balance sheet

Shareholders' equity, line by line

One company, three statements, fully tied together

Picking up here

Equity Value & Enterprise Value

The house analogy

Equity value, properly

Investor hierarchy

Enterprise value

Non-controlling interest

Valuation

Relative vs. intrinsic

Relative

Intrinsic

Trading comps

Build the peer set

Select, collect, calculate

Analyse & value

Build the peer set

Select, collect, calculate

Which metrics

From where

Then the easy part

Analyse & value

Analyse

Value

Transaction comps

Build the transaction set

Select, collect, calculate

Analyse & value

Build the transaction set

Select, collect, calculate

Which metrics

From where

Multiples, plus one extra

Acquisition premium

Control

Synergy

Competition

Incentive

Analyse & value

The discounted cash flow

Step 1 — unlevered free cash flow

Step 2 — WACC

Step 3 — terminal value

Perpetuity growth

Exit multiple

Steps 4–5 — discount, then bridge to share price

Practice

Accounting drills

Sort the line items

Equity & enterprise value drills

Valuation drills

Mixed drills

Build it yourself, start to finish

See the real thing

Apple

Coca-Cola

Netflix