Airlines, Hotels, Media + Synthesis

Phase 5 · Week 51 · ~20 hours

Where it sits. The tenth and final sector playbook of Phase 5, and the phase's capstone. The nine modules before this one each handed you a lens: P/B-vs-ROE for banks, P/EV for life insurers, EV/ARR for SaaS, EV/EBITDA-on-mid-cycle for commodities, P/NAV and FFO for real estate. This week does two jobs. First, it teaches the last three sector playbooks: airlines, hotels, and media/streaming, three consumer-facing, superficially glamorous industries whose economics run from structurally value-destroying (airlines) through deeply cyclical (hotels) to capital-hungry land-grab (streaming). Second, and more important, it synthesizes all seventeen sectors into one map. Given any business: which lens, which KPIs, which multiple, always checked against the one test that sits under every lens, ROIC > WACC. On the five-questions spine (in the master map), the whole of Phase 5 has been question 6 of the competency map, analyze any industry, and this is where you prove you can pick the right tool on sight rather than reciting a memorised list. It is the M6 milestone (in the master map): a right-lens exam plus two full sector-playbook applications, certifying competency C6.

What comes before / after. Before: Commodities, Real Estate & REITs. Airlines and hotels are cyclicals wearing service uniforms, so its capital-cycle and mid-cycle-normalisation machinery carries straight over. After: passing this gate completes Phase 5 and opens Phase 6 — Investing Philosophy, Mental Models & Behavioral Edge.

Suggested hour budget. Core content (three playbooks + the synthesis): 8h · worked examples: 1.5h · practice set: 3h · the right-lens capstone exam (20 vignettes, self-graded): 1.5h · applied mini-project (two full playbook applications, the module gate): 4h · flashcards + mastery check: 1.5h · teach-back + journal: 0.5h.

App link. The sector case engine's three hardest drills run on this week's material: the airline survival drill (a demand shock hits mid-simulation; you must read days-cash, unit profit, and lease-adjusted leverage to call who survives), the RevPAR-cycle drill (a hotel supply pipeline floods four years after a RevPAR peak; call the turn), and the centrepiece, the right-lens engine, where mixed company vignettes stream past and you tag lens + top-3 KPIs + the one red flag against the clock. Your completed panels seed the airlines, hotels, and media pages of the sector KPI cheat-sheet artifact. Passing the mastery check and both playbook applications certifies C6 and unlocks the next phase.

Learning objectives

By the end of the week you can:

1. Explain, from first principles, why the airline industry structurally destroys capital (a commodity product, perishable inventory, high operating leverage, powerful suppliers, and fuel/FX exposure), and why a disciplined low-cost survivor (IndiGo-style) can still compound while the graveyard fills around it.
2. Compute and interpret the full airline unit-economics panel: ASK, RPK, load factor (LCC breakeven ~80–85%), yield, RASK, CASK and CASK ex-fuel, unit profit (RASK−CASK), aircraft utilization, fuel % of opex, ancillary revenue %, net debt including capitalized leases, and days-cash. Then value an airline on EV/EBITDAR, explaining exactly why the "R" is there.
3. Run the hotel playbook: build RevPAR = ADR × occupancy and GOPPAR from scratch, read RevPAR Index versus a comp set (>100 = share gain), separate the owned/leased (EV/EBITDA, cyclical) business from the managed/franchised fee-income (P/E, asset-light) business, and explain the multi-year supply-lag cycle that drives the whole sector.
4. Run the media/streaming playbook: net adds, ARPU, monthly churn (<3% healthy), content spend and amortization %, engagement/watch-time as the leading churn predictor, and the ad-vs-subscription mix. Value a streamer on EV/Sales and per-sub, tracing the land-grab → scale → harvest arc.
5. Recite and apply all three red-flag panels (airlines, hotels, media) fast enough to use them in a timed teardown.
6. Reconstruct the full cross-sector synthesis map, every sector from M5.01–M5.10 with its model, KPIs, and multiple, and pick the correct lens in under 30 seconds via an explicit decision procedure.
7. Sanity-check every valuation against ROIC > WACC, explaining why a cheap multiple on a capital-destroyer is a trap and a rich multiple on a high-return reinvestor can still be cheap.
8. Pass the right-lens capstone exam (≥17/20) and execute two full sector-playbook applications on real companies to the rubric bar.

Prerequisites & connections

Builds on. Every earlier Phase 5 module, M5.01 to M5.09. This capstone is nothing but their synthesis, so you need to be able to run each of their panels before you can choose between them. M3.06 on relative valuation gives you the driver algebra behind every multiple, and why EV/EBITDA, P/E, P/B, and EV/Sales each price a different thing. M3.05's reverse DCF ("what's priced in") comes back when we ask what a streamer's per-sub value implies about future ARPU and churn. M3.07 and M2.03 supply ROIC vs WACC, the master test under all lenses, and show how to build NOPAT and invested capital for an asset-heavy hotel against an asset-light brand. M3.10 is mid-cycle normalisation. Airlines and hotels are cyclicals, so you value them on normalised earnings instead of trailing ones. M1.08 covers lease accounting (ASC 842 vs IFRS 16/Ind AS 116), the whole reason EV/EBITDAR exists. M2.02 treats net debt and coverage as survival metrics, along with days-cash. Finally, M4.03 to M4.06 bring Porter and the capital cycle. Airlines are the textbook five-forces disaster, hotels are the textbook supply cycle, and streaming is scale economies shared.

Feeds into. Phase 6 is next, and once you can classify any business, the schools of investing become choices about which good businesses to own. In Phase 7, airlines turn out to be a levered bet on fuel and the cycle, and hotels a bet on GDP and rates. All three are macro transmission made visible. Phase 8's 1 to 2 hour teardown of any company in any sector runs on the lens-picking reflex you build here. It's the sector half of the signature skill. Phase 9 picks up the fact that cyclicals and land-grab growth names demand different position sizing than compounders. And the synthesis table you build here is one of the most-used artifacts in the whole course.

Core teaching content

4.1 Three businesses, one meta-lesson

A flight, a hotel night, a show: three things a consumer can touch. All three attract capital because they look exciting and their end-products are visible and loved. And all three punish investors who value them on the wrong lens:

• Airlines are the purest demonstration in all of business that a wonderful product can be a terrible business. Everyone loves cheap flights, yet almost no one has made durable money owning the airlines that provide them. Value one on a trailing P/E and you'll buy the graveyard.
• Hotels teach that the same physical asset supports two utterly different businesses. Owning the building is capital-heavy and cyclical. Managing the brand on someone else's building is asset-light and annuity-like. The market pays wildly different multiples for each, and it's right to.
• Media/streaming teaches that a business can be worth a fortune before it earns a rupee of profit, or nothing despite millions of subscribers. The difference is unit economics (ARPU, churn, content cost), and a naïve subscriber-count headline hides it completely.

Everything in Phase 5 points to one lesson. The business model dictates the KPIs and the multiple, and the multiple is only ever grammar. The meaning is always ROIC > WACC, durably, with room to reinvest. Hold on to that as we go.

4.4 The airline balance sheet: leases, leverage, and days-cash

Airlines are lease machines. Most carriers lease a large share of their fleet instead of buying it, which preserves cash and shifts residual-value risk. That leaves you two analytical problems to handle:

• Net debt must include capitalised leases. Under Ind AS 116 / IFRS 16, effective 2019 for all Indian and IFRS reporters, almost every lease appears on the balance sheet as a right-of-use asset and a lease liability, and that liability is debt. Under US GAAP ASC 842, operating leases are also capitalised as a liability, though the P&L keeps a single straight-line lease expense inside operating cost. Either way, an airline's real leverage is gross debt + lease liabilities − cash. A carrier that looks ungeared on borrowings alone can be drowning once you add ₹40,000 to 50,000 crore of lease obligations. So go and find the lease note before you call any airline lightly geared. Ignoring leases is the most common airline-analysis error of all.
• Days-cash-on-hand = cash & equivalents / average daily operating expense, and it's the survival metric. An airline can lose a fortune in a quarter when demand vanishes: 9/11, COVID, a fuel spike. So don't ask whether it's profitable this year. Ask how many days of expense it can fund with no revenue. A well-run carrier holds several months, 150 to 250+ days. A fragile one holds weeks. Thin days-cash plus negative net worth is the pre-death signature. Negative net worth means accumulated losses have eaten the equity, and it's common in this sector. Kingfisher, Jet, and Go First all showed it before they stopped flying.

Two more things to read alongside. The first is fuel hedging. Does the carrier hedge any fuel at all? Going unhedged into a spike is a red flag, and being over-hedged into a crash is its own problem. Most Indian carriers hedge little, so they're near-fully exposed. The second is the fleet order book against demand. A fleet growing faster than the market means capacity dumped into weak demand and collapsing yields for everyone. That's red flag #5.

4.6 Airline red flags (the brief's five, expanded)

1. High load factor but falling yields. Fullness bought with price. Read RASK instead of load factor. If the plane is fuller but unit revenue is down, the carrier is subsidising traffic. It's a late-cycle tell.
2. Rising CASK ex-fuel. This is the one cost line management fully controls, and it's drifting up. You can forgive a fuel-driven CASK spike. A rising ex-fuel CASK means slack cost control or lost scale, and either one is a moat eroding.
3. Negative net worth + thin days-cash. Accumulated losses have eaten equity, and the liquidity buffer is measured in weeks. That's the pre-insolvency signature. Treat it as an emergency, and don't mistake it for a value opportunity.
4. Unhedged fuel into a spike. No hedge plus a crude rally converts thin unit profit into unit loss overnight. Check the hedging note before every fuel-cycle call.
5. Fleet growth into weak demand. When deliveries and lease commitments rise while the market softens, capacity gets dumped and yields get crushed across the whole industry and not only at this carrier. Watch the order book against demand.

4.9 Hotel valuation and red flags

For valuation, put an asset-heavy owned business on EV/EBITDA on mid-cycle EBITDA, normalised across the RevPAR cycle the same way you'd treat any cyclical. Cross-check it against net debt/EBITDA and against EV per key/room, which is a replacement-cost sanity check. An asset-light fee business goes on P/E or EV/EBITDA on the fee stream at a premium, driven by pipeline/room growth and fee margin. Mixed players get SOTP. Never put an asset-light multiple on an owned business, or the other way round.

The four hotel red flags:

1. RevPAR up only on ADR while occupancy falls. That's growth from price alone, with fewer heads in beds. It's often late-cycle and fragile, because the next downturn hits an occupancy that's already thinning.
2. RevPAR Index <100. You're losing share to your competitive set whatever the market is doing. That points to a brand or asset-quality problem, and the market cycle can hide it for a while.
3. Debt-funded owned expansion at the peak. Building capital-heavy, on borrowed money, just when RevPAR is highest is capital-cycle suicide. The new rooms open into the glut the building itself helped create. A careful operator expands asset-light and buys owned assets in the trough.
4. Surging supply pipeline. Heavy new supply coming into your micro-market means future RevPAR pressure. Read the pipeline data as well as today's numbers.

4.10 Media / streaming: subscriber economics

A streaming (SVOD) business is a subscriber machine, and its economics are a cohort story. The headline everyone quotes, total subscribers, means almost nothing on its own. The value is in the quality of those subs. Here's the panel:

• Net adds = gross new subscribers − churned subscribers, per quarter. It's the growth signal the market trades on. But it nets two very different forces, so always pull it apart. Strong gross adds with high churn is a leaky bucket. Weak gross adds with tiny churn can be a maturing, valuable base.
• ARPU, Average Revenue Per User = subscription (and increasingly ad) revenue / average subscribers. This is the price dial. A market can be high-ARPU, like US SVOD at ~$10 to 17/month. Or it can be low-ARPU, like India, often at tens of rupees or ad-supported free. The same subscriber count is worth completely different amounts depending on which. Rising ARPU is the harvest lever, and it comes through price hikes, mix, or ad monetisation.
• Monthly churn = subscribers lost / starting subscribers, per month. Below ~3% monthly is healthy for SVOD. Above it, the bucket leaks faster than marketing can fill it. Why is churn the most important quality metric? Because it drives customer lifetime: average customer life ≈ 1 / monthly churn. At 2% monthly churn the average subscriber stays ~50 months. At 5%, it's only ~20 months. That 2.5× difference in lifetime is a 2.5× difference in lifetime value on the same ARPU. So a low-ARPU, low-churn service can be worth more than a high-ARPU, high-churn one.
• Content spend & content amortisation %. Content is a streamer's cost of goods. It's capitalised when produced and amortised over its expected viewing life, so two numbers matter. Cash content spend is what actually leaves the bank, and it's the FCF drag. Content amortisation is the P&L expense. When cash spend runs well above amortisation, the company is front-loading investment. It's building a library and burning cash, which is the land-grab phase. When the two converge, spending has matured. Content as % of revenue (amortisation ÷ revenue) shows how much of every rupee of sales goes back into the screen, and 40 to 60% is common. Watch for the treadmill. If content spend has to keep outgrowing revenue just to hold subscribers, the business never generates owner cash.
• Engagement / watch-time is hours watched per member. It's the leading predictor of churn and the best early-warning metric in the sector. A subscriber whose viewing hours are falling is a subscriber about to cancel, months before they actually do. Netflix built its entire product strategy on maximising engagement. Engagement retains subs, and retention is the whole game. When a streamer stops disclosing engagement, or it's falling, start worrying before the churn number confirms it.
• Ad vs subscription mix, and SAC. The industry is pivoting from pure subscription toward advertising (ad-supported tiers, free ad-supported streaming, or FAST). Ad revenue diversifies ARPU and monetises the low-willingness-to-pay users who'd otherwise churn or never join. SAC (subscriber acquisition cost) is marketing spend per gross add, the CAC of this world. Rising SAC with slowing growth is the same warning it is in SaaS.

Every streamer travels the same road, from land-grab to scale to harvest. Where a company is on that road determines both its cash flows and its correct valuation:

1. Land-grab: spend on content and marketing ahead of subscribers to build a library and a base. Cash burns and losses are large. The bet is on future scale. Value it on EV/Sales and per-sub, because there are no profits to capitalise.
2. Scale: subscribers and ARPU rise until revenue covers content + marketing. The business crosses into profit and FCF.
3. Harvest: with the base built and content spend maturing, pull the monetisation levers (price hikes, ad tiers, cracking down on password sharing) and convert to real FCF and margin.

Netflix is the one large streamer that has reached harvest. It's profitable and FCF-positive, it's raising ARPU, and it's monetising ads and shared passwords. That's why it re-rated from an EV/Sales story to an earnings-and-FCF story. Much of legacy media (Disney+, Warner, Paramount's services) spent years still in the land-grab and scale phases, burning cash to chase Netflix.

India is the instructive opposite corner. It's an enormous, low-ARPU market where cricket rights (the IPL) drive huge subscriber spikes at tiny or zero ARPU. The JioStar entity, the 2024 Disney-Reliance media JV combining Hotstar and JioCinema, is a volume-and-advertising play instead of a high-ARPU subscription play. Same KPIs, opposite settings on the dials.

When it works, the moat is a flywheel of engagement and scale. More subs pay for a bigger content budget. That buys better and more content. Engagement rises and churn falls. Lower churn means more subs. It's scale economies shared.

4.11 Media/streaming valuation and red flags

For valuation, a growth-phase, pre-profit streamer goes on EV/Sales, with EV per subscriber as the cross-check. What's the market paying per sub, and does this service's ARPU and churn justify it against a benchmark like Netflix? At maturity, move to EV/EBITDA and P/E as real earnings appear. Always pair the per-sub number with churn and ARPU. 100 million subs at $4 ARPU and 6% churn are nowhere near 100 million subs at $14 ARPU and 2% churn. A single blended "EV/sub" that ignores the difference is how people overpay.

The five media red flags:

1. Decelerating or negative net adds. The growth engine stalling. In a business valued on growth, deceleration re-rates the multiple hard.
2. Rising churn. The bucket is leaking faster. Watch engagement for the leading signal before churn confirms it.
3. Content spend outgrowing revenue/subs. This is the treadmill, running faster to stand still with no path to owner cash.
4. Growth reliant on one-off levers. A single hit show, a sports-rights window, or a password-crackdown bump that lifts a quarter but doesn't repeat. Ask whether the run-rate survives the lever's removal.
5. Heavy content write-offs. Impairments of shelved or underperforming content are management admitting the spend was wasted. That gives you a direct read on content ROI, and it's often a sign the library was built on hope instead of data.

4.13 The one test under all of them: ROIC > WACC

Every lens in that table is a convenience, a shortcut to the same underlying question. And the question never changes. Does this business earn a return on invested capital above its cost of capital, durably, and can it reinvest at that spread? It's question 2 of the master map's five questions, and value is created only there.

Here's why the multiple is grammar and ROIC>WACC is meaning:

• A low multiple on a capital-destroyer is a trap. Think of a steelmaker at 5× peak earnings, an airline at a "cheap" EV/EBITDAR on peak-cycle profits, or a developer at a low P/E on revenue recognised from four-year-old sales. Each one looks cheap and isn't, because the underlying business earns less than its cost of capital through the cycle. The cheap multiple is the market correctly refusing to capitalise earnings that won't last.
• A high multiple on a high-return reinvestor can be cheap. Take an FMCG compounder at 45× earnings with 40% ROCE, reinvesting into a growing category. Or a fee-based hotel brand at 30× with 45% ROIC and a long room pipeline. Or a SaaS firm at 12× sales at Rule-of-40 with 125% NRR. Each can justify its multiple because the spread of ROIC over WACC is wide and reinvestable, so it compounds value for years. That's the McKinsey key-value-driver logic you met in the valuation phase: Value = NOPAT·(1 − g/ROIC)/(WACC − g). Value rises with ROIC and, when ROIC>WACC, with growth.

So the routine runs like this. Pick the lens from the model. Compute the KPIs, then read the multiple. Then always close the loop by asking what ROIC the business earns and whether the KPIs you just computed confirm it's durable. The sector KPIs aren't trivia. Each one is a leading indicator of whether the ROIC>WACC spread will persist. NIM and credit cost tell you a bank's ROA will hold. NRR tells you a SaaS firm's revenue compounds without re-spending. Cost-curve quartile tells you a miner survives the trough. And churn and engagement tell you whether a streamer's subs, and so its ROIC, are real.

Think of every KPI as a spy reporting on how durable the spread is.

4.14 Picking the lens in 30 seconds: the decision procedure

In a Phase 8 teardown you won't have time to deliberate. Run this flowchart until it's a reflex:

1. Is it a financial? Does it make money on a spread, on float, or on the value of a book? Then use a book-value or embedded-value lens. A bank or NBFC gets P/B vs ROE. A life insurer gets P/EV and VNB. A general insurer gets P/B plus the combined ratio. Financials are valued on equity instead of enterprise value, because debt is their raw material.
2. Is it capital-light and growing fast on recurring revenue? Then use EV/Sales or EV/ARR, with the Rule of 40. That covers SaaS, marketplaces (on net revenue), early e-commerce and pre-profit streaming.
3. Is it capital-heavy and cyclical or price-taking? Then use EV/EBITDA on mid-cycle earnings plus replacement cost, and never trailing P/E. Commodities, autos and owned hotels go here. So do airlines, with the lease twist that turns it into EV/EBITDAR.
4. Is it a stable, profitable compounder? Then use P/E, EV/EBIT, FCF yield. Think FMCG, IT services, generic pharma, asset-light hotel fees, mature software, most quality consumer names.
5. Is it a real-asset yield vehicle? Then use NAV, FFO/AFFO, cap rate. That's REITs. For developers, use P/NAV and EV/bookings, because the P&L lags reality by years.
6. Then, always: compute the sector KPIs, and close with ROIC > WACC, durable and reinvestable? If the multiple and the ROIC test disagree, trust the ROIC test and ask why the multiple is where it is. Usually the market is pricing a cycle turn or a durability change, and you should have a view on it.

Phase 5 compresses into those six questions. The rest of the week makes you fast at them.

Common mistakes & how experts think differently

1. Valuing an airline (or any deep cyclical) on trailing P/E. It's the most expensive error in the sector. Peak earnings at a normal multiple overvalue wildly, and the low P/E on record profits is a sell signal. Experts value airlines on through-cycle EV/EBITDAR. They check survival first (days-cash, lease-adjusted leverage) and valuation second.
2. Ignoring capitalised leases in airline leverage and EV. A carrier can look ungeared on borrowings while carrying tens of thousands of crore in lease liabilities. Since IFRS 16 they're in reported debt. Under US GAAP operating leases, you have to add the disclosed liability yourself. Net debt including leases is the only number to trust, and EV has to match EBITDAR.
3. Being seduced by a high load factor. A 92% load factor feels like a triumph, and it can come with a loss if yields were cut to fill the seats. Load factor is necessary but not sufficient. Experts read RASK (yield × load factor + ancillary), because it's the only summary that catches "fullness bought with price."
4. Blending an owned-hotel business with an asset-light one. Put a single multiple on a mixed hotel company and you misvalue both halves. The fee stream deserves ~30× and the owned cyclical asset ~12 to 14×. Experts value hotels sum-of-the-parts and treat fee-income growth as the de-risking story.
5. Confusing RevPAR growth from ADR with growth from occupancy. RevPAR = ADR × occupancy, but the two kinds of growth aren't equal. ADR-led growth with falling occupancy is fragile and often late-cycle. Occupancy-led growth still has room to price up. Always split RevPAR into its two drivers before you cheer.
6. Building owned hotels at the RevPAR peak. The supply lag is 3 to 5 years, so debt-funded construction when the numbers look best delivers rooms into the glut. Experts watch the supply pipeline as a leading indicator. They expand asset-light near peaks and buy owned assets in troughs.
7. Reading a streamer's subscriber count without ARPU, churn, and engagement. 100 million subs at $4 ARPU and 6% churn are worth a fraction of 100 million subs at $14 ARPU and 2% churn. Experts value per-sub only against churn and ARPU. They watch engagement as the leading churn signal, before the churn number moves.
8. Treating content spend as ignorable "investment." Content is a streamer's COGS, and cash content spend is the FCF drag. Spend that has to outgrow revenue to hold subs is a treadmill, and a treadmill builds no moat. Experts watch the gap between cash spend and amortisation. They also watch content write-offs, which are admissions of wasted spend.
9. Picking the multiple before understanding the model. The novice reaches for P/E by reflex. The expert first classifies the business (financial / capital-light growth / cyclical price-taker / compounder / yield vehicle) and derives the lens from the model, using the decision procedure above.
10. Forgetting the master check. Any lens can flatter a bad business for a while. The expert always closes by asking: ROIC > WACC, durable and reinvestable? Every sector KPI is evidence on that one question. None of it is trivia to recite.

Worked examples

WE1: Airline unit economics and EV/EBITDAR (an IndiGo-style LCC survivor)

Meet NimbusJet (synthetic, calibrated to ~FY24 IndiGo scale; pull IndiGo's actual annual report and investor presentation for real figures). It flies ASK 150 billion seat-km and sells RPK 127.5 billion. Passenger revenue is ₹58,000 cr and ancillary revenue ₹10,000 cr, for total operating revenue of ₹68,000 cr. Total operating cost is ₹60,000 cr, of which fuel is ₹24,000 cr. EBITDA is ₹18,000 cr, and aircraft lease-related cost (rentals / ROU depreciation + lease interest) is ₹9,000 cr. On the balance sheet, gross debt + lease liabilities come to ₹48,000 cr against cash & equivalents of ₹35,000 cr. Market cap is ₹1,60,000 cr.

Compute the panel:

• Load factor = RPK / ASK = 127.5 / 150 = 85.0%, comfortably above an ~80 to 85% LCC breakeven.
• Yield (passenger) = passenger revenue / RPK = ₹58,000 cr / 127.5 bn km = ₹4.55 per RPK.
• RASK (total unit revenue) = total revenue / ASK = ₹68,000 cr / 150 bn = ₹4.53 per ASK. Check the passenger identity: PRASK = passenger rev/ASK = 58,000/150 = ₹3.87 = yield × load = 4.55 × 0.85 ✓.
• CASK = total cost / ASK = ₹60,000 cr / 150 bn = ₹4.00 per ASK.
• CASK ex-fuel = (60,000 − 24,000)/150 = 36,000/150 = ₹2.40 per ASK. That's the controllable cost, and having the lowest in its market is the moat.
• Unit profit = RASK − CASK = 4.53 − 4.00 = ₹0.53 per ASK. Take the profit off the statements instead of off that figure: 68,000 − 60,000 = ₹8,000 cr operating profit, an 11.8% margin and a good airline year. Multiply the displayed 0.53 by 150 bn ASK and you get ₹7,950 cr, ₹50 cr light, because RASK is really 4.5333 and the third decimal is worth half a percent of the profit. That's the airline lesson in miniature. The margin is the difference between two large numbers, so it inherits every rounding you commit upstream.
• Fuel % of opex = 24,000/60,000 = 40%. Ancillary % = 10,000/68,000 = 14.7%, a healthy LCC extras engine.
• Days-cash = cash / (daily opex) = 35,000 / (60,000/365) = 35,000 / 164.4 = 213 days. That's deep liquidity, enough to survive a shock.
• Net debt incl. leases = 48,000 − 35,000 = ₹13,000 cr, well controlled.
• EV/EBITDAR = (market cap + net debt incl. leases) / (EBITDA + lease cost) = (1,60,000 + 13,000) / (18,000 + 9,000) = 1,73,000 / 27,000 = 6.4×.

Read it back. NimbusJet has the lowest CASK ex-fuel and an 85% load factor with healthy yield, so RASK − CASK is positive. It holds 213 days of cash and carries modest lease-adjusted leverage. This is the rare airline that can earn its cost of capital. It's the survivor.

WE2: The graveyard contrast (a Jet/Kingfisher-style legacy carrier)

Now the other end. TailwindAir (synthetic, calibrated to a distressed Indian full-service carrier) flies ASK 40 billion and sells RPK 31.2 billion. Total revenue is ₹16,800 cr against total cost of ₹18,800 cr (fuel ₹7,000 cr). It has ₹900 cr of cash, and accumulated losses have driven net worth to −₹5,000 cr. The fleet order book is expanding and fuel is unhedged.

• Load factor = 31.2/40 = 78%, below a ~82% breakeven.
• RASK = 16,800/40 = ₹4.20. CASK = 18,800/40 = ₹4.70. CASK ex-fuel = (18,800−7,000)/40 = 11,800/40 = ₹2.95, well above NimbusJet's ₹2.40. That's slack cost control or lost scale, red flag #2.
• Unit profit = 4.20 − 4.70 = −₹0.50 per ASK → × 40 bn = −₹2,000 cr operating loss.
• Days-cash = 900 / (18,800/365) = 900/51.5 = 17 days. That's weeks of cover instead of months, red flag #3.
• Negative net worth −₹5,000 cr + 17 days cash + unhedged fuel + fleet growth into a loss = every airline red flag firing at once.

Same industry, same KPIs. Put NimbusJet's ₹0.53 unit profit and 213 days of cash next to TailwindAir's ₹0.50 unit loss and 17 days of cash, and you're looking at the entire distance between compounding and the graveyard. An EV/EBITDAR "cheapness" screen on TailwindAir would be meaningless. The survival read comes first.

WE3: Hotels: owned vs asset-light, valued sum-of-the-parts

Apex Hotels (synthetic, IHCL-style mixed model) has two segments.

Owned segment: 5,000 owned rooms, ADR ₹12,000, occupancy 72%.

• RevPAR = ADR × occupancy = 12,000 × 0.72 = ₹8,640.
• Room revenue = RevPAR × rooms × 365 = 8,640 × 5,000 × 365 = ₹1,577 cr. Add F&B/banquets and total owned revenue ≈ ₹3,000 cr. At a 33% margin, owned EBITDA = ₹1,000 cr.
• GOPPAR: if gross operating profit is ₹1,400 cr over 5,000 rooms × 365 = 1.825 mn available room-nights, then GOPPAR = 1,400 cr / 1.825 mn = ₹7,671 per available room.
• RevPAR Index: Apex's owned RevPAR is ₹8,640 against a comp-set RevPAR of ₹8,000, so RGI = 8,640/8,000 × 100 = 108. Apex is taking share.

Managed/franchised segment: 15,000 managed rooms earn fee income (base + incentive) of ₹400 cr. At a 55% fee EBITDA margin, that's fee EBITDA of ₹220 cr, on near-zero capital.

SOTP valuation:

• Owned (capital-heavy, cyclical) at EV/EBITDA 14× → 14 × 1,000 = ₹14,000 cr.
• Fee (asset-light, annuity, high-ROIC) at 30× fee EBITDA → 30 × 220 = ₹6,600 cr.
• Enterprise value = 14,000 + 6,600 = ₹20,600 cr. Take off net debt of ₹3,000 cr and you're left with equity value of ₹17,600 cr.

Look at what just happened. The fee segment is one-fifth of EBITDA (220 of 1,220) but nearly one-third of value. That's because ₹1 of asset-light fee EBITDA is worth more than ₹1 of cyclical owned EBITDA. Fee EBITDA earns a higher ROIC and recurs more, and it's less cyclical. Blend the two at one multiple and you misprice the company. A pure asset-light peer, Marriott-style, would be all fee EBITDA, which explains its premium market multiple and its stock-buyback capital return.

WE4: Streaming: churn, content, and per-sub (Netflix-style vs a burning challenger)

StreamOne (synthetic, Netflix-scale, in harvest) has 260 million subscribers paying a blended ARPU of $11.50/month, with monthly churn of 2.3%. Cash content spend is $17bn and content amortisation $14bn. The operating margin is 22%, enterprise value is $250bn, and engagement runs at ~2 hours/day/member.

• Revenue = 260 mn × $11.50 × 12 = $35.9bn.
• Average customer life ≈ 1/monthly churn = 1/0.023 ≈ 43 months (~3.6 years).
• Content as % of revenue (amortisation) = 14/35.9 = 39%.
• Cash-vs-amortisation gap = 17 − 14 = +$3bn. It's still investing ahead of the P&L, but modestly, which fits a mature streamer instead of a land-grabber.
• Operating profit = 22% × 35.9 = $7.9bn. FCF is positive, so harvest is confirmed.
• EV per subscriber = $250bn / 260 mn = $962/sub.

Then there's BurnCo (synthetic challenger, still land-grabbing). It has 50 million subs at an ARPU of $7.00 and monthly churn of 5.5%, and it spends $8bn of cash on content. Revenue = 50 mn × 7 × 12 = $4.2bn.

• Average life ≈ 1/0.055 ≈ 18 months, less than half StreamOne's.
• Content spend ÷ revenue = 8/4.2 = 190%. BurnCo spends nearly twice its revenue on content just to compete. That's the treadmill, red flag #3.
• Operating loss (content + marketing > revenue), and negative FCF.
• Per-sub cross-check: applying StreamOne's $962/sub to BurnCo would be indefensible. BurnCo's subs churn 2.4× faster and earn 39% less ARPU, so each one is worth a fraction as much. The per-sub number only means something conditioned on ARPU and churn.

That contrast is the media playbook in miniature. The headline unit is the same (a subscriber), but the value is radically different, because unit economics (ARPU, churn, content cost) decide whether subscribers are an asset or a liability.

WE5: A US low-cost carrier in cents-per-ASM (the American convention)

SkyValue Airlines (synthetic, calibrated to a large US LCC ~Southwest scale: pull the actual 10-K; US carriers report per available seat MILE, in cents instead of per-ASK in rupees). It flies ASM (Available Seat Miles) of 160 bn and sells RPM (Revenue Passenger Miles) of 133.6 bn. Total operating revenue is $26,000m, made up of passenger $23,000m + ancillary/other $3,000m. Total operating cost is $24,800m, of which fuel is $6,800m. EBITDA is $3,400m and aircraft rent / lease-related cost $1,000m. SkyValue holds cash of $9,000m against gross debt + capitalised leases of $11,000m. Its market cap is $19,000m.

Compute the panel. US carriers quote unit metrics in cents per ASM:

• Load factor = RPM / ASM = 133.6 / 160 = 83.5%.
• Yield = passenger revenue / RPM = $23,000m / 133.6 bn = 17.22¢ per RPM.
• RASM (unit revenue) = total revenue / ASM = $26,000m / 160 bn = 16.25¢ per ASM. Check the passenger identity: PRASM = passenger rev/ASM = 14.375¢ = yield × load = 17.22 × 0.835 ✓.
• CASM = total cost / ASM = $24,800m / 160 bn = 15.50¢ per ASM.
• CASM ex-fuel = ($24,800 − $6,800)m / 160 bn = 11.25¢ per ASM, the controllable cost.
• Unit profit = RASM − CASM = 16.25 − 15.50 = 0.75¢ per ASM → × 160 bn ASM = $1,200m operating profit. That's a ~4.6% margin, the same razor-thinness as the rupee carrier of WE1, in a different currency.
• Fuel % of opex = 6,800/24,800 = 27.4%, notably lower than an Indian carrier's ~40%. US jet fuel is far less heavily taxed, and there's no rupee-vs-dollar mismatch on dollar-priced fuel. Ancillary % = 3,000/26,000 = 11.5%.
• Days-cash = 9,000 / (24,800/365) = 132 days, which is months of cover.
• Net debt incl. leases = 11,000 − 9,000 = $2,000m. EV/EBITDAR = (19,000 + 2,000) / (3,400 + 1,000) = 21,000 / 4,400 = 4.8×.

Read it and you'll see the same machine as WE1's Indian LCC. Unit profit is thin, fuel is the swing factor, EBITDAR is the right multiple, and days-cash is deep. The American reporting convention is cents per available seat mile, and the cost mix differs. Fuel is a smaller share because of lighter taxes and no FX mismatch. The lens doesn't care about currency. Only the units and the tax/FX backdrop change.

WE6: India streaming: the low-ARPU corner (JioStar-style volume vs a premium SVOD)

BharatStream (synthetic, calibrated to a large India ad-and-cricket-driven streamer, JioStar/Hotstar-style; pull the actual filings / JV disclosures) versus PrimeReel (synthetic, a premium India SVOD of curated originals). All figures in ₹.

BharatStream: 300 million subscribers on a blended ARPU of ₹55/month, with monthly churn of 4.5%. Cash content spend is ₹18,000 cr (sports rights dominate), and content amortisation is ₹15,000 cr. Enterprise value is ₹90,000 cr.

• Revenue = 300m × ₹55 × 12 = ₹19,800 cr.
• Average customer life ≈ 1 / 0.045 = 22 months, so lifetime revenue/sub = 55 × 22.2 = ₹1,222.
• Content as % of revenue (amortisation) = 15,000/19,800 = 76%, and cash content ÷ revenue = 18,000/19,800 = 91%. That's a punishing content load driven by cricket/IPL rights (the treadmill, red flag #3), with cash spend still ₹3,000 cr above amortisation (front-loading).
• EV per subscriber = ₹90,000 cr / 300m = ₹3,000/sub.

PrimeReel: 40 million subscribers at an ARPU of ₹180/month and monthly churn of 2.2%, with an enterprise value of ₹60,000 cr.

• Revenue = 40m × ₹180 × 12 = ₹8,640 cr.
• Average life ≈ 1 / 0.022 = 45 months, so lifetime revenue/sub = 180 × 45.5 = ₹8,182.
• EV per subscriber = ₹60,000 cr / 40m = ₹15,000/sub.

BharatStream has 7.5× the subscribers, yet each PrimeReel sub is worth ~6.7× more in lifetime revenue (₹8,182 vs ₹1,222). Higher ARPU and less than half the churn compound into a far longer life. The market pays ₹15,000 vs ₹3,000 per sub (5×), and it's right to. Apply BharatStream's ₹3,000/sub to PrimeReel and you'd badly undervalue it. Put PrimeReel's ₹15,000/sub on BharatStream's spiky, low-ARPU base and the result would be indefensible. This is the "instructive opposite corner" from the streaming section: the same KPIs (subs, ARPU, churn, content) with the dials set to opposite extremes. A raw subscriber count means nothing until you condition it on ARPU and churn.

Practice set

Work each one in full before you read the solution. Numeric answers are to ±3% unless stated. Synthetic companies are labelled, and real names are teaching specimens.

P1 (guided — airline unit economics). CedarAir reports ASK 80 bn, RPK 66.4 bn, total revenue ₹32,000 cr, total cost ₹30,400 cr (fuel ₹11,400 cr). Compute load factor, RASK, CASK, CASK ex-fuel, and unit profit. Is it above or below an ~83% breakeven?

Solution. Load factor = 66.4/80 = 83.0%, right at breakeven. RASK = 32,000/80 = ₹4.00. CASK = 30,400/80 = ₹3.80. CASK ex-fuel = (30,400−11,400)/80 = 19,000/80 = ₹2.375. Unit profit = 4.00 − 3.80 = ₹0.20/ASK → × 80 bn = ₹1,600 cr operating profit. That's a 5% margin, as thin as the operating-leverage warning predicts. At 83% CedarAir is only just above breakeven, and a 2-point load-factor slip or a fuel uptick wipes out the profit.

P2 (guided — days-cash and lease-adjusted leverage). CedarAir holds cash ₹4,000 cr; gross debt ₹6,000 cr; lease liabilities ₹22,000 cr; EBITDA ₹5,600 cr; lease cost ₹3,000 cr. Compute days-cash, net debt including leases, and EV/EBITDAR if market cap is ₹30,000 cr.

Solution. Days-cash = 4,000 / (30,400/365) = 4,000/83.3 = 48 days. That's thin, weeks of cover instead of months, and a survival concern. Net debt incl. leases = 6,000 + 22,000 − 4,000 = ₹24,000 cr. EV = 30,000 + 24,000 = ₹54,000 cr, and EBITDAR = 5,600 + 3,000 = ₹8,600 cr, so EV/EBITDAR = 54,000/8,600 = 6.3×. Notice how leases dominate the leverage. Leave them out and "net debt" is only ₹2,000 cr, so the airline looks nearly debt-free. That's the trap this KPI exists to catch.

P3 (guided — RevPAR decomposition). A hotel's RevPAR rose from ₹6,000 to ₹6,600 (+10%). In year 1: ADR ₹8,000, occupancy 75%. In year 2: ADR ₹9,167, occupancy 72%. Decompose the RevPAR growth into ADR and occupancy effects and comment.

Solution. Check first: Y1 RevPAR = 8,000 × 0.75 = ₹6,000 ✓; Y2 = 9,167 × 0.72 = ₹6,600 ✓. ADR grew +14.6% (9,167/8,000) while occupancy fell from 75% to 72%, which is −3 percentage points or −4% relative. So all the RevPAR growth, and more, came from rate, partly offset by falling occupancy. Comment: ADR-led growth with declining occupancy is fragile and often late-cycle. The hotel is pricing out volume, and the next demand dip lands on occupancy that's already thinning. Prefer occupancy-led growth, which leaves room to raise rate later, to rate-led growth, which is running out of pricing headroom.

P4 (independent — GOPPAR and RevPAR Index). BeachResort: 800 rooms; ADR ₹15,000; occupancy 68%; gross operating profit ₹190 cr for the year. Comp-set RevPAR is ₹9,500. Compute RevPAR, GOPPAR, and the RevPAR Index; interpret.

Solution. RevPAR = 15,000 × 0.68 = ₹10,200. Available room-nights = 800 × 365 = 292,000. GOPPAR = 190 cr / 292,000 = ₹1,90,00,00,000 / 292,000 = ₹6,507 per available room. RevPAR Index = 10,200/9,500 × 100 = 107, so BeachResort is outperforming its comp set by ~7% and gaining share. Interpretation: strong relative positioning (RGI >100) and solid GOPPAR conversion. The moderate 68% occupancy at a high ADR fits a rate-led luxury resort. Watch occupancy for cycle signals.

P5 (independent — owned vs asset-light multiple choice). Two hotel companies each earn ₹500 cr EBITDA. Company O owns all its hotels (capital-heavy, cyclical, ROIC ~10%, net debt/EBITDA 3.5×). Company M earns it entirely as management/franchise fees (asset-light, ROIC ~45%, minimal debt). Which deserves the higher EV/EBITDA, and roughly how different should the multiples be? Explain in ROIC terms.

Solution. Company M deserves the far higher multiple. The EBITDA is the same, but M's is asset-light, recurring, less cyclical, and earns ~45% ROIC, so its growth is cheap and value-accretive. O's is capital-heavy, cyclical, levered, and earns ~10% ROIC, near or below its cost of capital, so growth barely adds value. Illustratively, M might trade at ~25 to 30× EBITDA and O at ~10 to 14×, a 2 to 3× gap on identical EBITDA. The driver is the master formula: value rises with the ROIC-over-WACC spread and with reinvestable growth. M has a wide spread and cheap growth, while O has a thin spread and expensive growth. That's why mixed operators have to be valued sum-of-the-parts.

P6 (independent — churn to lifetime and value). Service A: ARPU $12/month, monthly churn 2%. Service B: ARPU $9/month, monthly churn 5%. Compute each service's average customer life and lifetime revenue per subscriber. Which subscriber is worth more, and why might B still trade at a lower EV/sub?

Solution. Average life ≈ 1/churn. A: 1/0.02 = 50 months, and lifetime revenue = 12 × 50 = $600/sub. B: 1/0.05 = 20 months, and lifetime revenue = 9 × 20 = $180/sub. So Service A's subscriber is worth ~3.3× more in lifetime revenue on only 33% higher ARPU, because low churn compounds into long life. B should trade at a much lower EV/sub. If it doesn't, it's overvalued. Never compare per-sub values without conditioning on churn and ARPU.

P7 (independent — the content treadmill). StreamCo: revenue $6bn, growing 15%; cash content spend $5.5bn, growing 22%; content amortization $4.5bn; operating loss −$0.4bn. Diagnose using the media red-flag panel.

Solution. Content spend ($5.5bn) is 92% of revenue and growing faster than revenue (22% vs 15%). That's the treadmill, red flag #3. StreamCo has to keep raising content spend faster than sales just to grow subs, so scale isn't producing operating leverage. Cash spend ($5.5bn) exceeds amortisation ($4.5bn) by $1bn, so it's still front-loading, consistent with land-grab instead of harvest. The operating loss confirms it hasn't reached the scale crossover. Verdict: value it on EV/Sales / per-sub instead of earnings. The investment case hinges entirely on whether spend growth slows below revenue growth (the path to FCF) before cash runs low. Check the balance sheet and the trajectory of net adds and engagement.

P8 (independent — the value-trap check). A steelmaker trades at 4.5× trailing P/E on record earnings; an airline at 5× EV/EBITDAR on a boom year; an FMCG compounder at 44× P/E with 42% ROCE. Rank them from most to least dangerous to buy on the multiple alone, and state the master check.

Solution. The most dangerous is the steelmaker at 4.5× P/E on record (peak) earnings. It's the classic cyclical trap, because the low P/E marks the top and earnings are about to mean-revert. Next comes the airline at 5× EV/EBITDAR on a boom year. It's the same cyclical, operating-leverage trap, plus the survival questions (days-cash, lease leverage) that the multiple ignores. Least dangerous, and possibly even cheap, is the FMCG compounder at 44×. That's a high multiple, but on a durable ~42% ROCE that reinvests accretively, so the ROIC>WACC spread and its persistence can justify it. The master check under all three: ROIC > WACC, durable and reinvestable? A cheap multiple on a capital-destroyer is a trap, and a rich multiple on a high-return reinvestor can be cheap.

P9 (timed, 6 min — right-lens mini). For each, name the lens, top-3 KPIs, and the one red flag to check: (a) an Indian private bank growing loans 20%; (b) a US enterprise SaaS firm, 35% ARR growth, not yet FCF-positive; (c) a commercial-office REIT.

Solution. (a) P/B vs ROE · NIM, GNPA/slippage & PCR, CASA (and ROA/CRAR) · red flag: rising SMA-2/slippages (asset quality deteriorating behind the loan growth). (b) EV/ARR (EV/Sales), Rule of 40 · NRR/NDR, CAC payback, Rule of 40 (& gross margin) · red flag: NRR <100% or billings/cRPO decelerating below reported ARR. (c) P/NAV, FFO/AFFO, cap rate · NOI & same-store NOI, occupancy & WALE, AFFO payout (& LTV) · red flag: AFFO payout >90% (distribution-cut risk) or occupancy slide with negative re-leasing spreads.

P10 (timed, 8 min — full airline mini-teardown). MonsoonAir: ASK 100 bn, RPK 88 bn, revenue ₹40,000 cr (ancillary ₹5,000 cr), cost ₹41,000 cr (fuel ₹16,400 cr), cash ₹2,500 cr, net worth −₹2,000 cr, fleet growing 25%/yr, no fuel hedges. Compute the panel, then give a 4-line verdict.

Solution. Load factor = 88/100 = 88%, which is high. RASK = 40,000/100 = ₹4.00 and CASK = 41,000/100 = ₹4.10, so unit profit = −₹0.10/ASK → −₹1,000 cr loss. CASK ex-fuel = (41,000−16,400)/100 = ₹2.46. Ancillary % = 5,000/40,000 = 12.5%. Days-cash = 2,500/(41,000/365) = 2,500/112.3 = 22 days. Verdict: (1) A high load factor (88%) but a unit loss. That's fullness bought with price, with yields too low to cover cost (red flag #1). (2) 22 days of cash + negative net worth. That's pre-insolvency liquidity (red flag #3). (3) Fleet +25% into a loss-making position, with unhedged fuel. Capacity is being dumped into weak economics, fully exposed to a crude spike (red flags #5 and #4). (4) This is a graveyard candidate. It's no value opportunity, and no multiple matters until survival is secured. Put it next to WE1's NimbusJet and you can feel the difference careful management makes.

P11 (timed, 8 min — hotel SOTP). Heritage Hotels: owned segment EBITDA ₹700 cr (ROIC ~9%, cyclical); fee segment EBITDA ₹180 cr (ROIC ~48%); net debt ₹4,500 cr; owned RevPAR ₹7,200 vs comp set ₹7,500. Value the equity (owned at 13×, fee at 28×) and comment on the RevPAR Index.

Solution. Owned EV = 13 × 700 = ₹9,100 cr. Fee EV = 28 × 180 = ₹5,040 cr. Total EV = ₹14,140 cr, and less net debt of ₹4,500 cr that leaves equity of ₹9,640 cr. RevPAR Index = 7,200/7,500 × 100 = 96 (<100), so Heritage is losing share to the comp set. That's a brand or asset-quality flag worth investigating, and the market cycle may be masking it. Notice the fee segment is 20% of EBITDA (180 of 880) but 36% of EV (5,040 of 14,140). It's the asset-light premium again. Strategic read: growing the fee segment de-risks the cycle and lifts group ROIC, and the sub-100 RGI on the owned book is the thing to fix.

P12 (timed, 10 min — cross-sector synthesis). You are handed five one-line businesses. Assign each a lens and the single most important KPI, then name which two are most likely value traps on a low trailing P/E. (i) A generic-drug maker with 60% US sales and an FDA Warning Letter on its main plant. (ii) An aluminium smelter at record profits, 3rd-quartile cash costs, expanding capacity with debt. (iii) A life insurer growing APE 18%. (iv) A fee-based hotel brand with a large room pipeline. (v) A telecom operator post-tariff-hike.

Solution. (i) P/E / EV/EBITDA · key KPI: USFDA plant status (a Warning Letter that leads to a possible Import Alert can wipe out US sales), plus US price erosion. (ii) EV/EBITDA on mid-cycle + replacement cost · key KPI: cash-cost quartile (3rd quartile = bleeds in the trough); never trailing P/E. (iii) P/EV, VNB · key KPI: VNB margin (is the growth profitable or ULIP-mix-driven?), plus persistency. (iv) P/E / EV/EBITDA on fees · key KPI: net room/pipeline growth (and fee margin, RevPAR Index). (v) EV/EBITDA · key KPI: ARPU (did the hike stick?), plus net debt/EBITDA and churn. Most likely value traps on a low trailing P/E: (ii) the smelter at record profits. Its earnings are peak-cycle, the low P/E is a sell signal, and debt-funded expansion at the top compounds the danger. And (i) the generic maker, if a low P/E reflects a market pricing in the plant risk and base erosion. That's a cheap multiple masking a deteriorating earnings base. The master check resolves both: through-cycle / post-risk ROIC > WACC?

P13 (guided — a US carrier in cents-per-ASM). SkyRoute (synthetic US LCC) reports ASM 40 bn (seat-miles), RPM 33.6 bn, total revenue $6,400m (ancillary $800m), total cost $6,080m (fuel $1,600m). Compute load factor, RASM, CASM, CASM ex-fuel, and unit profit (in cents per ASM and in $m). How does its fuel share compare with an Indian carrier's, and why?

Solution. Load factor = 33.6/40 = 84.0%. RASM = $6,400m / 40 bn = 16.0¢/ASM. CASM = $6,080m / 40 bn = 15.2¢/ASM. CASM ex-fuel = ($6,080 − $1,600)m / 40 bn = 11.2¢/ASM. Unit profit = 16.0 − 15.2 = 0.8¢/ASM → × 40 bn = $320m operating profit, a 5.0% margin. Fuel % of opex = 1,600/6,080 = 26.3%, against ~40% for an Indian carrier. US jet fuel carries far lighter taxes and there's no rupee-dollar mismatch on dollar-priced fuel, so the cost the world controls bites less. It's the same engine as WE1/WE5, quoted in cents per available seat mile, the US convention.

P14 (guided — the RevPAR supply-lag cycle turn). A hotel micro-market peaks at ADR ₹11,000 and 78% occupancy. Over the next three years a construction pipeline lifts room supply +30% while total room-nights sold grow only +6%, and pricing power fades so ADR slips 8%. Estimate the new occupancy and RevPAR, and call the turn.

Solution. Peak RevPAR = 11,000 × 0.78 = ₹8,580. Occupancy = sold ÷ available. Sold grows ×1.06 while available grows ×1.30, so new occupancy = 0.78 × 1.06 / 1.30 = 63.6%. New ADR = 11,000 × 0.92 = ₹10,120. New RevPAR = 10,120 × 0.636 = ₹6,436, a −25% fall even though room-nights sold actually rose 6%. That's the turn. Supply overwhelming demand collapses occupancy and pricing together, and the multi-year supply lag is the cyclicality you met in the hotel section. Debt-funded owned building into that ₹8,580 peak (red flag #3) delivers rooms straight into the glut it helped create.

P15 (independent — EV per subscriber, conditioned on ARPU and churn). IndiPlay (synthetic, India): 120m subs, EV ₹36,000 cr, ARPU ₹60/month, churn 4%/month. NorthStar (synthetic, US): 90m subs, EV $72bn, ARPU $12/month, churn 2%/month. Compute each EV/sub and each subscriber's lifetime revenue (use ₹85/$, as of mid-2026). Is IndiPlay's low EV/sub a bargain?

Solution. IndiPlay EV/sub = ₹36,000 cr / 120m = ₹3,000 (≈ $35), and lifetime revenue = 60 ÷ 0.04 = ₹1,500/sub (≈ $18). NorthStar EV/sub = $72bn / 90m = $800, and lifetime revenue = 12 ÷ 0.02 = $600/sub. So NorthStar's sub costs ~23× more (800 vs 35) but delivers ~34× more lifetime revenue (600 vs 18). On a lifetime-value basis NorthStar's subs are, if anything, the cheaper of the two. IndiPlay's low EV/sub is the market correctly pricing a low-ARPU, high-churn subscriber, so it's no bargain. Per-sub numbers mean nothing until you condition them on ARPU and churn.

P16 (independent — EV/EBITDAR with US-GAAP operating leases). A US carrier: market cap $12,000m; gross borrowings $5,000m; disclosed operating-lease liability $9,000m; cash $6,000m; EBITDA $2,800m; aircraft rent $1,400m. Compute net debt including capitalised leases and EV/EBITDAR. What would a lease-blind analyst wrongly conclude?

Solution. Net debt incl. leases = 5,000 + 9,000 − 6,000 = $8,000m. EV = 12,000 + 8,000 = $20,000m, and EBITDAR = 2,800 + 1,400 = $4,200m, so EV/EBITDAR = 20,000 / 4,200 = 4.8×. A lease-blind analyst nets only borrowings against cash: 5,000 − 6,000 = −$1,000m ("net cash"). They call the airline nearly debt-free and, using bare EBITDA, compute 11,000 / 2,800 = 3.9×. That understates both leverage and the multiple. Treating a lease-financed fleet as unencumbered is the most common airline mistake there is. Under US GAAP you have to add the disclosed operating-lease liability and match it with EBITDAR.

P17 (independent — hotel operating leverage / flow-through). An owned hotel earns GOPPAR ₹2,400 at RevPAR ₹6,000, with variable costs ~10% of room revenue and the remainder fixed at ₹3,000 per available room-night. RevPAR rises 15% to ₹6,900. Compute the new GOPPAR, its % change, and the incremental flow-through. What does this say about owned-hotel cyclicality?

Solution. New GOPPAR = 6,900 − (10% × 6,900) − 3,000 = 6,900 − 690 − 3,000 = ₹3,210, up from ₹2,400. That's a +33.75% jump on a 15% RevPAR rise, a ~2.25× amplification. Incremental flow-through = (3,210 − 2,400) ÷ (6,900 − 6,000) = 810 / 900 = 90%. With costs largely fixed, ~90 paise of every extra rupee of RevPAR falls to gross operating profit. And that operating leverage runs both ways. A 15% RevPAR fall would crater GOPPAR by ~34%. That's why owned hotels are so cyclical, and why you have to value them on mid-cycle EBITDA, never on a peak or trough year.

P18 (timed, 10 min — cross-sector lens + traps). Assign each a lens and its single most important KPI, then name the two most dangerous to buy on a low trailing multiple alone. (i) An Indian LCC at a "cheap" 5× EV/EBITDAR on a bumper post-pandemic travel year, fleet +20%. (ii) A US asset-light hotel-fee brand at 30× earnings with a long room pipeline. (iii) An Indian low-ARPU streamer, net adds +30%, still loss-making, content spend 120% of revenue. (iv) A US legacy airline at 4× trailing P/E on record profits. (v) An Indian owned-hotel operator building trophy assets with debt at a RevPAR peak.

Solution. (i) EV/EBITDAR on mid-cycle · unit profit (RASK−CASK) [and days-cash] · trap-worthy: a peak-year "cheap" multiple with fleet growth into the cycle. (ii) P/E (or EV/EBITDA on fees) · net room/pipeline growth · not a trap, just a durable high-ROIC fee compounder. (iii) EV/Sales, per-sub · net adds read with ARPU & churn · content spend >100% of revenue is the treadmill, so value it on sales instead of earnings. (iv) EV/EBITDAR on mid-cycle, never trailing P/E · through-cycle unit profit & days-cash · the low P/E on record profits is a peak/sell signal. (v) EV/EBITDA on mid-cycle · supply pipeline & occupancy · debt-funded owned build at the peak (red flag #3). Most dangerous on the low trailing multiple: (iv) the legacy airline at 4× P/E on record profits and (i) the LCC at 5× EV/EBITDAR on a bumper year. Both are deep cyclicals whose lowest multiples appear at the top, and both carry survival/capacity risks the multiple ignores. Master check for both: through-cycle ROIC > WACC, survival first.

Applied mini-project: the capstone (the module gate)

This is the Phase 5 capstone, and it certifies competency C6. It has two parts, and you need both to pass. Do Part A closed-book and grade yourself. Do Part B on real filings and grade yourself against the rubric.

Part A: the right-lens exam (20 vignettes)

For each business below, write down three things: (1) the primary valuation lens/multiple, (2) the top-3 KPIs you'd compute first, and (3) the one red flag you'd check before anything else. Give yourself ~45 seconds each, the same pace as the app's right-lens engine and the Phase 8 teardown reflex. Pass mark: ≥17/20 correct on lens + a defensible top-3 and red flag. Cover the answer key first.

1. An Indian private-sector bank growing its loan book ~20% a year, retail-heavy.
2. A US enterprise-software company, ARR growing 35%, not yet free-cash-flow positive.
3. An Indian FMCG staples major earning ~40% ROCE with a vast rural distribution network.
4. A global integrated steelmaker reporting record profits at a low trailing P/E.
5. An Indian NBFC in vehicle finance funded heavily by short-term commercial paper.
6. A US low-cost airline, single fleet type, high load factors.
7. A global hotel company that owns almost no hotels and earns management/franchise fees.
8. An Indian private life insurer, APE up 18%, shifting toward ULIPs.
9. A US streaming pure-play that recently turned profitable and FCF-positive.
10. An Indian large-cap IT services exporter to global banks.
11. A post-RERA Indian residential real-estate developer.
12. An Indian commercial-office REIT with marquee tech tenants.
13. A US generic-drug maker with 60% of sales in the US and one key manufacturing plant.
14. An Indian two-wheeler OEM heading into a festive season.
15. An Indian telecom operator after an industry tariff hike.
16. An engineering & construction (EPC) contractor bidding for infrastructure projects.
17. An owned-hotel-heavy operator carrying trophy city-centre assets, expanding with debt.
18. An online services marketplace still subsidising both sides to build liquidity.
19. A general (P&C) insurer growing premiums fast, with large investment gains this year.
20. An auto-ancillary supplier whose kit is almost entirely for internal-combustion engines.

Answer key (lens · top-3 KPIs · red flag):

1. P/B vs ROE · NIM, GNPA/slippage & PCR, CASA (ROA, CRAR) · rising SMA-2/slippages behind the loan growth.
2. EV/ARR (EV/Sales), Rule of 40 · NRR/NDR, CAC payback, Rule of 40 (gross margin) · NRR <100% or billings/cRPO decelerating below reported ARR.
3. P/E, EV/EBITDA (premium) · volume vs value growth, gross/EBITDA margin, A&P % & ROCE · value growth on flat/negative volumes with A&P cut to prop margin.
4. EV/EBITDA on mid-cycle + replacement cost · cash-cost quartile, EBITDA/tonne, net debt/EBITDA · never trailing P/E, because the low P/E on record earnings is a peak/sell signal; debt-funded expansion at the top.
5. P/B vs ROE (P/E) · spread/NIM, ALM gap, gearing & credit cost (cost-of-funds mix) · ALM mismatch + CP reliance (the IL&FS/DHFL failure mode).
6. EV/EBITDAR · unit profit (RASK−CASK), CASK ex-fuel, load factor vs breakeven (days-cash) · high load factor with falling yields; thin days-cash / negative net worth.
7. P/E, EV/EBITDA on fees · system RevPAR & RevPAR Index, net room/pipeline growth, fee margin (ROIC) · RevPAR Index <100 (share loss) or pipeline stalling.
8. P/EV, VNB · APE, VNB & VNB margin, persistency (EV/RoEV, solvency) · VNB margin falling on the ULIP tilt; persistency slide (mis-selling).
9. EV/Sales, then P/E/EV/EBITDA at maturity, per-sub · net adds, ARPU, churn (engagement, content amort %) · content spend outgrowing revenue; rising churn.
10. P/E · constant-currency revenue growth, EBIT margin, TCV/book-to-bill (attrition, utilisation) · CC growth decelerating while headcount/net-adds fall; book-to-bill <1.
11. P/NAV, EV/bookings (not P&L P/E) · pre-sales/bookings, collections vs bookings, net debt/equity (velocity, unsold inventory) · collections lagging bookings; aging unsold completed inventory.
12. P/NAV, FFO/AFFO, cap rate · NOI & same-store NOI, occupancy & WALE, AFFO payout (LTV, re-leasing spread) · AFFO payout >90% (cut risk); occupancy slide + negative re-leasing spreads.
13. P/E, EV/EBITDA · USFDA plant status, ANDA/Para IV pipeline, US price erosion (R&D %, complex mix) · Warning Letter leading to an Import Alert on the key plant; base erosion with no complex pipeline.
14. EV/EBITDA, P/E (cyclical) · volumes (retail vs wholesale), ASP & segment share, EBITDA margin (inventory days, EV mix) · wholesale-retail divergence + channel inventory build; rising per-unit discounts.
15. EV/EBITDA · ARPU, subs & net adds, net debt/EBITDA (churn, rev-market-share) · ARPU flat despite the hike; net debt rising with spectrum dues.
16. EV/EBITDA, EV/EBIT, ROCE · order inflow & backlog, book-to-bill, WC cycle & net debt (backlog coverage, ROCE vs WACC) · book-to-bill <1 for several quarters; backlog padded with low-margin orders; WC & net debt rising faster than revenue.
17. EV/EBITDA on mid-cycle (NOT asset-light P/E) · RevPAR (ADR × occ), GOPPAR, net debt/EBITDA (occupancy, supply pipeline) · debt-funded owned expansion at the peak; RevPAR up on ADR while occupancy falls.
18. EV/Sales on net revenue · GMV & take rate, contribution margin/txn, liquidity/active users (cohort GMV retention) · subsidy-bought GMV (negative CM); take-rate compression or disintermediation/leakage.
19. P/B, combined ratio · combined ratio (<100), loss & expense ratios, solvency (float yield, growth) · combined ratio >100 masked by investment gains; reserve under-provisioning.
20. EV/EBITDA, P/E · content/vehicle & kit value, client concentration, aftermarket/export mix (RM pass-through, ROCE) · ICE-only with no EV design wins; single-OEM concentration.

Score yourself strictly. If you got <17, list every miss, go back over that sector's playbook, and retake with the app's right-lens engine before you attempt Part B.

Part B: two full sector-playbook applications

Run two complete sector playbooks end-to-end on real companies from different sectors, with at least one Indian and one US/global. This is the "two playbook applications" half of the M6 gate. For each one, pull the actual data and write a one-to-two-page workup.

Candidate list. Pick two, from different sectors, with ≥1 India + ≥1 US. These are teaching specimens, and none of them is a recommendation:

• Airlines: InterGlobe Aviation (IndiGo), from the annual report + investor presentation on the BSE/NSE and screener.in. Or Southwest / Delta / Ryanair, from the 10-K/20-F on SEC EDGAR.
• Hotels: Indian Hotels (IHCL), from the annual report + investor presentation (BSE/NSE, screener.in). Or Marriott / Hilton, from the 10-K on EDGAR. Study the managed/franchised vs owned split.
• Media/streaming: Netflix, from the 10-K + the quarterly shareholder letter (the gold standard for this sector) on EDGAR / ir.netflix.net. Or an Indian listed media name (e.g. a Zee/Sun/Saregama) via BSE/NSE filings.
• Or any earlier sector playbook you want to prove you own (a bank, an NBFC, a SaaS name, a commodity producer, a REIT).

Structure each workup:

1. Classify the business in one line and name the lens with the lens-picking decision procedure.
2. Build the KPI panel for the sector, with the last 3–5 years of numbers pulled from filings, and each KPI marked healthy / watch / stress against the threshold.
3. Compute the primary multiple (e.g., EV/EBITDAR for the airline; SOTP for the hotel; EV/Sales and per-sub for the streamer) and state what it implies.
4. Run the red-flag panel for the sector explicitly: present or absent, with the evidence.
5. Close with the master check: estimate ROIC (or ROE/RoEV/ROA as the sector demands) versus a ~12% WACC hurdle (or cost of equity for financials), and state whether the spread is durable and reinvestable.
6. One-paragraph verdict: good/average/poor business, cheap/fair/expensive on the lens, and the one question you'd research next.

Rubric (0 / 1 / 2 per row; pass ≥ 10/14, i.e. ~71%):

File both workups in your company folder in the knowledge system. Passing Part A (≥17/20) and both Part B workups (≥10/14 each) certifies C6, and with the mastery check below it completes the Phase 5 gate.

Reading & resources

Airlines

• IndiGo (InterGlobe Aviation) annual report + quarterly investor presentations: BSE/NSE, screener.in. [Free][Beginner] The presentations spell out RASK, CASK, CASK ex-fuel, load factor, and fleet: the panel, pre-computed, to check your own build against.
• Southwest Airlines / Delta / Ryanair 10-K / 20-F: SEC EDGAR. [Free][Intermediate] Compare a US LCC, a legacy carrier, and a European LCC on the same panel.
• Rigas Doganis, Flying Off Course: Airline Economics and Marketing (4e). [Paid][Advanced] The standard text on why the industry economics are what they are.
• IATA Economics reports and the Airline Industry Economic Performance briefings. [Free][Intermediate] Industry-wide ROIC-vs-WACC data: the empirical proof of value destruction.
• Buffett's Berkshire Hathaway shareholder letters (1992, 2007) on airlines. [Free][Beginner] The "durable competitive disadvantage" and Kitty Hawk passages.
• Acquired podcast, episodes on airlines/aviation. [Free][Beginner]

Hotels

• Indian Hotels (IHCL) annual report + investor presentation (the asset-light pivot). [Free][Beginner] and Marriott International 10-K on EDGAR. [Free][Intermediate] Read the two side by side to feel the owned-vs-fee contrast.
• STR / CoStar and HVS RevPAR, occupancy, ADR, and supply-pipeline data and glossaries. [Some Free][Intermediate] The source of RevPAR Index benchmarks.
• Marriott / Hilton earnings-call transcripts, for how management talks about net unit growth and fee income. [Free][Intermediate]

Media / streaming

• Netflix 10-K + the quarterly shareholder letter: EDGAR / ir.netflix.net. [Free][Beginner→Advanced] The gold-standard sector disclosure: net adds, ARPU by region, content amortisation, FCF, engagement commentary. Read four consecutive letters to watch the land-grab-to-harvest arc.
• Matthew Ball, essays on streaming economics and The Metaverse and How It Will Revolutionize Everything (media-value chapters). [Some Free][Intermediate]
• Ben Thompson, Stratechery: "Aggregation Theory" and streaming pieces. [Some Free][Intermediate] Why scale + engagement is the moat.
• Gina Keating, Netflixed. [Paid][Beginner] The narrative history behind the numbers.

Synthesis / cross-sector

• Aswath Damodaran's data library (pages.stern.nyu.edu/~adamodar): industry-average multiples, margins, and ROIC by sector, updated each January. [Free][Intermediate] The empirical backbone of the synthesis table; pull the current numbers.
• Michael Mauboussin, What Does a Price-Earnings Multiple Mean? and the ROIC reports (Counterpoint Global). [Free][Advanced] Why each multiple encodes ROIC, growth, and risk.
• The sector KPI cheat-sheets artifact (artifacts/sector-kpi-cheatsheets.md), the one-page-per-sector companion for building the reflex.
• Re-read the M5.01–M5.09 modules' mastery checks as a spaced-review sweep before the exam.
• Going deeper — Michael Mauboussin, Measuring the Moat and Calculating Return on Invested Capital (Counterpoint Global / Morgan Stanley, free PDFs). [Free][Advanced] The deepest free treatment of why the ROIC>WACC spread persists or fades, which is the "meaning" under every lens in the ROIC > WACC section. Read it after you build the synthesis table, so the master check becomes rigorous instead of reflexive.
• Going deeper — Aswath Damodaran, Narrative and Numbers (Columbia Business School Publishing). [Paid][Advanced] How to attach a defensible business story, and a WACC, to each sector lens, turning the synthesis map into rigorous valuation. It's the natural bridge from Phase 5's classification into Phase 8's timed teardowns.

The Modern Analyst's Addendum

Everything above teaches this skill from first principles, by hand — that is how you learn it, and the mastery check still tests it that way. This addendum shows how a working analyst amplifies the same skill today. It adds; it never replaces. (R1/R10)

AI-Augment this skill

{
  "skill": "Running the airline RASK/CASK/EV-EBITDAR panel, the hotel owned-vs-fee SOTP, and the streaming ARPU/churn/lifetime-value chain (WE1-WE6) across a coverage list, and stress-testing the fixed-cost structure with a load-factor or occupancy sensitivity",
  "use": "An LLM turns quarterly RASK/CASK, RevPAR/GOPPAR, or subscriber/ARPU/churn disclosures into one clean table and drafts a first pass at which of the module's red flags are firing and which lens (SS4.14) applies to each name.",
  "tools": ["Claude/ChatGPT/Gemini to extract unit-economics figures from pasted investor decks or 10-K/annual-report excerpts into a table", "a notebook (Python/pandas) to hold the panel and run operating-leverage or lifetime-value sensitivity scenarios yourself", "DGCA's monthly traffic statistics for Indian carriers", "SEC EDGAR for US airline, hotel, and streaming 10-Ks"],
  "prompt": "Paste this airline's, hotel's, or streamer's last 4-8 quarters of unit-economics disclosures (RASK/CASK and load factor; or RevPAR/GOPPAR and owned-vs-fee EBITDA split; or subscribers/ARPU/churn/content spend). Build the panel, flag which of the module's red flags are firing, and run a load-factor (or occupancy, or churn) sensitivity showing how much the fixed-cost structure amplifies a small demand move into a larger profit move.",
  "verify": "Recompute the base-case unit profit directly from raw revenue and cost (never from a rounded per-unit rate multiplied back out -- WE1's own stated operating profit is a small case study in exactly that trap, corrected below) and check the tool's sensitivity table assumes only the cost share you actually told it was fixed.",
  "diy": "The mastery gate is unaided: build the RASK/CASK or RevPAR/GOPPAR or ARPU/churn panel from raw disclosures, apply SS4.14's lens-picking procedure, and state whether the business clears ROIC>WACC through a cycle -- no tool in the room."
}

Modern Data Analysis

By hand first. WE1-WE6's panels are hand-built for one airline, hotel, or streamer at a time, one quarter or one year.

Today's workflow. A generalist covering all seventeen of Phase 5's sub-sectors doesn't rebuild each panel from scratch every reporting season. In practice you hold the unit-economics KPI panel for every name in one table, and you run the same operating-leverage or lifetime-value sensitivity across the whole coverage list at once. A five-point load-factor shock or a churn move becomes a parameter sweep instead of a worked example you redo by hand each time.

Tools & sources (IN + US). For India, start with DGCA's monthly traffic statistics at dgca.gov.in. They give domestic/international traffic, load factor, and market share by carrier, all publicly downloadable. They're also the primary check on any Indian carrier's disclosed load factor. Company investor decks (IndiGo, Indian Hotels) give RASK/CASK and owned-vs-fee splits, and screener.in has multi-year financials. For the US, the DOT Bureau of Transportation Statistics' Air Carrier Statistics (T-100/Form 41 data) at transtats.bts.gov are the primary source for RASM/CASM-equivalent traffic and financial data by carrier. They're the direct analogue of DGCA's release. SEC EDGAR's XBRL company-facts API has Southwest's, Marriott's, and Netflix's historical revenue, cost, and subscriber-adjacent disclosures. And STR/CoStar provide US RevPAR benchmarks, already cited in Reading & resources.

# recomputed in-session -- R3
# Part 1: verification anchor -- WE1 NimbusJet, computed directly rather than via a rounded per-unit rate
pax_rev, anc_rev, tot_cost, ask = 58000, 10000, 60000, 150   # Rs cr, Rs cr, Rs cr, bn ASK
tot_rev = pax_rev + anc_rev
operating_profit_direct = tot_rev - tot_cost                 # the unambiguous way: revenue minus cost
rask, cask = tot_rev/ask/100, tot_cost/ask/100                # Rs per ASK
operating_profit_via_rounded_unit_rate = round(rask - cask, 2) * ask
print(f"operating profit, direct subtraction: Rs{operating_profit_direct}cr")
print(f"operating profit, via rounded RASK-CASK x ASK: Rs{operating_profit_via_rounded_unit_rate:.0f}cr")
# -> Rs8,000cr direct vs Rs7,950cr via the rounded rate -- the module states Rs7,950cr (see Check)

# Part 2: NEW -- quantify "margin is a leveraged bet on load factor" (the brief's framing for this module)
load_old, load_new = 0.85, 0.80                               # a plausible demand shock, cost held fixed
scale = load_new / load_old
rev_new = (pax_rev + anc_rev) * scale                          # passenger + ancillary both load-linked
profit_new = rev_new - tot_cost                                 # cost assumed ~fixed within one schedule
load_chg_pct = (load_new/load_old - 1)
profit_chg_pct = (profit_new/operating_profit_direct - 1)
print(f"\nload factor {load_old:.0%} -> {load_new:.0%} ({load_chg_pct:+.1%}):  "
      f"profit Rs{operating_profit_direct}cr -> Rs{profit_new:.0f}cr ({profit_chg_pct:+.1%})")
print(f"operating-leverage multiplier: {profit_chg_pct/load_chg_pct:.1f}x")

Verify. The direct-subtraction check exposes a small internal inconsistency in WE1. Its given revenue (₹68,000 cr) minus its given cost (₹60,000 cr) is unambiguously ₹8,000 cr, but the prose states "₹7,950 cr". That's the figure you get by rounding RASK to ₹4.53/ASK before multiplying back out by 150 bn ASK. The gap is ~0.6%, and it doesn't change any red-flag conclusion in WE1. Still, it's a real disagreement between WE1's stated total and its own given inputs, and no rounding choice on my part, so it's reported below as a defect candidate. Every other WE1 to WE6 figure recomputed for this addendum matched with no discrepancy: load factors, yields, CASK ex-fuel, days-cash, EV/EBITDAR for both NimbusJet and SkyValue, the Apex Hotels SOTP, and the StreamOne/BurnCo/BharatStream/PrimeReel subscriber economics. The load-factor sensitivity is new. A 5.9% relative fall in load factor turns a ₹8,000 cr profit into a ₹4,000 cr one, a 50% profit swing and an ~8.5× operating-leverage multiplier, on the same fleet, the same fares, the same cost base.

Verify, at full precision. Every worked figure restated unrounded, so the next reader can catch a rounding drift the way GR-220 was caught. NimbusJet (WE1): load factor 85.000000%, yield ₹4.549020 per RPK, RASK 4.533333, CASK 4.000000, PRASK 3.866667, CASK ex-fuel 2.400000, unit profit 0.533333, operating profit 8,000.00 taken as revenue minus cost, margin 11.764706%, fuel 40.000000% of opex, ancillary 14.705882% of revenue, days-cash 212.916667, net debt including leases 13,000, EV/EBITDAR 6.407407. Multiplying the displayed 0.53 by 150 bn ASK instead returns 7,950.00, fifty crore light, because RASK is 4.533333 and the third decimal is worth half a percent of the profit. TailwindAir (WE2): load factor 78.000000%, RASK 4.200000, CASK 4.700000, CASK ex-fuel 2.950000, unit loss −0.500000, operating loss −2,000, days-cash 17.473404. Apex (WE3): RevPAR 8,640.00, room revenue 1,576.80, GOPPAR 7,671.232877, RevPAR index 108.000000, fee EBITDA 220.00, sum-of-the-parts 14,000 plus 6,600 = 20,600 and equity 17,600, fee share 18.032787% of EBITDA and 32.038835% of enterprise value. One flag inside Apex: owned revenue is stated as approximately 3,000, so 33% of it is 990.00 and the sum-of-the-parts uses the round 1,000; the ₹10 cr sits inside the "≈" and changes nothing here, but it is the same species of drift as WE1's and is recorded rather than hidden. StreamOne (WE4): revenue $35.88bn, average life 43.478261 months, content amortisation 39.018952% of revenue, cash-over-amortisation gap +3.000000, operating profit 7.893600, enterprise value per subscriber 961.538462. BurnCo (WE4): revenue 4.200000, average life 18.181818 months, cash content 190.476190% of revenue. SkyValue (WE5): load factor 83.500000%, yield 17.215569 cents per RPM, RASM 16.250000, CASM 15.500000, PRASM 14.375000, CASM ex-fuel 11.250000, unit profit 0.750000 cents, operating profit $1,200m, margin 4.615385%, fuel 27.419355% of opex, ancillary 11.538462% of revenue, days-cash 132.459677, EV/EBITDAR 4.772727. BharatStream (WE6): revenue ₹19,800 cr, average life 22.222222 months, lifetime revenue per subscriber ₹1,222.22, content amortisation 75.757576% and cash content 90.909091% of revenue, enterprise value per subscriber ₹3,000.00. PrimeReel (WE6): revenue ₹8,640 cr, average life 45.454545 months, lifetime revenue per subscriber ₹8,181.82, enterprise value per subscriber ₹15,000.00, so the lifetime-revenue ratio is 6.694215 and the per-subscriber price ratio 5.000000 on a subscriber base 7.5 times smaller.

Quantitative lens

Refuse a static margin; use the operating-leverage multiplier. NimbusJet earns a comfortable ~12% operating margin at 85% load factor. Drop load to 80% with fares and cost unchanged, and operating profit halves. That's a demand shock well short of TailwindAir's 78% graveyard case, and it gives an ~8.5× multiplier between the percentage move in load factor and the percentage move in profit. It's the arithmetic behind the claim at the start of the airline section that the industry "structurally destroys capital". The same machine is a good business at 85% and a value-destroyer a few points lower, with nothing else in the story changing. So think of margin as a property of an airline at a given load factor, and never of the airline alone.

Refuse a rounded per-unit rate multiplied back out; use the direct subtraction. WE1's own ₹7,950 cr is this error in miniature. Divide revenue and cost by ASK, round each to two decimals, subtract, and multiply back by ASK, and you won't recover the true total (₹8,000 cr) when the division doesn't terminate cleanly. The per-unit rate (RASK, CASK, ARPU, RevPAR) is for comparison. The total is always a direct subtraction.

Refuse a blended multiple on a mixed owned-plus-fee operator; use sum-of-the-parts. Apex Hotels' fee segment earns ~18% of group EBITDA. At the worked example's own multiples (14× owned, 30× fee), it's close to a third of enterprise value. That's because a rupee of asset-light, high-ROIC fee income is worth several times a rupee of capital-heavy, cyclical owned income. One blended multiple erases that difference in both directions.

The limits. The load-factor sensitivity above assumes cost is entirely fixed within one scheduling period, which is a simplification. Some costs do flex with passenger count: catering, handling, a sliver of fuel burn from a lighter aircraft. So 8.5× is the right order of magnitude and the right direction, but it isn't a precise elasticity to trade on. And operating leverage is symmetric. The same multiplier that turns a 6% load-factor shortfall into a 50% profit collapse turns a load-factor recovery into an outsized profit rebound. The sensitivity measures the amplifier. It can't tell you which direction the next quarter moves.

Do it in code: hold the panel in G2/DA1.02, run the load-factor/occupancy/churn sensitivity as a distribution rather than one scenario with G1/QM1.01, chart the amplification with G2/DA1.04, and pull DGCA's or BTS's traffic data programmatically with G2/CS2.05 instead of reading monthly releases by hand.

Where this goes next: galaxy cross-links

• G2 · DA1.02 Pandas data wrangling: the multi-sector KPI panel above, built properly across every name in a coverage list rather than one worked example at a time.
• G2 · DA1.04 EDA & visualization: charting the load-factor-to-profit amplification (or the RevPAR-to-GOPPAR one from the hotel section) is how "operating leverage" becomes a picture instead of an adjective.
• G1 · QM1.01 Probability distributions & estimation: turns a single load-factor or churn scenario into a distribution of outcomes, so the 8.5× multiplier comes with a range of plausible profit outcomes, not one point estimate.
• G2 · CS2.05 APIs, HTTP & JSON: the mechanics of pulling DGCA's and BTS's published traffic data programmatically instead of reading monthly PDF releases by hand.
• X1 · AI0.07 The AI-augmented analyst: the general version of the AI-Augment box above, covering what to delegate, what to verify, and why the mastery gate stays unaided.

Teach it back & journal

Teach it back (Feynman). Explain to a smart 15-year-old, in plain language and on one page: "How can an airline that fills 90% of its seats still go bankrupt, while a hotel company that owns almost no hotels prints money?" Use no jargon they don't already have. Build up load factor, yield, and unit cost for the airline, and the owned-vs-fee distinction for the hotel, from everyday intuition: a bus that's full but sells tickets too cheap, or a franchise that collects a royalty without buying the shop. If you can make the paradox obvious to them, you own the material.

Journal reflection. Look back over all ten Phase 5 playbooks and answer this in your decision journal, without flattering yourself: Which valuation lens did I most often reach for wrongly, and why? Most learners have one reflex error. Usually it's defaulting to P/E, forgetting to normalise cyclicals, or blending asset-light and asset-heavy. Name yours, write the two-line rule that fixes it, and add it to your permanent checklist. Then write your own one-paragraph version of the lens-picking procedure, in your words, the way you'll actually run it in a teardown next week.

This module's flashcards and mastery quiz are wired into the app: see the node's Quiz and Reviews.