alirezarezvani/claude-skills / finance/skills/stock-analysis
stock-analysis
UnverifiedAnalyze, research or value a listed stock or company (US, global, NSE or BSE) with verified data, ratios, scores and valuation. Use whenever asked if a stock is a buy, what it is worth, or how a company looks.
references/02-core-factors.md
View on GitHub# Core factors — business model, moat, industry and growth
Use this when: you are at Stage 4 and need the universal, non-financial half of the analysis — what the business actually is, whether it can defend its economics, and whether it can reinvest at high returns for long enough to matter.
Everything the financial statements show you is an *output*. This file covers the *inputs* that determine whether those outputs persist: the monetisation architecture, the moat mechanism and its direction, the structure of the industry, and the length of the reinvestment runway. Get these wrong and the ratio work is arithmetic about a business you have misunderstood. Two disciplines carry over from the governing principle: no factor here has a universal "good" level — read every number against sector peers and the company's own 5–10 year record — and never let one strong factor stand in for the set.
## Contents
- [How to work through this file](#how-to-work-through-this-file)
- [Part A — Business model and competitive moat](#part-a--business-model-and-competitive-moat)
- [Part B — Industry, market and competitive dynamics](#part-b--industry-market-and-competitive-dynamics)
- [Part C — Growth prospects and reinvestment runway](#part-c--growth-prospects-and-reinvestment-runway)
- [Verify against something the company did not write](#verify-against-something-the-company-did-not-write)
- [Where this generic frame breaks](#where-this-generic-frame-breaks)
- [Checklist](#checklist)
## How to work through this file
Three passes, in order. Each answers one question, and each depends on the one before it.
1. **What is this business?** (Part A) — how it makes money, from whom, and why that is hard to take away.
2. **What game is it playing?** (Part B) — industry structure sets the ceiling on returns more often than management skill does.
3. **How much longer can it compound?** (Part C) — growth only creates value above the cost of capital, and only while the runway lasts.
Write down the answers as short factual claims with a source and period attached, not as adjectives. "Top customer = 31% of FY25 revenue, disclosed under Ind AS 108 segment note" is analysis. "Customer concentration is a concern" is not.
Do not run all forty-odd factors below at equal depth. For most companies three or four decide the outcome. Identify them early, evidence those properly, and use the rest to confirm nothing disqualifying was missed.
**Source map.** US/global: 10-K Item 1 (Business), Item 1A (Risk Factors), Item 7 (MD&A), the segment footnote (ASC 280), 8-Ks, proxy, investor-day decks, earnings-call transcripts, EDGAR full-text search across competitors and customers. **India:** the annual report's Management Discussion & Analysis (mandatory content under SEBI LODR Schedule V — industry structure, opportunities and threats, segment performance, outlook, risks), the segment note under Ind AS 108, the Directors' Report and its annexures, the Business Responsibility & Sustainability Report, and — often the single richest source — the quarterly earnings **concall transcript**, which listed companies must publish on the website and file with the exchanges. Indian investor presentations carry order books, capacity, volumes and market-share claims that appear nowhere in the audited accounts; use them, and label them unaudited.
---
## Part A — Business model and competitive moat
### A1. Define the business before judging it
Map exactly what products or services generate revenue, **who actually pays** (which is often not who uses), and the job being done for that payer. Break revenue *and gross profit* down by segment and by individual product line — profit concentration is almost always more extreme than revenue concentration, and one product frequently produces the overwhelming majority of economic profit while the narrative describes the whole portfolio.
Test yourself: can you state the business and its profit engine in two plain sentences? If not, you do not yet know enough to assess the moat. Separate the *story* from where money is actually made.
Judge mission-criticality: a product embedded in the customer's workflow, regulated process, or production line has structurally more durable economics than one that can be deferred a quarter without consequence.
*Red flags:* refusal to disclose segment-level economics; a "diversified" structure masking one weak core; frequent unexplained pivots; reliance on a single hit product; you cannot articulate what the customer is buying.
### A2. Monetisation mechanics
Identify the architecture precisely — unit sale, subscription, usage/consumption, transaction take-rate, licence/royalty, advertising, leasing, freemium conversion, razor-and-blade with high-margin consumables, or marketplace commission. Then ask three things: who sets price, how often it is billed, and whether revenue scales with value delivered or is structurally capped.
This matters because the architecture, not the industry label, drives margin structure, capital intensity, predictability and cyclicality. Consumption models capture upside but add volatility. Subscriptions add predictability and are evidence of switching costs. Advertising is cyclical and concentration-prone. In razor-and-blade and aftermarket models the real profit pool sits in the consumable — analyse *that* line, not the installed-base sales that look like the business.
For platforms, quantify the take rate and interrogate its headroom. A rising take rate is frequently how a platform masks stalled volume growth.
### A3. Revenue quality and recurrence
Split revenue into contracted/recurring (subscriptions, maintenance, long-term service contracts, consumables) versus transactional, project or one-off. Then check contract length, auto-renewal terms, minimum commitments, remaining performance obligations (RPO) or backlog, and the deferred-revenue trend. Cohort retention tells you whether the installed base expands or leaks.
Recurring revenue is worth more because it does not have to be re-won each period; it is also the cleanest observable evidence that switching costs exist. Be sceptical of the label: one-time hardware or perpetual-licence sales relabelled as "ARR", or annually re-solicited business described as recurring, is a common dressing-up.
### A4. Unit economics
Model one incremental customer or unit end to end: gross margin per unit, fully loaded customer acquisition cost, lifetime value, payback period, contribution margin after all variable costs. Ask whether the economics improve or degrade with scale, and — critically — whether they hold without promotional subsidy.
A company can grow fast and still be uninvestable if each customer is unprofitable or payback is dangerously long. Sound unit economics are the precondition for self-funded growth; broken ones mean growth destroys value, and faster growth destroys it faster.
Interrogate the LTV assumption itself. It is built on an assumed churn rate and discount rate, both chosen by the company.
### A5. The moat sources — name the mechanism
A moat is not an adjective. Identify which specific mechanism produces the excess return, or conclude there is none.
**Pricing power.** The clearest external evidence of a moat. Examine the history of price increases against volume response, gross-margin behaviour through cost-inflation periods, whether price is a small and low-salience share of the customer's total cost, and whether pricing is value-based. The recent global inflation episode is a natural experiment: did the company pass input costs through cleanly, or discount to hold volume?
**Switching costs.** Quantify what it actually costs a customer to leave — money, time, retraining, data migration, integration rework, certification, operational risk. Retention and renewal rates are the empirical test; footprint depth (modules or seats per account) is the leading indicator.
**Network effects.** Establish whether each additional user increases value for existing users, and of which kind: direct, indirect/two-sided, or data. Then test the two things that break them — multi-homing, and networks that are locally dense but do not compound nationally.
**Brand.** Test whether the brand changes buying behaviour and commands a price premium, not whether it is well known. Awareness is not a moat. Rising advertising spend merely to hold share is evidence against a brand moat, not for one.
**Scale and cost advantage.** Establish whether size produces a structural unit-cost advantage — purchasing scale, manufacturing or logistics density, fixed-cost leverage, distribution reach, proprietary process — and verify it shows up in the margin gap versus smaller peers. If the claimed advantage is not visible in the numbers, it is not there. Also consider *efficient scale*: niches profitably served by one or two players where entry would destroy everyone's returns (pipelines, rail, regional utilities, a single-city cement market).
**Intangibles: IP, patents, licences, regulatory exclusivity.** Inventory patents with revenue-weighted expiry dates, approvals, spectrum, franchises and concessions. Map the cliff explicitly. Legal exclusivity can be monopoly-grade, but it is time-bounded, can be litigated away, and can be legislated away — never model it as permanent.
### A6. Dependencies that can end the business
**Customer concentration.** Quantify revenue from the top 1/5/10 customers. In the US this surfaces via the 10-K and ASC 280 disclosure of any customer above 10% of revenue; **in India, Ind AS 108 requires the same entity-wide disclosure of major customers at the 10% threshold** — look for it in the segment note, and if it is absent for a company that is obviously concentrated, treat the silence as information. Assess contract length, tenure, and the risk that a large customer insources or vertically integrates. Include channel concentration: one distributor or one retailer can be the real dependency.
**Supplier and input dependence.** Identify single- or sole-sourced critical inputs, one contract manufacturer, one foundry, one API supplier, one geography. Assess substitutability, second-source qualification, inventory buffer, and hedging. No demand-side moat survives an input that cannot be obtained. In India, Schedule III requires disclosure of cost of materials consumed, which lets you size input dependency even where suppliers are not named.
**Geographic mix.** Break revenue *and operating profit* by region. Ask whether the moat travels — many models that dominate a home market fail abroad, which both caps the runway and turns expansion into value destruction. Flag single-country dependence for the bulk of profit or upside, unhedged FX exposure, and repatriation or capital-control risk.
### A7. Moat synthesis — width, durability, direction
Now combine. Width is how large the excess return; durability is how many years it persists; **direction is usually the more valuable judgement.** A narrowing wide moat is often a worse investment than a widening narrow one, because the market has already paid for the width.
Measure how long the company has earned returns above its cost of capital and whether that spread is widening, stable or narrowing. Then name the primary disruption vector explicitly — technological change, business-model disruption, or counter-positioning where the incumbent cannot respond without cannibalising its own economics. If you cannot name a plausible way the moat is attacked, you have not looked hard enough.
Mechanics of ROIC and the ROIC–WACC spread are in `references/05-returns-and-dupont.md`; use those numbers here as evidence, and do not double-count them as a separate score.
### Part A metric set
Ranges are **indicative only**. They shift by market, by sector, by cycle and by period; peer and own-history comparison overrides every band below.
| Metric | Definition / how to compute | Indicative healthy range | Why it matters |
|---|---|---|---|
| **Revenue and gross profit by segment/product** | Each segment or product line as % of total revenue *and* of total gross profit, from the segment note (ASC 280 / Ind AS 108) | No single product >~50% of profit without a credible follow-on; profit concentration usually exceeds revenue concentration | Locates the real profit engine, which is frequently not the one the narrative describes |
| **Recurring revenue mix** | Contracted/subscription/consumable revenue ÷ total revenue | >70% is a genuinely subscription business; 30–70% hybrid; interrogate any "recurring" claim below that | Recurring revenue does not have to be re-won each period and is direct evidence of switching costs |
| **Net revenue retention (NRR/NDR)** | Revenue this period from last period's cohort ÷ that cohort's prior revenue, including expansion, contraction and churn | >120% exceptional, 110–120% strong, 100–110% adequate, <100% the base is leaking | The single best test of lock-in; above 100% the business grows with zero new customers |
| **Gross retention / logo churn** | Cohort revenue retained before any expansion; annual customer churn % | Gross retention >90% for enterprise, >80% SMB; churn trend matters more than level | Strips out the flattering effect of upsell into a shrinking base |
| **LTV/CAC and CAC payback** | Lifetime gross profit per customer ÷ fully loaded acquisition cost; months of contribution margin to repay CAC | LTV/CAC >3x; payback <12 months consumer/SMB, <24 months enterprise | Determines whether growth is self-funding or a subsidy; lengthening payback is an early warning |
| **Contribution margin per unit** | Revenue less *all* variable cost per unit or transaction | Positive and improving with scale | Negative contribution margin means growth accelerates value destruction |
| **Realised price growth vs volume growth** | Split reported revenue growth into price, volume, mix and FX (percentage points) | Both positive is the franchise signature; price-only with falling volume is demand destruction | The cleanest quantitative read on pricing power |
| **Gross margin level and 5–10 yr trend** | Gross profit ÷ revenue, against peers of similar model | Stable or rising through an input-cost shock | Margin held through inflation is pricing power demonstrated, not asserted |
| **Take rate (platforms)** | Net revenue ÷ GMV or gross transaction value | Level is sector-specific; the *trend* and its driver are what matter | A rising take rate frequently masks stalled volume |
| **Customer concentration** | Top-1, top-5, top-10 customers as % of revenue | Top customer <10%; >20% is a single-point-of-failure | Concentrated buyers extract margin and can end revenue abruptly |
| **Single-sourced critical inputs** | Count of critical inputs with no qualified second source; key input as % of COGS | Zero sole-sourced critical inputs, or a qualified alternate | Supply failure is existential and not offset by any demand-side moat |
| **Revenue exposed to patent/licence expiry** | % of revenue from products losing exclusivity within 3–5 years, with dates | Low, and covered by a late-stage pipeline | Cliffs are dateable, foreseeable, and routinely under-modelled |
| **Years of ROIC > WACC, and spread trend** | From `05-returns-and-dupont.md`; plot the spread over 10 years | A long run with a stable or widening spread | Duration and direction of excess returns is the largest driver of intrinsic value |
---
## Part B — Industry, market and competitive dynamics
Industry structure sets the ceiling on returns more reliably than management quality does. A capable operator in a structurally bad industry usually loses to a mediocre one in a good industry.
### B1. Market size, and whether the sizing is honest
Decompose into TAM, SAM (serviceable addressable) and SOM (serviceable obtainable). Establish whether the estimate is bottom-up (units × price × realistic penetration) or a top-down number lifted from a slide. Note who produced it and when — company IR, a sell-side bank, or an independent body. In India, industry sizing usually traces to CRISIL, ICRA, industry associations (SIAM, IPA, CREDAI, IBEF) or a company-commissioned consultant; treat commissioned sizing as marketing until independently checked.
Then do the step that matters: **back-calculate the market share the company must reach to justify the current price.** This converts a vague "huge market" into a testable claim. Cross-reference the reverse-DCF in `references/06-valuation.md`.
*Red flags:* TAM revised upward to defend a falling stock; adjacent-market TAM stacking; "we only need 1% of a giant market"; a bottom-up build implying more than 100% of any realistic segment.
### B2. Growth rate and position on the S-curve
Locate the industry on the adoption curve — early adopters, mass market, or saturation. Decompose industry growth into volume, price and mix. Separate durable secular demand from pull-forward (pandemic, subsidy, pre-buy ahead of a regulation change) that will normalise. Compare industry growth to nominal GDP: much of what is marketed as a growth market is GDP plus inflation.
Mistaking a maturing market for a growth market is one of the most expensive errors available, because the multiple and the growth rate de-rate together.
### B3. Structure, concentration and the rationality of competitors
Count the players and map the share distribution. Compute CR4/CR8 and, where you have share data, the Herfindahl-Hirschman Index. Establish the trajectory — consolidating or fragmenting.
Then ask the question that concentration statistics miss: **are the competitors rational?** A single player that does not maximise profit — state-owned, subsidy-funded, PE-backed and buying share to exit, or a subscale player pricing for survival — can compete away the profit pool for everyone in a structurally concentrated industry. This is a live issue in Indian PSU-heavy sectors and in any market where a well-funded entrant is buying share.
### B4. Rivalry intensity and how share actually moves
Trace market share over 5–10 years, and establish *how* it moves — through price, product or distribution. Share earned without price concessions is owned; share bought with rebates and promotions is rented. Watch whether R&D and advertising intensity must keep rising just to hold position: that is a moat being consumed to look stable.
### B5. The five forces, applied concretely
**Supplier power** — supplier concentration relative to the industry, substitutability, single-sourcing, the threat of a supplier forward-integrating, and the industry's demonstrated ability to pass input costs through and with what lag.
**Buyer and channel power** — buyer concentration, price sensitivity, backward-integration ability, and the gatekeepers sitting between the company and the end user (mega-retailers, distributors, group purchasing organisations, app stores, marketplaces, hospital chains, e-commerce platforms). A dominant channel quietly becomes the industry's real profit-taker; watch its take rate.
**Barriers to entry** — enumerate and stress-test each one: minimum efficient scale versus market size, capital intensity, licences, IP, network effects, brand, switching costs, proprietary distribution, learning-curve cost. Judge whether each is widening or eroding. High sustained margins with low barriers are an invitation, not a moat.
**Substitutes and disruption** — map what else does the same job, and track the *price-performance trajectory* of the alternative rather than its current position. Apply the low-end and new-market disruption lens. Substitution destroys value permanently rather than cyclically, and the financials look fine until the inflection.
### B6. Cyclicality, and where in the cycle you are standing
Classify demand drivers as cyclical (GDP, credit, capex, housing, commodity price), seasonal, or secular. Locate the position in the cycle and estimate mid-cycle normalised margins and earnings rather than trusting the current print.
**The peak-cycle trap:** record margins alongside an optically low P/E is the classic value trap, not a bargain. The inverse error — reading a structural decline as a cyclical dip — destroys just as much capital. Deep cyclicals get the overlay in `references/13-situations.md`.
### B7. The capital cycle — watch supply, not demand
Track industry-wide capacity additions against demand growth, allowing for capex lead times. Monitor utilisation, whether capital is flooding in (competitor expansions, IPOs, PE and VC funding, announced greenfield projects) or leaving, and channel inventory.
Returns are driven more by changes in supply than by demand forecasts. Heavy investment in good times sows the next glut; capital fleeing a hated industry sets up the next up-cycle. Aggregate industry capex ÷ depreciation is the cheapest single supply-side indicator available. This lens is decisive in cement, chemicals, steel, shipping, semiconductors, hotels and airlines.
### B8. Regulation, policy and political economy
Map the whole regime — antitrust, price control, tariffs, environmental and emissions rules, data and privacy, licensing, reimbursement, safety — plus pending legislation. Quantify dependence on subsidy, tax incentive or regulatory arbitrage.
**India-specific:** price control and trade-margin rationalisation (NPPA/DPCO in pharma), tariff and anti-dumping orders, GST rate changes, PLI scheme dependence, sectoral regulators (RBI, IRDAI, TRAI, CERC/SERCs, RERA), and the pattern of retrospective or mid-stream policy change. Sizeable chunks of reported profit in PLI-supported sectors are policy income, not franchise income — say so explicitly.
**US/global:** FDA and reimbursement, FTC/DOJ antitrust posture, IRA/CHIPS-style incentives, tariff schedules, state-level utility rate cases, EU DMA/DSA obligations.
### B9. Where the profit pool sits, and which layer the company occupies
Map profit along the value chain and, more importantly, the direction it is migrating — hardware to software, OEM to platform, network to content, manufacturer to brand owner. Compare ROIC across peers and against WACC. Establish whether the whole industry earns above its cost of capital or only one or two players do.
Some industries destroy capital regardless of management quality; much of airline and dry-bulk shipping history is the proof. Owning the layer the profit pool is moving *toward* is frequently the entire call.
### B10. Commoditisation
Commoditisation is the default state of most products. Test whether differentiation is real: can the industry price above inflation without losing volume, and pass input costs through, and with what lag? Track premium versus private label or generic, and the degree of spec interchangeability. Slow gross-margin compression across an industry is what commoditisation looks like in the accounts, years before anyone calls it that.
### Part B metric set
Indicative only — vary by market, cycle and period; peer and own-history comparison overrides them.
| Metric | Definition / how to compute | Indicative healthy range | Why it matters |
|---|---|---|---|
| **Penetration (revenue ÷ TAM)** | Company revenue as % of a credibly built, bottom-up TAM | Low penetration only counts if the TAM is independently sourced | Sets the ceiling; the number most often inflated to justify a multiple |
| **Implied share needed** | Back-solve from the reverse DCF: what share must be reached to justify today's price | Comfortably below the current leader's share | Turns a narrative into a falsifiable claim |
| **Industry CAGR vs nominal GDP** | 3/5/10-yr industry growth ÷ nominal GDP growth | >1.5x GDP is a genuine growth industry | Distinguishes secular tailwind from inflation |
| **CR4 / HHI** | Top-4 combined share; sum of squared shares (US DOJ convention: >2,500 concentrated, 1,500–2,500 moderate, <1,500 unconcentrated) | Concentrated *and* rational is the profitable combination | Structure predicts industry profitability better than any single firm's strategy |
| **Market-share trend** | Basis points gained/lost per year over 5–10 yrs, and the mechanism | Flat-to-rising share *without* price concession | Share bought with discounts is rented and reverses |
| **Industry capacity utilisation** | Output ÷ installed capacity, industry-wide | 80–85% is typically the pricing-power threshold; below ~70% price discipline breaks | The supply-side variable that sets industry pricing |
| **Industry capex ÷ depreciation** | Aggregate peer capex ÷ aggregate depreciation | ~1.0 = replacement; sustained >1.5 across an industry warns of a coming glut | The single best early indicator in the capital cycle |
| **Book-to-bill** | Orders received ÷ revenue billed in the period | >1.0 growing, <1.0 for two-plus quarters is a genuine warning | Forward-looking where backlog exists |
| **Industry ROIC vs WACC** | Median peer ROIC less WACC; also the dispersion across peers | Median above WACC, with the company in the upper quartile | Reveals whether the industry creates value at all, or only its best player does |
| **Current margin vs 10-yr mid-cycle** | Current EBIT margin ÷ 10-year average margin | Near 1.0 for normalisation; well above 1.0 means you are underwriting peak | Prevents the peak-cycle low-P/E trap |
| **Realised price vs input inflation** | Company/industry realised price growth less input-cost inflation | ≥0 through a cost shock | Direct test of commoditisation |
| **Revenue dependent on subsidy or regulated price** | % of revenue or EBIT from subsidy, incentive scheme or a regulated tariff | Known, quantified, and modelled to expiry | Policy income is not franchise income and can end on a date |
---
## Part C — Growth prospects and reinvestment runway
Growth is not intrinsically good. Growth at returns below the cost of capital destroys value, and the faster it grows the more it destroys. This section establishes three things: was past growth real, is incremental capital earning a return, and how many years of that are left.
### C1. Decompose historical growth — never accept a CAGR
Compute 3/5/10-year revenue CAGR, then break it apart:
- **Organic vs inorganic.** For any acquisitive company this is the whole analysis. Strip acquisitions and ask what the base business did.
- **Organic into volume, price, mix and FX**, in percentage points.
- **By segment and geography**, to see whether growth is broad-based or one product in one region.
- **Sequential (QoQ) as well as YoY**, because inflections show up sequentially first.
Adjust for revenue-recognition changes (ASC 606 / Ind AS 115), divestitures, and one-off commodity or pandemic spikes that distort the base. Volume-led organic growth is the highest-quality kind; price-only, FX-only or acquisition-fuelled growth is far less repeatable and often conceals a stagnating core.
For serial acquirers specifically, project what reported growth looks like when M&A pauses — and scrutinise purchase accounting (fair-value step-ups, restructuring reserves created in acquisition accounting then released to earnings, contingent-consideration remeasurement, "one-off" integration costs that recur every single year). Roll-ups can show rising reported EPS while the acquired businesses shrink, because each deal resets the baseline. Detail in `references/07-forensic-red-flags.md`.
### C2. Earnings growth quality
Separate EPS growth into revenue growth, margin expansion, share-count reduction, tax-rate change, and below-the-line items. Compare EPS CAGR to net-income CAGR to revenue CAGR — EPS can outgrow net income purely through buybacks. Then verify earnings convert to cash (see `references/03-earnings-quality.md` and `references/04-balance-sheet-and-cashflow.md`), and read the GAAP-to-adjusted bridge for add-backs that are growing, recurring, and real (stock-based compensation above all).
Growth manufactured by financial engineering is not evidence of a compounding business.
### C3. Reinvestment rate and incremental return on capital
This is the most important factor in Part C and the most commonly skipped.
Reinvestment rate ≈ (net capex + acquisitions + change in net working capital + capitalised R&D or growth opex) ÷ NOPAT.
ROIIC ≈ Δ NOPAT ÷ Δ invested capital, measured over rolling 3–5 year windows so a single year's lumpiness does not dominate.
Then compare ROIIC to WACC and to the company's own average ROIC. **ROIC tells you about capital already deployed; ROIIC tells you about the capital being deployed now** — which is what you are actually buying. ROIIC well below reported average ROIC means a good legacy business is subsidising poor new investment, and reported ROIC will fade toward the incremental rate over time. Full mechanics in `references/05-returns-and-dupont.md`.
### C4. Runway length and the sustainable growth identity
Anchor on: **sustainable growth ≈ reinvestment rate × ROIC.** A company returning most of its earnings cannot also compound at a high rate; if the guidance implies otherwise, something in the model is wrong.
Then estimate how many years of high-return reinvestment remain: TAM penetration, remaining white space, store or plant or route density versus a realistic ceiling, and whether the pipeline of projects clearing the hurdle rate is expanding or shrinking. Two companies with identical ROIC differ enormously in value if one has twenty years of runway and the other three. Runway is what justifies — or refutes — a premium multiple.
A specific tell: high ROIC with cash piling up and buybacks replacing capex usually means the reinvestment opportunity set has closed, whatever the growth narrative says.
### C5. Backlog, order book and RPO — the forward evidence
For industrials, capital goods, defence, EPC, construction, semis and subscription software, examine backlog or order book size and growth, coverage (backlog ÷ trailing revenue, expressed in months or years), book-to-bill, RPO and current RPO, de-booking and cancellation rates, and the *margin* embedded in the backlog.
These are among the few forward-looking, semi-verifiable indicators available. A shrinking backlog warns that reported revenue growth is about to roll over while the income statement still looks strong.
Interrogate quality: backlog padded with non-binding letters of intent or framework agreements, backlog growing only because delivery times lengthened, or a large low-margin cancellable order presented as firm. **India note:** order-book figures for EPC, capital goods and defence companies come from investor presentations and concalls and are unaudited — check whether the definition (firm orders, L1 orders won but not awarded, framework agreements) changed between periods.
### C6. Capex intensity and the maintenance/growth split
Split capex into maintenance and growth. Management sometimes discloses the split; otherwise estimate maintenance as depreciation adjusted for inflation and asset-base growth, and say plainly that it is an estimate.
Then review announced expansion: new plants, fabs, stores, data centres, mines — greenfield versus brownfield, budgeted cost, capacity added, timeline, ramp curve, expected return and payback. Compare against the previous announcement to catch cost overruns and slippage, which are chronic in Indian infrastructure, cement, metals and specialty-chemicals expansions.
Growth capex is where value is most often destroyed: mistimed, over-budget, or built at cycle-peak equipment prices into a market that will be oversupplied by the time it commissions. Cross-check against the industry capital cycle in B7 — the moment to be suspicious is when the whole industry is expanding at once.
### C7. The forward bridge — make growth arithmetic, not narrative
Take management's stated drivers (new products, secular tailwind, share gain, price, geographic entry) and build an additive bridge from today's revenue to the target, driver by driver, in percentage points. Then check the drivers actually sum to the target, and that each one is independently plausible.
Separate secular from cyclical demand in that bridge. Extrapolating cyclical peak demand as structural is how peak-cycle multiples get paid.
### C8. Innovation: pipeline and R&D productivity
Assess the vitality index — revenue from products launched in the last 3–5 years — alongside R&D as % of revenue *and its productivity* (incremental revenue or gross profit per R&D rupee/dollar). R&D spend is an input; productivity is the output that separates innovators from cash-burning labs.
Review launch cadence, the hit rate of past launches, the patent and exclusivity timeline, and pipeline depth by stage. A falling vitality index means the company is living off legacy products, which is a growth cliff with a date on it even when current earnings look fine.
### C9. Adjacency and new-market expansion
Ask whether there is a demonstrated right to win — a transferable capability, brand or distribution — or whether this is diworsification. The evidence is the track record: did earlier new-market entries reach target economics, or quietly stall and retreat? Compare unit economics and payback in new markets against the mature core.
A repeatable expansion playbook (a retailer that reliably hits store economics in each new region) is a powerful and durable driver. Serial entry-and-retreat means the runway is narrower than claimed.
### C10. Price, volume and mix in forward growth
Determine how much future growth is priced versus volume, and whether price sticks without volume loss. Price-led growth carries close to 100% incremental margin and is highly durable when backed by brand, switching costs or scarcity — and is a warning sign when volumes fall as prices rise. Check contractual escalators, and test whether a "premiumisation" narrative is actually visible in mix data.
### C11. Guidance credibility
Build a scorecard of guidance versus actuals across the last 8–12 quarters and of multi-year targets versus delivery. Does management beat, meet, sandbag or miss? How does initial full-year guidance evolve — serial cuts or raises? Were the previous analyst-day roadmap and "catalysts" delivered?
Their historical accuracy is the best available prior for the current forecast. Serial over-promisers should have their projections discounted heavily regardless of how good the current story sounds. **India note:** formal numeric guidance is less common; the equivalent evidence is concall commentary — capacity commissioning dates, margin guidance, order-inflow expectations — checked against what subsequently happened. Watch for goalposts moving and for KPIs quietly dropped when they turn unfavourable.
### C12. Capital allocation track record
Over a decade this is often the largest single driver of per-share value. Assess the full history across organic reinvestment, M&A, buybacks, dividends and debt paydown.
For M&A: prices paid, goodwill impairment history, synergy realisation, post-deal ROIC. For buybacks: were they executed at low or high valuations? A buyback at a stretched multiple is value destruction dressed as shareholder return. Judge whether capital consistently flows to the highest-return use and whether management can articulate a hurdle rate.
Governance dimensions of the same question — related-party leakage, promoter-group transactions, empire building — are in `references/08-governance.md`.
### C13. Deceleration risk and the law of large numbers
Compute the **absolute** revenue the company must add each year to sustain its growth rate, and judge whether that is plausible given market size and share. High percentage growth becomes arithmetically harder as the base compounds.
Watch the deceleration signals: sequential growth slowing, leading indicators (bookings, backlog, app downloads, hiring) decelerating ahead of reported revenue, comps getting harder, share nearing a ceiling. Then distinguish a comp-driven air pocket from structural maturation.
Identifying the inflection from hyper-growth to mature growth is one of the highest-value calls in equity analysis, because multiples de-rate violently and simultaneously with the growth rate.
Apply the **outside view** here. Before accepting a bottom-up forecast, ask what the base rate is: how often do companies of this size actually sustain 20%+ growth for a decade, how often do turnarounds and roll-ups work, and how fast does excess ROIC historically fade toward the cost of capital? Bottom-up models produce systematic inside-view optimism; base rates are the cheapest correction available. More in `references/17-process-and-epistemics.md`.
### C14. How the growth is funded
Establish whether growth is funded by internally generated cash, debt, or equity issuance. Compute the funding gap (reinvestment need less operating cash flow) and how it is plugged. Track share-count growth from issuance and stock-based compensation, convertible and warrant overhang, and leverage rising to fund capex.
**Per-share growth is what accrues to owners.** A company growing revenue 20% while issuing 15% more shares is barely compounding for you. Growth requiring continual external capital is fragile because it depends on capital markets staying open and cheap. Self-funded growth is the gold standard of a durable runway.
### Part C metric set
Indicative only — vary by market, cycle and period; peer and own-history comparison overrides them.
| Metric | Definition / how to compute | Indicative healthy range | Why it matters |
|---|---|---|---|
| **Organic vs inorganic growth split** | Revenue growth excluding acquisitions completed in the last 12 months, vs total | Majority organic; know the number before judging anything else | Acquired growth is bought, not earned, and stops when M&A stops |
| **Volume / price / mix / FX contribution** | Decomposition of organic growth into pp of each | Positive volume contribution in most years | Volume-led growth is the most durable form |
| **ROIIC (incremental ROIC)** | Δ NOPAT ÷ Δ invested capital over rolling 3–5 yr windows | Above WACC with a clear margin; ~15–20%+ marks a genuine compounder in most sectors | Prices what the capital being deployed *now* will earn |
| **Reinvestment rate** | (Net capex + acquisitions + Δ NWC + capitalised growth spend) ÷ NOPAT | High is good *only* when ROIIC is high; high with weak ROIIC is the worst combination | Determines how much of the return compounds internally |
| **Sustainable growth rate** | Reinvestment rate × ROIC | Should roughly reconcile to guided growth | Exposes guidance that is arithmetically impossible |
| **Backlog coverage** | Backlog ÷ trailing 12-month revenue, in months or years | Sector-dependent; the *trend* and de-booking rate carry the signal | Forward visibility, and the first place a slowdown appears |
| **Book-to-bill** | Orders ÷ billings for the period | >1.0; below 1.0 for consecutive quarters is a real warning | Leads reported revenue by two to four quarters |
| **RPO / cRPO growth** | Remaining performance obligations, and the portion due within 12 months | cRPO growth ≥ revenue growth | Detects subscription growth rolling over before revenue does |
| **Capex ÷ depreciation** | Total capex ÷ depreciation and amortisation | ~1.0 maintenance-only; 1.5–2.5 genuine expansion; sustained >2 with flat returns is a warning | Separates real expansion from treadmill spending |
| **Growth vs maintenance capex** | Disclosed split, or estimate maintenance as inflation-adjusted depreciation | Growth capex tied to specific projects with stated returns | Reveals the true free cash flow of the business as it stands |
| **Vitality index** | % of revenue from products launched in the last 3–5 years | 15–30% for consumer and industrial innovators; higher in pharma/tech | Leading indicator of organic growth; falling means living off legacy |
| **R&D productivity** | Incremental gross profit ÷ trailing R&D spend | Rising, or at least stable | R&D spend is an input; only the return on it is evidence |
| **Guidance hit rate** | Beat/meet/miss across trailing 8–12 quarters, plus multi-year targets set vs achieved | Consistent meet-or-modest-beat, no serial in-year cuts | The best available prior on the credibility of the forward case |
| **Share count growth** | Diluted share count CAGR, including SBC and convertibles | ≤0 for mature businesses; <2–3%/yr for growth names | Aggregate growth means nothing if per-share growth is diluted away |
| **SBC as % of revenue** | Stock-based compensation ÷ revenue | Low single digits outside early-stage software | A real cost, routinely adjusted out, and a direct transfer from shareholders |
| **Self-funding test** | FCF after growth capex; funding gap = reinvestment need − OCF | Non-negative, or a credible dated path to it | Growth dependent on open capital markets is fragile by construction |
| **Absolute revenue add required** | Current revenue × target growth rate, in currency | Small relative to the realistic addressable market | The law of large numbers, made concrete |
---
## Verify against something the company did not write
Everything above can be answered entirely from company-produced documents — which is exactly how a fabricated or deteriorating business passes a desk-based checklist. Add at least one independent cross-check for any thesis you intend to act on.
- **Alternative data.** Web traffic and app-download trends, app-store and review-site ratings, job postings and headcount (LinkedIn), employee-review sentiment and attrition, card/transaction panels, freight and customs data. **India-specific proxies:** GST and e-way-bill volumes, VAHAN vehicle-registration data for autos, SIAM dispatch numbers, DGCA traffic data for airlines, AWACS/IQVIA secondary sales for pharma, RBI sectoral credit data, port and rail freight statistics. Divergence between reported growth and every external proxy for it is the earliest tell available.
- **Scuttlebutt.** Use the product. Read customer reviews. Talk to customers, distributors, suppliers or ex-employees where feasible. Financial statements are lagging; ground-level observation moves quarters earlier and is one of the few genuine edges available.
- **Proof of existence.** For asset-heavy or geographically remote claims, corroborate that the plants, mines, stores or data centres exist and operate — satellite imagery, environmental clearances and building permits, power draw, shipping and customs records, statutory filings of subsidiaries (MCA/ROC filings in India) reconciled to consolidated claims.
- **Per-employee productivity.** Revenue and gross profit per employee versus peers and over five-plus years. Hard to fabricate, and it quietly validates or refutes claimed scale. A hiring freeze that contradicts a growth story is an independent warning.
- **Bespoke KPI audit.** Catalogue every company-invented metric (ARR, bookings, GMV, MAU, "cash EBITDA", contribution margin ex-marketing, order book) with its exact definition, and check whether the definition changed year over year and whether it reconciles to audited revenue or cash. A shifting denominator keeps a growth narrative alive after the audited numbers have rolled over.
- **Year-over-year redline.** Diff consecutive annual reports and 10-Ks for changes in risk-factor wording, segment definitions, accounting-policy and useful-life assumptions, and — most valuable — disclosures and metrics quietly *dropped*. Management highlights what it adds and never what it removed. Cheap, and high-yield.
- **Vendor and customer financing.** Check whether the company lends to, guarantees, or extends unusually long credit to its own customers or channel to enable purchases (captive finance arms, channel financing, seller notes). Revenue funded by the seller's own credit reads as clean organic growth until the receivables sour.
- **Read the other side.** Find the strongest existing bear case — short-seller reports, sceptical sell-side notes, forum criticism with actual numbers in it — and state what the informed money on the other side sees before deciding it is wrong.
---
## Where this generic frame breaks
Apply the sector playbook before using any of Part C's growth metrics. In these sectors the generic frame is not merely less useful — it is inverted or undefined.
- **Banks and lenders** — above-system loan growth is the best leading indicator of the *next* credit cycle's losses, not an achievement. There is no meaningful capex or FCF, so reinvestment rate and capex/depreciation are noise; the reinvestment constraint is regulatory capital (CET1), and sustainable growth = ROA × leverage × retention. The moat is the deposit franchise, not the loan book. See `references/sectors/banks.md` and `nbfc.md`.
- **Insurers** — top-line premium growth can be bought by underpricing risk, and the loss shows up years later. Growth is measured in VNB, APE and new-business margin, not revenue. See `references/sectors/insurance.md`.
- **REITs, InvITs and real estate** — growth is same-store NOI plus acquisitions, and acquisitions only create value if the cap rate exceeds the cost of capital. Use AFFO and NAV, not EPS. See `references/sectors/realestate-reit.md`.
- **Miners, commodity producers, refiners** — there is no pricing power by construction; price is exogenous. The moat is position on the industry cost curve, reserve life and grade. TAM analysis is meaningless; the capital cycle in B7 is the whole game. See `references/sectors/metals-mining.md` and `oil-gas.md`.
- **Regulated utilities** — growth is rate-base growth at an allowed return, so the regulator is the business model. See `references/sectors/utilities-power.md`.
- **Holdcos and conglomerates** — analyse the underlying businesses separately and value sum-of-the-parts; group-level growth and margin are aggregation artefacts. See `references/sectors/holdco-assetmgr.md`.
---
## Checklist
**Business model and moat**
- [ ] Business and profit engine stated in two plain sentences, with revenue *and gross profit* split by segment/product.
- [ ] Monetisation architecture named; who sets price identified; take rate quantified for platforms.
- [ ] Recurring vs one-off revenue split; NRR, gross retention and churn sourced, not assumed.
- [ ] Unit economics modelled end to end: contribution margin, LTV/CAC, payback — and tested without subsidy.
- [ ] Pricing power evidenced by realised price vs volume and by gross margin through a cost shock.
- [ ] Switching costs quantified in money and time; retention used as the empirical test.
- [ ] Network effects classified (direct/indirect/data) and multi-homing checked.
- [ ] Brand tested for price premium and behaviour change, not awareness.
- [ ] Cost advantage verified in the margin gap versus smaller peers, not asserted.
- [ ] Patents, licences and exclusivity inventoried with revenue-weighted expiry dates.
- [ ] Top-1/5/10 customer concentration pulled from the segment note (ASC 280 / Ind AS 108); channel concentration included.
- [ ] Sole-sourced critical inputs and single-geography supply identified.
- [ ] Revenue and EBIT split by geography; single-country dependence and FX exposure flagged.
- [ ] Moat width, durability **and direction** stated, with the primary disruption vector named.
**Industry and competitive dynamics**
- [ ] TAM traced to its source and its build method; implied share needed back-solved from the current price.
- [ ] Industry growth decomposed and compared to nominal GDP; pull-forward demand identified.
- [ ] Concentration measured (CR4/HHI) *and* competitor rationality assessed.
- [ ] Market-share trend traced over 5–10 years, with the mechanism of share movement identified.
- [ ] Supplier power, buyer/channel power, entry barriers and substitutes each assessed concretely.
- [ ] Position in the cycle established; mid-cycle normalised margin estimated before any multiple is applied.
- [ ] Industry capacity additions, utilisation and capex/depreciation checked — the supply side, not just demand.
- [ ] Regulatory regime mapped; subsidy- and policy-dependent profit quantified (India: DPCO/NPPA, PLI, tariff orders).
- [ ] Profit-pool location and migration direction identified; peer ROIC dispersion vs WACC compared.
- [ ] Commoditisation tested via realised price vs input inflation and premium vs private label.
**Growth and reinvestment runway**
- [ ] Revenue growth decomposed: organic vs inorganic, then volume/price/mix/FX, by segment and geography, YoY and QoQ.
- [ ] EPS growth decomposed; buyback, tax and one-off contributions separated; GAAP-to-adjusted bridge read.
- [ ] Reinvestment rate and ROIIC computed over rolling windows and compared to WACC and to average ROIC.
- [ ] Sustainable growth (reinvestment rate × ROIC) reconciled against guided growth.
- [ ] Backlog/order book/RPO, coverage, book-to-bill and de-booking checked; definition changes caught (India: unaudited).
- [ ] Capex split into maintenance and growth; announced projects checked against prior announcements for overrun and slippage.
- [ ] Forward growth bridge built driver by driver in percentage points, and the drivers actually sum to the target.
- [ ] Vitality index and R&D productivity computed; patent cliff dated.
- [ ] Track record of prior adjacency and geographic expansions assessed.
- [ ] Guidance-vs-actual scorecard built over 8–12 quarters (India: concall commitments vs outcomes).
- [ ] Capital allocation history judged: M&A prices and post-deal ROIC, impairments, buyback valuations.
- [ ] Absolute revenue add required to sustain the growth rate computed; leading indicators checked for deceleration.
- [ ] Base rate for the claimed growth durability checked against the reference class, not just the model.
- [ ] Funding gap, share-count growth and self-funding test completed; per-share growth compared to aggregate growth.
**Independent verification**
- [ ] At least one non-company data source cross-checked against the reported growth curve.
- [ ] Bespoke KPIs catalogued and definition changes checked.
- [ ] Consecutive filings redlined for dropped disclosures.
- [ ] The strongest bear case read and answered on its own terms.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |