15.2 · Templates

Why this matters: a template is a checklist that looks like a document. It makes sure the same questions are asked of every company, so that cases can be compared with each other and with their outcomes. It also makes it obvious when a question has been skipped. The templates below are the ones in the casebook's templates/ folder, reproduced here so you can see them without cloning the private repository.

Learning objectives — after this lesson you can:

• Say which template belongs to which stage of the case-study method, and what each field is for.
• Fill a field well rather than just filling it, using the example lines given for each template.
• Scaffold a new case in the casebook with one command.

Prerequisites: 15.1 The case-study method · Time: ~45 min

---

1. The set

Stage	Template	File in the casebook	Worked on Kaveri in
1	Kill test	templates/00-kill-test.md	15.3 §2
2–3	Full case study	templates/01-full-case-study.md	15.3 §3–4
3	Forensic checklist (40 tests)	templates/06-forensic-checklist.md	09.7
4	One-page thesis	templates/02-one-page-thesis.md	15.3 §5
4	Decision journal	templates/04-decision-journal.md	15.3 §5
5	Quarterly update	templates/03-quarterly-update.md	15.3 §6
6	Post-mortem	templates/05-post-mortem.md	15.3 §7

To start a case, run this from the casebook root:

python tools/new_case.py GENUSPOWER.NS "Genus Power Infrastructures"

It creates cases/GENUSPOWER.NS/ from the templates, filled with the company name, ticker and date. Then pull the data with python tools/pull.py GENUSPOWER.NS, which prints the ratio dashboard and forensic quick scores that the kill test asks for.

---

2. Stage 1 — kill test

Thirty minutes, six boxes, one verdict. The discipline is in box 6: you must tick a reason for mispricing. "None obvious" is a legitimate answer, and it usually means KILL.

Field	Filled badly	Filled well
Business in three sentences	"Leading pump maker with strong brand"	"Makes agricultural and industrial pumps and motors, sold through 1,800 dealers in 14 states; since FY23 also sells solar pumping systems to state agencies under PM-KUSUM, now 27% of revenue."
Why mispriced	"Undervalued"	"Misunderstanding: the market capitalises tender-driven solar growth at an agri-franchise multiple."
Verdict	"Looks interesting"	"TRIAGE — the one question: is the solar growth creating or destroying value, given 96 receivable days?"

Kill test — templates/00-kill-test.md

# Kill test — {{COMPANY}} ({{TICKER}})

Date: {{DATE}} · Price: ₹ · Market cap: ₹ Cr · Source: Screener.in + `tools/pull.py`

## 1. The business in three sentences

1.
2.
3.

## 2. Numbers (5 years)

| Metric | Value | Comment |
|:--|--:|:--|
| Revenue CAGR (5y) | | |
| EBITDA margin (range) | | |
| ROCE (range) | | |
| CFO / PAT (avg) | | |
| Net debt / EBITDA | | |
| Receivable days (trend) | | |
| Dilution (share count 5y) | | |

## 3. People

| Check | Finding |
|:--|:--|
| Promoter holding & 3-year trend | |
| Pledge (% of promoter holding) | |
| Auditor changes / resignations (5y) | |
| SEBI / regulatory actions | |
| Related-party transactions (% of revenue or cost) | |
| Promoter remuneration (% of PAT) | |

## 4. Forensic quick scores (from `tools/pull.py`)

| Score | Value | Threshold | Flag? |
|:--|--:|--:|:--|
| Beneish M (8-var) | | > −1.78 | |
| Piotroski F | | < 4 | |
| Accruals ratio | | > 10% | |
| Implied yield on cash | | < 3% with high cash | |

## 5. Valuation snapshot

| Multiple | Now | 5y median | Peers (3) |
|:--|--:|--:|--:|
| P/E (TTM) | | | |
| EV/EBITDA | | | |
| P/B | | | |

## 6. Why might it be mispriced?

- [ ] Neglect / low coverage   - [ ] Complexity   - [ ] Forced selling   - [ ] Cyclical trough
- [ ] Misunderstanding of the business   - [ ] None obvious

## Verdict

- [ ] **KILL** — reason:
- [ ] **TRIAGE** — the one question triage must answer:

---

3. Stages 2–3 — the full case study

This file starts as the triage section and grows into the complete study. Its sections follow the course's modules, so each one has a lesson behind it. The two sections most often skimped are:

• section 3 (reading the filings), where the notes and the say-do table live;
• section 5 (scuttlebutt), the only evidence the market may not already have.

Full case study — templates/01-full-case-study.md

# {{COMPANY}} ({{TICKER}}) — full case study

Started: {{DATE}} · Analyst: Pravar · Status: triage / deep dive / memo / monitoring / closed

> One-paragraph summary of the situation and the question this study answers.

## 0. Sources used

| Document | Date | Link / path | Notes |
|:--|:--|:--|:--|
| Annual report FY | | | |
| Concall transcripts (Q… to Q…) | | | |
| Credit-rating rationale | | | |
| DRHP / offer documents (if any) | | | |
| Regulator / industry data | | | |

## 1. Business (Module 05)

- What it sells, to whom, how it makes money; unit economics.
- Segment table (revenue, margin, growth, share of profit).
- Value chain position; customer and supplier concentration.
- Industry structure (five forces summary), capital cycle stage, regulation.
- Moat: source, evidence in the numbers (ROIC > WACC for how long? pricing power? share trend?), erosion risks.

## 2. History and numbers (Module 04)

Paste the ratio dashboard from `tools/pull.py` (10 years where available).

| Metric | Y-9 | … | Y-1 | Y0 | Comment |
|:--|--:|--:|--:|--:|:--|

- Growth decomposition (volume × price × mix; organic vs acquired).
- Margin structure and operating leverage.
- Returns on capital (DuPont; incremental ROIC over 3/5 years).
- Working capital and cash conversion (CCC trend; CFO/EBITDA; FCF).
- Leverage, liquidity, maturity profile, off-balance-sheet items.
- Quality of earnings scorecard.

## 3. Reading the filings (Module 03)

- Auditor's report: opinion type, KAMs, EoM, CARO remarks.
- Notes that matter: related parties, contingent liabilities, borrowings & covenants, receivables ageing, CWIP ageing,
  revenue policy changes, exceptional items, tax reconciliation.
- Concall themes over 8 quarters; say-do table.

| Quarter | Guidance / claim | Delivered | Score |
|:--|:--|:--|:--|

## 4. Forensics and governance (Modules 05.6, 09)

- Forensic checklist score (attach `forensic-checklist.md`): overall RAG and the top three items.
- Governance checklist: board, auditor, RPTs, pledges, remuneration, capital-raising history, group structure.
- Standalone vs consolidated cash; loans to subsidiaries/related parties.

## 5. Scuttlebutt and data checks (Module 05.7)

| Date | Source (type, no names needed) | What I asked | What I learned | Confidence |
|:--|:--|:--|:--|:--|

Alternative data / government data checked:

## 6. Valuation (Modules 06–07)

- Model: `model/` (three statements, base/bull/bear; assumptions table).
- DCF summary: WACC, terminal growth, value/share by scenario, TV share of EV, sensitivity grid.
- Multiples: vs own history bands and peer table.
- **Reverse DCF:** the price implies … ; my base case is … ; the gap is explained by …
- Probability-weighted value and payoff asymmetry (upside/downside ratio, expected value).

## 7. Thesis (Module 11)

- **Variant perception** (one sentence).
- Thesis in three bullets.
- Catalysts with expected timing.
- Key risks; pre-mortem ("it is 2029 and this lost 50% — why?").
- **Thesis-breakers** (specific, observable, dated).

## 8. Decision

- Action, size (quarter-Kelly vs max-loss → lower; forensic overlay; liquidity check), entry plan, monitoring KPIs and schedule.
- Journal entry: `journal/{{DATE}}-{{TICKER}}.md`.

## 9. Updates

Quarterly updates live in `updates/`. Log of material changes to the thesis:

| Date | Change | Action |
|:--|:--|:--|

Section 4 attaches the forensic checklist: the 40-test sheet from 09.7, with the same thresholds and RAG bands:

• Green: 12 or less, with no 3s and at most two 2s;
• Amber: 13–30, with no 3s;
• Red: any 3, or over 30.

The template has blank score and note columns. The Kaveri copy in cases/KAVERIPMP/ is fully scored, at 18 (Amber).

---

4. Stage 4 — one-page thesis and decision journal

The thesis is the memo of 11.3 compressed to one page. The journal records the decision and the state of mind it was taken in, before the order is placed.

Field	Filled badly	Filled well
Thesis-breaker	"If fundamentals deteriorate"	"Receivable days at 30-Sep-2026 (H1 balance sheet, due Nov-2026) below 80, with > 6-month overdues falling"
Confidence	"High"	"65% that the price is within the base-case range (₹280–₹340) in two years"
Mood	(left blank)	"Irritated at missing the 2024 rally; watch for wanting to be 'in' something"

One-page thesis — templates/02-one-page-thesis.md

# {{COMPANY}} ({{TICKER}}) — one-page thesis

Date: {{DATE}} · Price ₹ · Mcap ₹ Cr · Recommendation: BUY / WATCH / AVOID · Horizon: … · Size: …% of portfolio

**What the market believes:** …

**What I believe (variant perception):** …

**Why I'm right (three bullets, each with a number):**
1. …
2. …
3. …

**Key numbers**

| | FY-2 | FY-1 | FY0 | FY+1E | FY+2E |
|:--|--:|--:|--:|--:|--:|
| Revenue (₹ Cr) | | | | | |
| EBITDA margin | | | | | |
| PAT (₹ Cr) | | | | | |
| ROCE | | | | | |
| CFO / PAT | | | | | |
| Net debt / EBITDA | | | | | |

**Valuation:** base ₹… (method), bull ₹…, bear ₹…, probability-weighted ₹…; price implies … growth / … margin.
Upside/downside ratio: …

**Catalysts:** (1) … by …; (2) … by …

**Thesis-breakers — I am wrong if:** (1) …; (2) …; (3) …

**Monitoring:** KPIs …, sources …, cadence …

**Pre-mortem (one line):** …

**Position management:** entry plan …; add if …; trim if …; exit if …

Decision journal — templates/04-decision-journal.md

# Decision — {{DATE}} — {{TICKER}} — {{ACTION}}

**Decision:** buy / add / trim / sell / pass — … shares at ₹… (≈ ₹… , …% of portfolio)

**What I believe:** (thesis in two lines)

**What the market believes:** 

**Why the market is wrong / why now:** 

**What I expect to see, and by when:**

| Expectation | By when | How I'll know |
|:--|:--|:--|

**What would make me reverse this decision:** 

**Sizing logic:** quarter-Kelly … vs max-loss … → lower … → forensic overlay … → chosen …; liquidity check: ADV ₹… Cr, position = … days

**Confidence (0–100):** …   **Mood / context (be honest):** …

**Pre-mortem:** it is three years later and this went badly because …

---
*Review this entry at the next quarterly update and at exit. Grade the process, not the outcome.*

---

5. Stage 5 — quarterly update

One hour, filled in before you read the broker notes or look at the price reaction. The thesis-breaker table carries forward from the thesis unchanged. You may not rewrite a breaker because it fired.

Quarterly update — templates/03-quarterly-update.md

# {{COMPANY}} ({{TICKER}}) — quarterly update {{QUARTER}}

Date: {{DATE}} · Price ₹ · Position: …% · Cost ₹ · P&L since entry: …%

## Results vs expectations

| Metric | Expected | Actual | Delta | Comment |
|:--|--:|--:|--:|:--|
| Revenue | | | | |
| EBITDA / margin | | | | |
| PAT | | | | |
| KPI 1 | | | | |
| KPI 2 | | | | |
| Working capital / cash | | | | |

## Concall: what management said (and didn't)

- Guidance change:
- Say-do update (add a row to the case file):
- Questions not answered / evasive answers:

## Thesis-breaker check

| Breaker | Status (green/amber/red) | Evidence |
|:--|:--|:--|

## Valuation now

Price vs scenarios (bear ₹… / base ₹… / bull ₹…); what the price implies now.

## Action

- [ ] Hold   - [ ] Add   - [ ] Trim   - [ ] Exit
Why (two lines):
Journal entry: `journal/{{DATE}}-{{TICKER}}.md` (only if action ≠ hold)

---

6. Stage 6 — post-mortem

Write it on exit, or three years after any decision, including decisions to pass. The key box is "one change to the process". A post-mortem that changes nothing has only told a story.

Post-mortem — templates/05-post-mortem.md

# Post-mortem — {{COMPANY}} ({{TICKER}})

Entry: {{DATE}} at ₹… · Exit: … at ₹… · Holding period: … · Return: …% (vs Nifty 500 …%) · Peak/trough drawdown: …

## What I said would happen vs what happened

| Expectation (from the journal) | Outcome | Right / wrong / partly |
|:--|:--|:--|

## Which scenario played out, and why

## Was the result process or luck?

- Process errors (analysis, sizing, timing, behaviour):
- Bad/good luck (macro, unforecastable events):

## What was knowable at entry that I missed?

## One change to the process

(The single filter, checklist item or rule that would have caught this earliest.)

## Lessons mapped to course modules

| Lesson | Module |
|:--|:--|

---

Key terms

Term	Meaning
Template	A fixed set of questions in document form, used for every case so that cases can be compared
Scaffold	The case folder created from the templates by tools/new_case.py
Say-do table	Guidance against delivery, quarter by quarter; lives in section 3 of the case study
Thesis-breaker	A dated, observable test that would falsify the thesis; carried unchanged into every update

Check your understanding

1. Which template do you open after a quarterly result, and what must you not do before filling it in?

Answer

`03-quarterly-update.md`. Do not read the sell-side's take or look at the price reaction first. Otherwise the update records the market's interpretation instead of testing your thesis.

1. Rewrite "management seems competent" as a case-file entry that can be checked.

Answer

A say-do row, for example: "May-2025: guided FY26 growth 15–18% and margin 15%+; delivered 12.5% and 13.8% — miss on both." Competence is then scored from the record, not asserted.

1. Why does the kill test ask for the reason a stock might be mispriced, and not just whether it looks cheap?

Answer

A low multiple usually has a reason the market already knows. Naming a specific source of mispricing — neglect, complexity, forced selling, a cyclical trough or a misunderstanding — is what makes the idea researchable and falsifiable. Without one, there is no edge to find.

Go deeper

• Atul Gawande, The Checklist Manifesto — why checklists work in complex, expert domains.
• The casebook's PROCESS.md — the operational rules each template serves.

---

← Previous: 15.1 The case-study method · Module index · Next: 15.3 Worked example — Kaveri →