11.3 · Writing an investment thesis & memo¶

Why this matters: an investment memo is where research becomes a decision. Writing it forces the variant perception into one sentence, the valuation into a range with named drivers, the risks into a pre-mortem, and the future into thesis-breakers with dates. A memo that cannot be written is a position that should not be taken. This lesson gives the structure and a complete sample memo on Kaveri Pumps — whose conclusion, consistent with the whole course, is "not at this price".

Learning objectives — after this lesson you can:

Write a thesis in three bullets and a variant perception in one sentence.
Structure a memo: key value drivers and KPIs, valuation summary, catalysts, risks and pre-mortem, thesis-breakers, monitoring plan, recommendation and size.
Write the "I'm wrong if…" section so that it is falsifiable and dated.
Produce a one-page version for decisions and a full version for the case file.

Prerequisites: 06.8 Margin of safety, 09.7 The forensic checklist, 11.2 The research process · Time: ~75 min

1. The structure¶

Section	Length	Content	Test
Header	3 lines	Company, ticker, date, price, market cap; recommendation (Buy / Watch / Avoid / Sell); target range; horizon; proposed size	Could a reader act on this line alone?
What the market believes	2–3 sentences	The consensus narrative and the expectations implied by the price (from the reverse DCF)	Is it stated fairly — the strongest version, not a straw man?
Variant perception	1 sentence	What you believe that differs, and the evidence class it rests on	Would a sceptic know exactly what you are claiming?
Thesis	3 bullets	Each bullet: a claim + a number + the source	Is each bullet checkable?
Key value drivers and KPIs	Table	The 3–5 drivers that move value (from the tornado), current level, your forecast, consensus/implied, the KPI that tracks each	Does each driver have a monitoring KPI?
Valuation	Table + 3 sentences	Base/bull/bear with probabilities and weighted value; the key sensitivities; what the price implies; multiples cross-check	Is the range honest and the asymmetry stated?
Catalysts	List with dates	Events that would move the price toward value, and when	Are they specific and dated?
Risks and pre-mortem	List + paragraph	The bear case's mechanisms; "it is three years later and this lost 50% — why?"	Does the pre-mortem name things the thesis section did not?
Thesis-breakers ("I'm wrong if…")	3–5 items	Observable, dated tests that would falsify the thesis	Would you actually sell if one fired?
Monitoring plan	Table	KPI, source, frequency, threshold	Can it be done in an hour a quarter?
Forensic and governance summary	3 lines	RAG rating; top items; adjustments made to reported numbers	Is the discount applied somewhere specific?
Recommendation and sizing	3 lines	Action, size logic (from 11.4), entry plan
Appendices (full version)		Model summary, checklist scores, scuttlebutt log, sources

The one-page version is the header through thesis-breakers, compressed; the case file has everything.

2. Writing the variant perception¶

Bad: "Kaveri is a quality pump company that will benefit from solar." (Not variant; not testable.) Better: "The market expects ~14% growth for ten years; I expect 9–11%." (Variant, testable — but why?) Good: "The market prices Kaveri (₹390) for ~14% revenue growth over ten years with margins recovering to 15.5%; I expect 9–11% because the solar segment that produced the FY23–26 surge is tender-driven, concentrated in two slow-paying states and now 27% of sales, and because management has missed its receivables guidance four quarters running — evidence the market is under-weighting because the headline growth still looks intact. I will know by Q3 FY27 (Feb-2027)."

The form: market expects X · I expect Y · because Z (evidence the market under-weights) · value if right · known by when. Every clause is necessary.

3. Sample memo — Kaveri Pumps & Motors (fictional), 22-Sep-2026¶

Kaveri Pumps & Motors Ltd (KAVERIPMP, fictional) · Price ₹390 · Mcap ₹2,340 Cr · Recommendation: WATCH (do not buy above ₹260; avoid at current price) · Value range ₹168–416, probability-weighted ₹306 · Horizon 2–3 years · Proposed size at ₹260: 4% of portfolio

What the market believes. Kaveri is a well-run Coimbatore pump maker riding the PM-KUSUM solar-pump opportunity, with FY23–26 revenue CAGR of 14.7%, a new motors plant, and a temporary working-capital bulge from state agencies that will normalise. At ₹390 (25.9x trailing EPS, 13.7x EV/EBITDA) the price implies ~14.1% revenue growth for ten years with EBITDA margins recovering to 15.5% and receivables normalising to ~80 days — essentially a continuation of FY23–26.

Variant perception. The FY23–26 growth was a mix shift into a segment with structurally worse economics — tender pricing, concentrated slow-paying state buyers, bank-guarantee intensity — and the market is capitalising that growth at a franchise multiple. I expect 9–11% growth with margins plateauing near 14%, receivables normalising only partly, and a ₹10–15 Cr provisioning catch-up; management's four consecutive misses on receivable guidance and the withdrawal of the KPI in Q1 FY27 are the evidence the market under-weights.

Thesis (why I am right). 1. The growth engine is the weak segment. Solar went from 7% to 27% of revenue (FY23–26) while agri grew ~5% a year; solar carries ~22–24% gross margins vs ~40% agri, and 6–12-month payment cycles vs ~60 days (segment note; running-example §5, §8). 2. The cash has not followed the profit. CFO/PAT fell from 100% (FY24) to 63–72% (FY25–26); receivable days rose 61 → 96; > 6-month overdues doubled to ₹62 Cr with 6.4% coverage; FCF was ₹13 Cr on ₹90 Cr of PAT (§4, §6, §8). 3. Returns on new capital are below cost. Incremental ROIC FY24–26 was 4.6% vs a 12.2% WACC; ROIC fell from 15.1% to 12.0%; the reference DCF's ₹320 already assumes a recovery to 15.5% margins and 18% terminal RONIC.

Key value drivers and KPIs.

Driver	FY26	My base FY27–29	Price-implied	KPI / source
Revenue growth	12.5%	9–11%	~14%	Quarterly results; segment revenue
EBITDA margin	13.8%	13.5–14.5%	15.5%	Results; gross margin vs copper
Receivable days	96	85–90	~76–80	Half-yearly BS; ageing note; management disclosure (withdrawn in Q1)
ECL coverage of > 6m overdues	6.4%	25% (my adjustment)	—	Annual report note
Hosur motors utilisation	57%	65–70%	75%+	Annual report; concall
RPT purchases (% of materials)	10.4%	stable	—	RPT note

Valuation.

	Bear	Base	Bull	Weighted	Price
Value / share	₹168	₹320	₹416	₹306 (25/50/25) or ₹286 (35/45/20)	₹390
Return from ₹390	−57%	−18%	+7%	−22%

Most sensitive to terminal margin and RONIC (±₹40/share each), then WACC (±₹27 per 50 bps); the working-capital normalisation is worth ₹23/share and is the near-term test. The whole WACC × g grid (11–13%, 4–6%) sits below ₹390. Peers trade at 30–40x (median 35x) on higher ROCE; Kaveri's ~26x is roughly fair within the group after adjusting for returns and receivables risk — the group itself is priced for growth the value-driver formula cannot support at a 12% cost of capital. Reverse DCF: ₹390 requires 14.1% growth for ten years (holding base margins), or a 17.4% flat margin, or an 11.1% WACC.

Catalysts (toward value, i.e. downward from here). Q2 FY27 results (Nov-2026): receivable days and any provision; Q3 FY27 (Feb-2027): confirmation or refutation of the 15–18% guidance; FY27 annual report (Jul-2027): ageing schedule and ECL; any SAST filing on the pledge. Upward catalysts (bull case): a state clearing dues; an industrial-motor contract win; a copper price fall.

Risks and pre-mortem. Bear mechanisms: state dues stuck → provisions and working-capital debt → growth cut → de-rating to 15–18x on lower EPS (₹13 × 16 = ~₹210). Pre-mortem: "It is September 2029 and Kaveri is at ₹180. The two state agencies disputed ₹40 Cr of claims in FY27; the auditor qualified; the promoter's real-estate venture needed cash and the pledge rose to 25%; a rival won the FY28 tenders at lower prices; Hosur never passed 65% utilisation; the stock de-rated to 14x." Risks to a short or to staying out: a large state pays in full (stock +15% in a day); management is right about 15–18% growth; the Indian small-cap multiple regime stays elevated.

I'm wrong if… (thesis-breakers). 1. Receivable days at 30-Sep-2026 (H1 balance sheet, due Nov-2026) are below 80 with > 6-month overdues falling. 2. Q2–Q3 FY27 revenue growth is above 14% YoY with EBITDA margin above 14.5%. 3. Provisioning is raised to ≥ 20% of overdues and the stock does not fall (the market already priced it). 4. Hosur utilisation is disclosed above 70% by FY27 year-end. Any two of these by Feb-2027 and the bear probability drops to 15%; the weighted value rises to ~₹330 and the verdict becomes "fair, not cheap".

Monitoring plan. Quarterly: revenue by segment, EBITDA margin, receivables (half-yearly BS), provisioning commentary, Hosur utilisation, guidance vs delivery (say-do table), pledge (SAST filings, continuous), RPT share (annual), auditor's report (annual). One hour per quarter; template in the casebook.

Forensic and governance. Checklist Amber (18 points, six 2s, no 3s): revenue/receivables cluster; RPT trend; new pledge. Adjustments: EPS ₹15.1 → ₹13.6 for provisioning; bear probability 25% → 35% for governance; WACC unchanged. No evidence of fabrication.

Recommendation and sizing. Watch. Entry trigger ₹260 (expected return +18%, upside/downside 3.0 on the reference scenarios) with receivable days < 85; at that price and evidence, size 4%: quarter-Kelly on a scenario tree that includes a 10–15% governance-jump tail gives ~5–12%; the 3% max-loss rule on a −35% bear gives ~8.5%; take the lower and halve it for the Amber forensic rating (11.4). Kill: pledge > 15%; auditor qualification; CFO exit.

That is the whole memo. Note what it does not contain: a target price to the rupee, a paragraph about India's irrigation potential, or a recommendation the numbers do not support.

4. Writing well¶

Lead with the conclusion. Header first; the reader may stop there.
Numbers with sources in every claim; no adjectives doing the work of evidence.
The strongest opposing case, stated fairly, in "what the market believes" and in "risks".
Falsifiable, dated thesis-breakers. "If execution disappoints" is not one; "receivable days > 90 at Sep-2026" is.
One page for the decision, the rest for the file. A memo nobody finishes is not a memo.
Date and freeze it. Later updates are new documents that reference the original; the original is the record of what you believed and why.

Trader's lens

The memo is a trade ticket with a thesis attached: instrument, direction, size, entry, stop (thesis-breakers), target (value range), and the reason. Traders who write the ticket before the trade fill fewer bad ones; the memo is the same pre-commitment, with the added discipline that the "stop" is a fact about the company, not a price.

Common mistakes

A thesis that is a description of the company.
No variant perception (the memo agrees with the price and recommends buying anyway).
Risks listed as boilerplate ("execution risk, competition, macro").
Thesis-breakers that are unobservable or undated.
A target price with no distribution.
Writing the memo to justify a position already taken.

Key terms¶

Term	Meaning
Investment memo	The written case for a decision: thesis, drivers, valuation, catalysts, risks, breakers, sizing
Variant perception	The specific belief that differs from price-implied expectations, with evidence and a test
Thesis	The three claims, each with a number and a source, on which the case rests
Catalyst	A dated event expected to move price toward value
Pre-mortem	A narrative of how the position failed, written before it is taken
Thesis-breaker	An observable, dated outcome that would falsify the thesis and trigger exit
Monitoring plan	The KPIs, sources, frequency and thresholds used to track the thesis
One-pager	The compressed decision version of the memo

Check your understanding¶

Rewrite "Kaveri has strong growth prospects in solar pumps" as a thesis bullet.

Answer

"Solar pumping systems grew from ₹61 Cr (7% of revenue) in FY23 to ₹356 Cr (27%) in FY26 on state PM-KUSUM tenders; the unexecuted solar order book is ₹410 Cr (FY26 annual report, notes) — which supports ~₹400–450 Cr of FY27 solar revenue if execution and payments hold." Claim, number, source, and the condition.

Why does the sample memo put "what the market believes" before the thesis?

Answer

Because the thesis is only valuable relative to the price; stating the market's view (and the implied expectations from the reverse DCF) first defines what the memo must disagree with, and forces the author to represent the opposing case fairly.

Write one more thesis-breaker for the Kaveri memo.

Answer

"The FY27 annual report shows RPT purchases from Kaveri Castings falling below 9% of materials with disclosed supplier margins in line with third-party foundries, removing the margin-migration concern" — observable, dated (Jul-2027), and it would raise the base probability.

What distinguishes a pre-mortem from a risk list?

Answer

A pre-mortem is a specific causal story told from the failed future, which surfaces interactions (state dispute → provisions → pledge rises → auditor) that a list of independent risks misses; it also produces the thesis-breakers directly.

Go deeper¶

Michael Steinhardt, No Bull — the origin of "variant perception".
Gary Klein, "Performing a Project Premortem", Harvard Business Review (2007).
Howard Marks, memos (Oaktree) — models of writing that leads with the conclusion and states the opposing case.
The casebook's templates/02-one-page-thesis.md and 01-full-case-study.md — the memo as a form.

← Previous: 11.2 The research process · Module index · Next: 11.4 Position sizing & portfolio construction →