Most equity research is a forecast wearing a spreadsheet. The model projects revenue, the analyst picks a multiple, and the price target is the output of two assumptions that were never independently justified. It reads as rigor and functions as opinion.

The alternative is to work backwards. Start with the price, extract the growth and returns the market is already paying for, and then ask a much narrower question: is that expectation too high or too low, and what evidence would settle it? That framing survives being wrong, because being wrong means a specific expectation failed, not that a forecast was off.

01The four questions

Every company note answers the same four questions in the same order. If a note cannot answer the first one, the rest is decoration.

  1. 01How does this business convert effort into cash?Unit economics, capital intensity and the reinvestment runway. Not the story — the mechanics.
  2. 02What are normal earnings?Mid-cycle margins and maintenance capital, adjusted for one-offs that flatter or depress the reported figure.
  3. 03What is already in the price?The growth and returns implied by the current multiple, derived rather than asserted.
  4. 04What would prove me wrong?Specific, observable evidence that would change the view — written before the position, not after.

02Business quality

Quality is not a multiple. It is the ability to reinvest at high returns for a long time, and it shows up in a small number of places: incremental return on invested capital, the stability of gross margin through a cycle, the share of revenue that recurs, and the bargaining power visible in working capital.

I look for evidence rather than adjectives. A wide moat should be visible as pricing power in a downturn, switching costs in retention data, or scale advantages in a unit cost curve that declines with volume. If the moat only appears in the management presentation, it is not a moat.

What I check first

  • Five years of incremental ROIC, not just the level.
  • Gross margin behavior in the last two downturns.
  • Capitalized costs, lease treatment and stock compensation — the three places where cash flow can be flattered.
  • Insider ownership and the compensation metrics the board actually uses.

03Normalizing the earnings

A valuation built on a peak margin is a valuation of the peak. Normalization is the least glamorous and most important step: estimate the margin the business would earn at mid-cycle utilization, deduct maintenance capital, and treat growth capital separately. For cyclicals, use a multi-year average of through-cycle margins, not the trailing twelve months.

The adjustment list is short but decisive: add back non-recurring charges only if they are genuinely non-recurring, deduct stock compensation as the real expense it is, and normalize the tax rate rather than accepting a one-year benefit. Each adjustment is disclosed so that a reader can disagree with it.

04Reverse DCF

The reverse DCF is the core of the process. Instead of forecasting cash flows and discounting them to a value, hold the discount rate and terminal growth fixed, and solve for the free-cash-flow growth rate that makes the present value equal to today’s price. That growth rate is the market’s embedded expectation.

The table below shows the mechanic on a stylized company: $100 of current free cash flow, an 8% discount rate, and 2.5% terminal growth. The only question being asked is how much growth a given multiple already assumes.

Implied FCF growth (10 yrs)Present valueMultiple of FCFRead-across
2%$1,79217.9×Mature, low-growth business
4%$2,09320.9×Roughly nominal GDP plus a little
6%$2,44924.5×Durable share gains required
8%$2,86328.6×Reinvestment runway must be long
10%$3,35033.5×Exceptional, and usually priced as such
12%$3,91239.1×Requires a decade of near-flawless execution

Stylized and illustrative; not a valuation of any company. The point is the inversion of the question. At 24× free cash flow, the market is not paying for the past — it is paying for roughly 6% growth for ten years. The research task becomes: is 6% credible, and what evidence would show it is not?

The useful question is not what a company is worth. It is what the price requires, and whether that requirement is realistic.

05Catalysts and horizon

A cheap stock with no mechanism for repricing is a value trap with a spreadsheet. Every note states the specific events or disclosures that could close the gap: a segment disclosure, a capital allocation decision, a cycle inflection, a management change, a regulatory outcome. The horizon should match the catalyst. If the thesis needs three years and the catalyst is next quarter, one of them is wrong.

06Sizing and disconfirmation

Position size comes from the size of the gap between price and value, adjusted for how confident I am in the normalization, and capped so that a mistake in the model is survivable. A stock where the thesis depends on a single margin assumption gets a smaller position than one with several independent supports, even if the upside looks larger.

Disconfirmation is written into the note before the position is opened: the two or three observations that would cause an immediate review. This is the same principle as a stop-loss, applied to reasoning rather than price.

07The write-up template

Every company note follows the same structure, so that notes can be compared and reviewed quickly. The full notes run longer; this is the skeleton.

SectionContentLength
ThesisWhat the market believes, where I disagree, and why now200 w
Business mechanicsUnit economics, capital intensity, reinvestment runway500 w
Normalized earningsAdjustments, mid-cycle margin, maintenance capital400 w
Embedded expectationsReverse DCF, sensitivity, comparison to history500 w
Risks and disconfirmationSpecific evidence that would change the view300 w
Valuation and sizingScenario values, position size, review triggers300 w

08When I pass

Most companies fail before the valuation stage, and that is the desired outcome. I pass when the accounts cannot be understood without management’s narrative, when the returns depend on a single customer or contract, when the reinvestment opportunity is smaller than the cash the business generates and capital allocation is poor, or when the embedded expectation is so demanding that being approximately right still loses money.

Company notes written to this standard will be published in this section as they are completed. If you would like to see a full note before then, ask.

For discussion only. Not investment advice, and not a recommendation regarding any security. The valuation table is a stylized illustration and does not describe any company.

Back to work