The House Method, in brief
Every company we publish on is run through both legs of the House Method — always
both, no exemptions. Leg A asks how cheaply the market prices a company's quality
growth. Leg B asks what the company's cash flows are actually worth. A name earns
high conviction only when both legs agree.
Leg A — Compounder Score (relative value)
Score = EV ÷ [ Normalized Fwd Gross Profit × Quality Growth × Durability ]
Lower means cheaper for quality growth. Every input is printed next to every published score:
- Normalized gross profit — next-twelve-months revenue × a normalized gross margin: the average of the consensus forward margin and the 5-year mean margin, never above the forward margin. A peak-cycle margin is halved instead of capitalised, while a structural shift that shows up in both history and consensus still counts.
- Quality Growth — the 3-year forward revenue CAGR, capped at 40% and floored at 0%, plus half the trailing free-cash-flow margin — growth weighted by the cash it produces.
- Durability (0.6–1.3) — a fixed rubric, never a freehand number. Base 0.8. Add +0.2 for market leadership (top-2 share in the core market, stable or rising — only with a sourced share figure behind it), +0.2 for revenue visibility (contracted backlog, royalties or subscriptions at least 50% of forward revenue), +0.1 for a strong balance sheet (net cash, or net debt/EBITDA under 1), +0.1 if gross margin held within 15 points through the last downturn. Subtract −0.2 if a single product or customer exceeds 30% of revenue, −0.1 for negative trailing free cash flow, −0.1 if the value hinges on one pending decision.
Tiers: T1 Deep Value (0–8) · T2 Quality Discount (8–16) · T3 Fair Compounder (16–30) · T4 Stalwart Fair / Expensive (30–50) · T5 Bubble / Not Applicable (50+).
Hard gates: negative EBITDA and negative free cash flow caps a name at Tier 4; no gross profit (pre-revenue) or Quality Growth at or below zero makes Leg A not applicable. If a reader cannot recompute our number from the published inputs, we do not publish it.
Leg B — Margin of safety (absolute value)
MoS% = [(Average Intrinsic Value − Price) ÷ Price] × 100
Two parallel two-stage discounted-cash-flow models — one on operating cash flow,
one on free cash flow, each with a five-year consensus growth stage and 3% terminal
growth — are averaged into a single intrinsic value per share, then
compared with the market price:
- +40% or more — strong undervaluation
- +20% to +40% — meaningful undervaluation
- +5% to +20% — modest undervaluation; selective, quality businesses only
- 0% to −5% — fairly valued
- Below −5% — overvalued on this method
If a company produces no cash flow, the model honestly returns roughly zero — which means the method cannot value it, not that it is worthless.
How the legs combine: we publish both numbers and a quadrant —
never a blended composite score. Tier 1–3 with a margin of safety above +5% is
BEST (high conviction). Tier 1–3 at or below +5% is WATCH
(quality, wrong price). Tier 4–5 above +5% is INVESTIGATE (cheap, but
the engine is in question). Tier 4–5 at or below +5% is AVOID.
We publish the disagreements, not just the agreements.
Lineage, stated plainly: the relative leg is an EV/Gross-Profit
method in the PEG tradition, rebuilt with mid-cycle normalization, a quality-growth
term, and a durability rubric. The absolute leg is a standard two-stage DCF. The
method is ours in its rules, normalization, and disclosure — not in pretending the
underlying math was invented here.
