Width. It takes two inputs: the size of the spread (ROIC minus WACC, through the cycle) and its expected duration. Morningstar's operational grammar is worth adopting because it forces probabilistic language. A narrow moat means excess returns are more likely than not to persist 10 years out. A wide moat means more likely than not 20 years out, with 10 nearly certain. No moat is anything less.
Some rules of thumb help with calibration, though none of them is a law. Multiple reinforcing sources beat one (Visa: network + scale + brand). Customer-level lock-in (switching costs, efficient scale) tends to outlast product-level advantages like patents with dates and process leads that diffuse. Regulated tolls cap width whatever the barrier. And the wider your claimed CAP, the more the valuation depends on it. So the next module makes CAP explicit, and Phase 8's teardown demands "moat (source + strength)" in one line, never just "yes."
Advertisement
How do you read the direction of a moat?
Direction: the derivative matters more than the level. Prices already capitalize the known width. Your edge usually comes from spotting width changing before the fade (or the widening) is consensus. Read direction from leading indicators. You already own most of them from earlier work:
Widening signals. Price increases stick with volumes intact, so FMCG value and volume growth are both positive. Retention or NRR trends up. Incremental ROIC is at or above average ROIC, meaning the new rupee earns more than the old, as when DMart adds stores at cluster economics. Take rate rises with share stable. Competitors exit, or refuse to match capex. And the flywheel spins, with scale metrics improving unit economics.
Eroding signals. Price hikes fail or discounts creep. Watch realizations and trade-spend lines for this. Gross margin slips while share is defended, which is paying to stand still, the Voltas signature. S&M or A&P per rupee of revenue rises just to hold position. DSO creeps up as customers gain power, which is the Phase 2 lens re-aimed. Incremental ROIC drops below average ROIC. Asian Paints' newest capital earns less than its old capital, so the erosion shows up at the margin first. A capitalized, irrational entrant arrives, like Jio in 2016 telecom or Birla Opus in 2024 paints. ("A moat is only as strong as the rationality and poverty of your richest potential entrant.") Regulation re-prices the toll (UPI vs card MDR, IEX and market coupling). And a technology shift resets the basis of competition.
That's Nokia's lesson: platform changes don't attack the moat, they drain the lake around it.
What do a moat's fade and recovery look like in telecom?
Telecom as the full movie (illustrative). Bharti Airtel entered FY16 with scale, the best network, and ~8% ROCE. Jio's late-2016 entry came with free voice, near-free data, and ~$30bn+ pre-committed. It collapsed industry ARPU from ~₹180 to 200 toward ~₹100, drove a dozen operators into three, and pushed Bharti's returns to ~zero by FY19. Scale and share were no defense against an entrant with different objectives and a deeper pocket. Then the surviving three raised tariffs repeatedly (2019, 2021, 2024). Bharti's ARPU climbed toward ~₹245+ and its ROCE got back into the low-to-mid teens by ~FY25.
Both halves teach something. The destruction shows that "network + scale" in a commodity-service industry is narrow protection. The recovery shows the capital cycle, which you'll study with moat dynamics: returns are made by industry structure, and structure is made by entry and exit. Direction analysis would have caught both turns years before the ROCE row did. The entrant's spectrum purchases were public across 2013 to 15, and the tariff-hike coordination was visible in 2019.
How often should you re-check a moat?
The standing routine. For any company you cover, re-run the five-row numbers test annually. Re-read the direction signals quarterly, since results calls make most of them visible. Then write the moat verdict down in your knowledge system: source, width, direction, and the one thing that would change your mind. Your capstone dossier and every Phase 8 teardown will demand this artifact.