Lore

DHM Model (Delight, Hard-to-Copy, Margin-Enhancing)

Overview

DHM is Gibson Biddle's core product-strategy hypothesis framework: a product or feature bet should Delight customers in ways that are Hard-to-copy and Margin-enhancing. Biddle frames strategy itself as "a route to continuing power in a significant market" (Hamilton Helmer), and DHM turns that route into a testable hypothesis rather than a fixed plan ("will this delight in hard to copy margin enhancing ways").

Durable advantage compounds even from a low hit rate

Many of Netflix's delight bets — especially social/network-effect features — reportedly succeeded only about a third of the time. Biddle argues DHM still produced a world-class company because the hard-to-copy leg compounds: content-spend advantage sustained over years builds economies of scale a new entrant can't match ("if I gave you 500 million bucks to start a startup to compete with Netflix, what would you find hard about that?"). The batting average on individual bets matters less than whether each win is structurally hard to replicate.

Illustration: Netflix Party (a demanded feature Netflix declined to build)

A "what would you do" case Biddle uses to pressure-test DHM. A fast-growing third-party extension let friends watch Netflix in sync; a live poll initially favored building it, 75%–25%. Netflix passed, reasoning through all three DHM legs: the room of tech-savvy panelists were "not our customers," prior first-party social attempts (Netflix Friends, Xbox party mode, international sharing) had each plateaued around 2–5% adoption, it cut against Netflix's "movie enjoyment made easy" positioning ("your friends have really sucky movie tastes"), and — decisively — it was easy to copy by definition, since rivals already had similar features. Once this evidence was presented, live sentiment shifted from ~75% in favor to 42%.

Illustration: auto-canceling inactive accounts (forgoing $100M)

Netflix auto-canceled the ~0.5% of members who hadn't watched anything in 12 months, giving up a projected $100M in revenue. No A/B test can price the brand trust gained, or the downside of an AOL-style cancellation backlash avoided, so Biddle treats this as a case where Delight and Hard-to-copy must be judged by human reasoning where consumer-science can't reach. Against $25B in revenue, $100M is framed as a low-stakes price for that hard-to-copy trust advantage.

Illustration: hard-to-copy advantage lost

Biddle cites his own product, Creative Wonders/The Learning Company (sold to Mattel for $4B, spun out a year later at $400M), as a case where commercial success without hard-to-copy advantage evaporated almost immediately.

Related

glee-model sequences DHM bets into a multi-year vision; gem-model force-ranks growth/engagement/monetization when they compete for the same bet; smt-model turns a single DHM hypothesis into a measurable proxy metric and tactics.

Tracking Bets Over Time

Tracking Bets Over Time

Netflix keeps a running green/red chart of its major DHM bets and their outcomes across its history. The chart shows the company being "right" only roughly a third of the time — used to argue that a strong long-term record doesn't require a high hit rate, only that wins compound and losses get killed via Fail Fast, Recover Fast.