In proven, crowded markets, the wedges that worked were aimed not at the incumbent's feature list but at its business model:
- Plausible vs Google Analytics — attacked surveillance-in-exchange-for-free (the model), not the feature set;
- Tally vs Typeform — attacked per-response billing that punishes the user's success;
- TypingMind vs ChatGPT Plus — attacked flat-rate pricing vs pay-per-use;
- the Anki-export route vs LingQ/Migaku — attacks retention-by-captivity (downgrade data loss, blocked export): free export is a wedge the incumbents cannot match without giving up the lock-in they live on (the-anki-parasite).
Why it holds. Features can be copied by the incumbent in a quarter; the business model cannot be matched without self-harm — Google cannot stop monetizing data; Typeform cannot stop charging per response without destroying its revenue; OpenAI would not un-bundle its subscription. A wedge lodged in the incumbent's revenue structure is a wedge it is institutionally forbidden to close.
Open extension test. flomo: does "we refuse hoarding features" attack the incumbents' growth-by-accumulation model in the same structural way? If strained, the pattern may be limited to pricing structures specifically.
Cousin: parasitic-growth mechanism 6 — a host whose model forbids it from closing your gap is the same structure found inside ecosystems.