Thin Edge of the Wedge

Empires that launched embarrassingly narrow. The point: the wedge is supposed to look unambitious. That's why incumbents ignore it and you get to win it. 2026-05-31.

The pattern: none of these started as the thing they became. Each picked one tiny, almost-laughable slice — and the ambition was the empire they built from it, never the entry point. If the wedge looks big and exciting on day one, it's probably already contested.

The wedge → the empire

Company	Started as (the laughably narrow wedge)	Became
Amazon	An online bookstore. Bezos picked books off a list of 20 candidates — pure commodity, infinite catalog, easy to ship.	"The everything store." Books were always a Trojan horse for retail + AWS.
Uber	Hail a black town car from your phone — in one city (SF), premium-only. "UberCab."	UberX, global rides, Eats, freight. The luxury niche seeded the network.
DoorDash	PaloAltoDelivery.com — a landing page with PDF menus from a handful of Palo Alto restaurants and the founders' cell number.	National logistics layer for local commerce.
Airbnb	Air mattresses on their apartment floor + breakfast, for designers when SF hotels sold out during a conference.	The global lodging marketplace.
Stripe	"7 lines of code" — payments for the one developer who hated every existing integration.	Internet's financial infrastructure (billing, treasury, lending).
Facebook	TheFacebook — Harvard students only, then one campus at a time.	3B+ users. The deliberate scarcity is what made it spread.
Shopify	Software to run one snowboard store (Snowdevil). They couldn't find good tooling, built their own, sold that.	~10% of US e-commerce.
Square	One little white card reader in the headphone jack, for the micro-merchant who couldn't take cards (a glassblower).	Full SMB commerce + banking stack.
Toast	A better POS for restaurants with fewer than two locations.	Payroll, payments, lending — the restaurant OS.
AppFolio	Cloud PM software for small residential managers, when Yardi & RealPage only cared about enterprise. Your exact vertical.	Public company (IPO 2015), now expanding into AI + payments + screening.

Sources: Toast / land-and-expand, wedge strategy, company histories (Amazon, Uber, DoorDash, Airbnb, Stripe, Facebook, Shopify, Square, AppFolio).

The three things every one of these did

Picked a wedge that looked too small to matter. "Books." "Air mattresses." "7 lines of code." The unsexiness was the moat — big players didn't bother defending it.
Earned the right to expand — customers pulled them. Nobody started as a platform. The empire was added after they dominated the slice, and it was existing customers asking for the next thing.
The ambition lived in the destination, not the door. Bezos always meant "everything." He just knew you don't open with everything. Vision ≠ roadmap.

What this means for Bedrock

Maintenance isn't your ceiling — it's your door. "AI agent that runs maintenance over iMessage for SMB residential PMs" is a perfectly respectable wedge — arguably narrower and sharper than "books." The God-Mode company brain is your Amazon-everything. You don't abandon it; you earn it by owning the door first.

"Crowded" is the wrong read. The funded players (EliseAI, Super, Beam) are charging at enterprise multifamily. Your slice — SMB residential, iMessage-native, agent actually closes the loop — is the part of the market they're not defending. Crowded at the top ≠ crowded at your wedge. Empty markets are the real death sentence; a crowded one means demand is proven.

The open question isn't "maintenance vs. ambitious." It's "is maintenance a painkiller for the right customer?" A vitamin for Vanessa can be a painkiller for the PM who just lost a tenant over a missed repair. You don't know yet — because n=1 warm customer. resolve this first