Building for a market with an expiration date
When the market itself is a depreciating asset, speed becomes the strategy — and pre-written gates matter more, not less, because deadline pressure is exactly when discipline gets rationalized away.
- In February 2026 the Supreme Court ruled IEEPA tariffs unlawful (Learning Resources, Inc. v. Trump), leaving roughly $127 billion potentially refundable across ~330,000 importers.
- Refund rights expire continuously on a fixed legal schedule: the addressable market shrinks an estimated 8–10% every month.
- The strategy priced the decay in from day one, and the app went from empty repository to working scaffold in a single ~40-commit day.
- The discipline lesson: a launch gate written in advance — no broker promise without a signed partner — held even under deadline pressure.
Most products are built for markets you assume will still be there next year. In April I started building for one with a countdown clock bolted to it.
The window
In February 2026, in Learning Resources, Inc. v. Trump, the Supreme Court ruled that a class of tariffs imposed under the IEEPA — the International Emergency Economic Powers Act, the emergency-powers statute the tariffs leaned on — had no legal authority behind them. The practical consequence: roughly $127 billion in duties, paid by some 330,000 importers, became potentially refundable. For a single small importer that can mean six or seven figures back — if they navigate the customs process to claim it in time.
"In time" is the whole game. Refund rights expire entry by entry on a fixed legal schedule, so the total addressable market shrinks on one too — an estimated 8 to 10 percent of it gone every month. Most markets reward patience. This one taxes it.
const CAPE_DEADLINE = "2026-07-19T23:59:59"the clockThe market's depreciation schedule, rendered as product UI — a live days/hours/minutes counter at the top of every page. When the addressable market shrinks 8–10% a month, the deadline isn't marketing; it's the product's central fact.
What a countdown does to product decisions
Speed stops being a virtue and becomes the strategy. The app went from empty repository to working scaffold — authentication, privacy-scrubbed analytics, the tariff rate tables — in a single forty-commit day. Not because rushing is good, but because every week of build time was, literally, another percent or two of the market evaporating.
The thesis has to be testable just as fast. The original idea was self-serve software: pay $79, get your refund paperwork formatted. The evidence said otherwise — government data showed more than 12,000 refunds already rejected over a single missed enrollment step. The real gap wasn't access to software; it was operational competence in a process an importer runs once in a lifetime. So the product pivoted: free analysis that shows an importer what they're owed, then a hand-off to licensed professionals who actually file.
And the gates matter more under deadline pressure, not less — because deadline pressure is exactly when "just ship it" starts corroding standards. So the one that mattered was written down before it could be tested against temptation: no broker-match promise on the site without a signed partner behind it, because a call to action that points nowhere destroys trust with the people deciding whether to hand you their customs data. That gate held, even when honoring it cost me.
Every market has a decay rate
Most are just slow enough to ignore. Building inside an explicit one forces an honesty that permanent-feeling markets let you defer: what a month actually costs, which assumption would break the thesis, and what you won't do even at speed. The countdown didn't change what good discipline looks like. It changed how fast the bill for skipping it would arrive.