Trends · 8 min read
The earliest signal is rarely the loudest
How emerging technologies, startups, hiring patterns, and online conversations reveal market shifts years before they become obvious.
Every mainstream trend leaves a trail for years before it breaks. The trail is rarely hidden — it is simply spread thin across places nobody reads together: preprints, niche forums, procurement filings, job postings, import records.
The three phases of a signal
- Fringe. A handful of practitioners solve a problem for themselves.
- Transfer. The solution jumps domains, usually via a single translator.
- Consensus. The press names it, and the opportunity is mostly priced in.
Most organizations only have instrumentation for phase three.
A trend is not a prediction. It is a distribution of evidence that has started to lean in one direction.
Building instrumentation
The practical work is unglamorous: pick ten sources per domain, normalize them, and track rate of change rather than volume. Volume tells you what is popular. Acceleration tells you what is arriving.
- Track vocabulary, not topics — new words precede new markets.
- Weight sources by how early they were last time.
- Attach evidence to every claim, or the claim will not survive a boardroom.
What to do with lead time
Nine months of lead time is only valuable if something changes because of it. Convert each signal into one of three moves: a hiring decision, a product bet, or a deliberate pass. Anything else is trivia with a chart attached.