Why every card measures

Most apps measure at the resolution of the dashboard: you opened it, you spent eleven minutes, you came back Tuesday. That's coarse telemetry — useful for counting users, useless for knowing one. Swapp measures at the resolution of the interaction. Every round of a game emits a start and a completion with a score. Every page you turn in a book emits progress. Every answer in a quiz emits the answer. The unit of record is the thing you actually did.

Each of those signals names what it measures — its construct. A reaction game doesn't just report a score; it reports reaction time. A memory grid reports working memory. The number means something specific, which is what separates a measurement from a leaderboard.

Constructs make difficulty honest too. A card's difficulty isn't a designer's guess or a slider someone set at launch — it's computed from how everyone actually performed on it. If a card turns out to be harder than it looks, the population's results say so, and the feed treats it accordingly.

That's also why progress on Swapp is a fact rather than a feeling. You didn't "seem sharper this week" — your reaction times moved, on these cards, by this much, against how everyone else did. Rank and progress are computed numbers with a paper trail, and the feed personalizes from that same history: what you see next comes from what you've done.

Under all of it sits Empress, an immutable behavioral ledger. Every interaction lands there as an xAPI statement and stays there. The measurements aren't a feature bolted onto a game feed — the feed is a surface on top of the ledger. That ordering is the whole product.