C.W.K.
Stream
Lesson 03 of 04 · published

No Urgency Mechanics

~11 min · non-goals, anti-patterns, calm-changes, engagement

Level 0Open Gate
0 XP0/36 lessons0/12 achievements
0/100 XP to next level100 XP to go0% complete
"Every 'act now' feature is a bet that your reaction is worth more than your judgment. Keep takes the other side."

The non-goals are load-bearing

Most architecture docs list features. Keep's most important list is the opposite — an explicit set of things it will never build. Among them: no unsolicited alerts, no buy/sell scores, no emotional urgency mechanics, no order entry, no automated trading. These aren't "maybe later" items. They're permanent boundaries, and naming them out loud is what keeps a well-meaning future contributor from cheerfully adding a "⚠️ your portfolio dropped 3% today!" banner because it seemed helpful.

Why urgency mechanics are the specific enemy

An "urgency mechanic" is any design element whose purpose is to make you feel that you must act now: streaks, red badges, countdown timers, "X people are selling," pulsing alerts, daily-mover leaderboards. They're borrowed straight from slot machines and social feeds, and they work — they reliably increase engagement. That's exactly the problem. In a portfolio, increased engagement means increased trading, and increased trading is, for a long-term holder, almost always value destroyed. The mechanic that's great for a game's retention is poison for an investor's returns.

Engagement is the wrong metric for a stronghold. Optimizing a financial tool for time-on-app optimizes for the exact behavior — frequent reactive trading — that hurts the user. When your success metric and your user's success point in opposite directions, you must refuse the metric. Keep measures itself by accuracy and calm, never by opens.

The calm 'changes' projection

Keep still lets you see what moved — it just does it without the adrenaline. There's a projection called changes that compares the present against the most recent prior snapshot and reports the difference. Read that sentence carefully: it reports, it never raises. There's no threshold that trips an alert, no "significant move" siren. It's a quiet diff you can look at when you choose to, presented as a fact — "here is what's different since last time" — with no implied instruction to do anything about it.

The seductive feature to reject: the 'smart alert.' Someone will always propose "just notify Dad when something important happens — surely that's helpful?" But 'important' is a judgment the app is forbidden to make, and a notification is unsolicited by definition. The helpful-sounding alert is the exact thing the whole ethic exists to prevent. Reject it every time.
This is why Keep will never 'grow into' a trading app. The pressure to add alerts, scores, and one-tap actions is enormous — it's what every competitor does, it's what drives engagement metrics, and it always sounds like a feature request. Naming these as permanent non-goals is the guardrail that survives the pressure, long after the person who wrote the ethic has moved on.

Code

The changes projection reports; it never raises·python
def changes_since_last_snapshot(owner, as_of):
    now = current_snapshot(owner, as_of)
    prev = most_recent_prior_snapshot(owner, before=as_of)
    # A calm diff. Note what is ABSENT: no threshold, no alert,
    # no 'if delta > X: notify()'. It states the difference and stops.
    return {
        "prev_date": prev.date,
        "delta_value": now.value - prev.value,   # a fact, not a siren
        "delta_by_position": diff_positions(now, prev),
    }

# There is deliberately no alerting layer above this.
# 'Important' is a judgment the automatic surface must never make.

External links

Exercise

List three urgency mechanics you've seen in any app (streaks, red badges, countdowns, 'others are doing X', movers lists). For each, name the behavior it's engineered to increase, and ask: is that behavior good for the user, or good for the app's engagement metric? Then design the calm alternative — the version that reports the same underlying fact without manufacturing the urge to act.
Hint
The clean test for an urgency mechanic: remove it, and does the user lose any INFORMATION, or only lose the PRESSURE? If they still have every fact and only the adrenaline is gone, it was never informing — it was nudging, and a stronghold cuts it.

Progress

Progress is local-only — sign in to sync across devices.
Spotted a bug or have feedback on this page?Report an Issue

Comments 0

🔔 Reply notifications (sign in)
Sign inPlease sign in to comment.

No comments yet — be the first.