Freek Borghgraef

The through-line

Across the posts, Freek Borghgraef keeps returning to one claim: engagement is not a feature you bolt on, it's an infrastructure you build around real user behavior. Streaks, badges, points, and leaderboards are treated as surface mechanics that fail unless they're wired to signals the product actually tracks and to the timing of when a user needs them. This holds steady from the earliest posts on Apple Activity Rings through to Khan Academy, Duolingo, and Kahoot [1][2][5][7].

The position sharpens and widens over time. Early posts analyze consumer health and learning apps to show how event-driven systems beat generic reminders [1][2][5][7]. Later posts pivot to fintech, arguing that the same mechanic, once it's visible and user-owned, becomes identity rather than loyalty, and identity is what actually resists competitive copying [10][11][13]. Most recently, the argument turns urgent: vibecoding is flooding the market with functional-but-unusable apps, churn is rising, and the value of understanding user behavior is going up, not down [9].

On event infrastructure

Borghgraef's recurring target is the team that "ship[s] engagement as a feature list: streaks, challenges, leaderboards, badge collections," launches it in a sprint, and watches the "drop-off curve" fail to move [2]. His counter-model is a system where "every workout logged, every goal set, every session completed generates a signal" that "feeds a system that decides what to surface next, and when" [2]. Apple's three rings are his standing example of this: "a single daily target hides too much... One number can't catch that. Three signals can" [1]. He extends the same logic to reward design at Khan Academy, where a "video view earns fewer points than an exercise" and a mastery challenge "earns the most," so the "design logic: reward the behavior that predicts long-term retention, not the behavior that's easiest to complete" [7][8].

On timing

A second, closely linked theme is that when an intervention fires matters as much as what it says. Khan Academy's Mastery Challenge "only appears when the system calculates that a previously learned skill is due for review," and Borghgraef calls this "load-bearing": "a challenge that's always available gets ignored" [5]. He applies the same standard to Apple's nudge copy, "you usually close your rings by now," noting it "reacts to the one ring that's actually falling short that day, not to a generic daily reminder on a fixed schedule" [1].

On streaks as identity

Borghgraef repeatedly reframes the streak mechanic as something owned by the user rather than manufactured by the app. On Duolingo: "That's not manipulation. That's the architecture of identity. The streak isn't Duolingo's streak. It belongs to the user" [6]. He generalizes the mechanic beyond language learning: "Every app category has the same mechanism available. A ride streak. A savings streak. A workout streak. A reading streak" [6]. Applied to a 150-year-old bank's badge-and-leaderboard program, the same logic holds: "That's not a loyalty program. That's identity. The user isn't responding to a rate or a cashback offer. They're protecting a score they built themselves" [11]. The buyable Streak Freeze on Duolingo fits this pattern too, since it is "buyable before the behavior breaks, not after" [4].

On competing on identity, not price

In his fintech posts Borghgraef argues that price and feature competition is a losing long game because it can be copied fast, while identity cannot. On Monzo: "Features get matched in a quarter. Identity takes a decade to copy" [10]. On Cash App's rotating Boosts: "Boosts aren't giving away revenue. They're paying for the only thing that matters in a high-churn category: a reason to open the app tomorrow" [13]. He frames the common mistake explicitly: "The mistake I see fintech teams make is reaching for 'more cashback' or 'higher APY'. Both are price wars" [13].

Takeaways

• Replace fixed-schedule reminders with event-driven signals tied to actual in-app behavior, the way Apple's ring nudge reacts to "the one ring that's actually falling short that day" [1].
• Time interventions to the moment of need, not availability: Khan Academy's Mastery Challenge only surfaces at "predicted skill decay," which is why it works while always-available prompts get ignored [5].
• Weight rewards by what predicts retention, not by what's easiest to complete, as Khan Academy does by giving mastery challenges the highest point value [7][8].
• Treat streaks as user-owned identity assets rather than engagement gimmicks, and give users a way to protect them, as with Duolingo's Streak Freeze [4][6].
• In competitive categories like fintech, build identity signals (a card, a badge, a visible streak) rather than compete on rate or cashback, since "features get matched in a quarter. Identity takes a decade to copy" [10][11][13].
• Expect app quality to decline and churn to rise as vibecoding lowers the barrier to shipping functional-but-unusable apps, making user-relevance work more valuable, not less [9].