Are streak features still spreading beyond Duolingo?
They spread, then plateaued. Streak screens rose from 6.8% of captured apps in 2023 to a 14.0% peak in 2024, then fell back to 8.1% in 2025 (59 of 730 companies) [1]. Streaks broke out beyond Duolingo in 2024 but did not keep climbing as a share. If you are considering a streak mechanic, it is a proven-but-not-rising pattern - adopt it for fit, not because it is trending up.
Streak screens peaked at 14.0% of captured apps in 2024, then settled to 8.1% in 2025 (July 2026).
The trend: a 2024 breakout that receded
Share of captured apps showing a streak (gamification) screen, by capture year:
| Capture year | Companies with pattern | Cohort size | Prevalence |
|---|---|---|---|
| 2023 | 20 | 293 | 6.8% [1] |
| 2024 | 56 | 399 | 14.0% [1] |
| 2025 | 59 | 730 | 8.1% [1] |
Streaks roughly doubled in share into 2024 as the mechanic spread beyond language-learning, then slipped back as a share in 2025 [1].
How to apply it
The 2024 spike-and-retreat suggests streaks are best treated as a fit-dependent mechanic, not a universal growth lever. They work where a daily-habit loop is genuine (learning, fitness, journaling); the share pullback in 2025 hints that apps without that loop tried streaks and did not sustain them. Adopt streaks because your retention model rewards daily returns, not because the pattern is climbing - it is not.
Caveats
Prevalence is deduped by distinct company and matched on %streak%; it is a lower bound bounded by tag recall [1]. Note the absolute company count barely moved 2024->2025 (56->59) while share fell, because the 2025 cohort was much larger (730 vs 399) - a reminder to read share, not counts [2]. Partial-2026 (11 apps) is excluded as a thin-capture artifact [2].
The numbers
| Stat | Computed from |
|---|---|
| 6.8% (2023, 20/293) -> 14.0% (2024, 56/399) -> 8.1% (2025, 59/730) | streak_prevalence_over_time |
| 2025 cohort 730 co vs 2024 399 co; 2026 partial 11 apps excluded; read share not raw counts | streak_prevalence_over_time / capture_cohort_by_year |
Sources & citations
- [1] Lazyweb Research analysis of 809 tracked apps (mobile-app corpus, 47,578 capture-dated screenshots), July 2026. Prevalence = share of companies captured each year whose screens carry a `%streak%` tag, deduped by company_name. ↩
- [2] Lazyweb Research analysis of 809 tracked apps (mobile-app corpus, 47,578 capture-dated screenshots), July 2026. Capture cohort sizes by year; trend read as share, not raw counts. ↩
Source: Lazyweb Research — proprietary analysis of real, in-market app screens. Cite as Lazyweb Research, 2026-07-07.