YouTube Shorts Algorithm: Signals, Loops, and the Swipe Metric

The YouTube Shorts algorithm is designed to solve a unique user experience problem: how to hold the attention of a mobile viewer equipped with an infinitely scrollable feed. Unlike horizontal formats that require conscious selection, Shorts are delivered via predictive auto-play. Because the effort required to skip a video is reduced to a single upward thumb flick, the algorithmic margin for error is razor-thin.

To win distribution in the Shorts feed, creators must understand the mechanical signals that govern the platform's vertical recommendation architecture.

The Core Gatekeeper: "Viewed vs. Swiped Away"

The most consequential metric inside YouTube Shorts analytics is the ratio of Viewed vs. Swiped Away. This metric measures the percentage of viewers who actively watched the video for at least two seconds versus those who immediately swiped to the next clip.

Algorithmic performance thresholds typically fall into three brackets:

• Under 60% Viewed: Algorithmic distribution halts almost instantly. The video is flagged as lacking immediate hook value.
• 60% to 75% Viewed: Standard performance; yields modest distribution across niche testing pools.
• 80%+ Viewed: High distribution signal. The algorithm will push the video outward to broader audiences.

The 100%+ Retention Metric (Loop Velocity)

In traditional long-form video, an audience retention rate of 50% represents strong performance. In Shorts, 50% retention often leads to stalled distribution. Because vertical videos loop automatically, top-performing Shorts frequently achieve average percentage viewed figures of 105% to 130%.

Achieving these retention rates requires deliberate script architecture:

1. Seamless Looping: Structure your final sentence to flow directly into your opening line. Connecting the ending phrase to the opening hook keeps the viewer watching through a second cycle before they realize the video has restarted.
2. High Information Density: Eliminate pauses, breath breaks, and transitional hesitations. Every second of video must advance the narrative, present a visual change, or deliver information.
3. Visual Layering: Display dynamic background details, on-screen text overlays, or fast-moving visual elements that encourage viewers to replay the clip to catch details they missed the first time.

The Testing Pool Cycle: How Shorts Scale

When you publish a Short, it follows a standardized distribution loop:

Phase 1: The Initial Test Batch

Within minutes or hours of upload, the algorithm serves the video to an initial evaluation group of 1,000 to 3,000 users. During this burst, the system closely monitors Viewed vs. Swiped Away rates, completion rates, likes, and shares.

Phase 2: The Evaluation Pause

The sudden view surge often plateaus abruptly. During this pause, the algorithm processes the comparative metrics against category benchmarks. If your completion rate drops below baseline or your swipe-away rate climbs, distribution stops.

Phase 3: Broad Tier Scaling

If your initial test signals register in top percentiles, the system pushes the Short to broader, general audiences. This process repeats across progressively larger cohorts until performance signals fall below system thresholds.

Conclusion

The Shorts algorithm is a mathematical reflection of viewer attention. Master the initial two-second visual hook to suppress swipe-aways, engineer clean audio loops to drive retention above 100%, and strip away narrative fluff to build vertical videos that scale consistently across the platform.