Rebuffer ratio = (total time spent rebuffering) / (total session time). A session that played 60 minutes with 30 seconds of rebuffer has a ratio of 0.83 %. Industry benchmarks: under 0.5 % is excellent, 0.5–1 % is good, 1–2 % is acceptable, above 2 % causes measurable drop in engagement. Akamai, Conviva and Mux all publish industry-wide averages each year — typical 2025 values across major OTT services hover around 0.3–0.8 % for desktop, 0.5–1.5 % for mobile.
The metric matters because rebuffering directly damages user behaviour. Studies from Conviva (consistently across years) show that for every 1 % increase in rebuffer ratio, the average viewer watches roughly 5–10 minutes less per session and is more likely to abandon. A live event with rebuffer ratio above 3 % can lose 20 %+ of its audience within the first 10 minutes. Every QoE improvement project at every OTT eventually targets rebuffer ratio.
Reducing rebuffer ratio means improving the weakest link in the chain. Common interventions: lower starting rendition (less risk of buffer drain at startup), more aggressive downshift in the ABR (avoid the wait of fetching a too-high segment), larger player buffer (more cushion against network dips), better CDN selection (multi-CDN steering to faster paths), and better encoding (lower bitrate at same quality means less stress on the network). Each lever typically gives 5–20 % rebuffer-ratio reduction in the right context.
