Breadcrumbs

What is EBVS and how do I reduce a high EBVS rate?

EBVS (Exits Before Video Start) is the QoS metric in AV Insights that counts the share of playback attempts where the user left before the video actually started playing.

What EBVS measures

EBVS represents AV sessions where the user triggered play but the content never actually started — they exited before the first frame rendered. The canonical signal is the gap between two distinct events:

  • av.play — fires when the user clicks the play button.

  • av.start — fires when the first frame of content actually appears on screen.

If your av.play count is meaningfully higher than your av.start count, the difference is users who clicked play but left during the loading / pre-roll / autoplay-block / startup-buffer window before content rendered. EBVS quantifies that gap.

Under the hood, sessions are flagged as "bounced" (av_session_bounce = True) when the sum of av_cursor_difference across the session is 0 — i.e., the cursor never moved through the content. These bounced sessions are what EBVS reflects.

It's the closest AV Insights equivalent to "user clicked play and got nothing," and it's the first metric to look at when investigating poor video quality of experience.

Why EBVS is high

Common causes, in rough order of frequency:

  • Slow initial buffering. The video player took too long to load the first segment, and the user gave up. The fix is on the delivery side — CDN configuration, manifest loading, ad-server response time.

  • Pre-roll ad timeouts. A pre-roll ad failed to load or took too long to start, so the session ended before the main content began.

  • Autoplay blocked. Browsers (especially mobile Safari and Chrome) block autoplay with sound. The session starts but the player is paused waiting for user interaction; if the user scrolls away, EBVS counts the session.

  • Player misconfigurations. A player that fires av.play on page load but doesn't actually start streaming until a user click will count an EBVS session for every visitor who scrolls past without clicking.

  • Quality switching at startup. Adaptive bitrate players that spend the first few seconds switching quality levels can produce enough delay to lose impatient viewers.

How to reduce EBVS

  1. Confirm the EBVS rate is genuinely high. Industry baselines vary by content type — short clips often have higher EBVS than long-form because the click-through population is more exploratory. Compare against your own historical baseline before declaring a problem.

  2. Break down EBVS by content type, player, and device. A single problematic content type or player version usually explains most of the elevated rate.

  3. Inspect the time-to-first-frame on the delivery side. Use your CDN's logs to confirm whether the initial segment is loading within the user's tolerance window (~2-3 seconds for most audiences).

  4. Review autoplay and ad-loading behavior. If pre-rolls are involved, measure the pre-roll fill time and timeout settings separately.

  5. Use av_position to confirm true zero-consumption. Sessions where av_position is 0 at session end are unambiguous EBVS; sessions where the position moved slightly may indicate a different issue (early abandon, not failure to start).

Distinguishing EBVS from genuine user abandonment

EBVS is a technical metric — it represents sessions where the player likely failed to deliver, not sessions where the user deliberately left. If you see EBVS rising alongside a fall in content quality scores or a rise in av.error events, the cause is delivery-side. If EBVS rises alongside changes in your content mix or landing pages, the cause may be user behavior rather than technical failure.