Last updated: 8 May 2026 · By Luke Lv, Director, Lumira Studio
Video analytics is the use of platform data to understand how video content is performing and what to do about it. The challenge is that platforms (YouTube, LinkedIn, Vimeo, Facebook, native website) all expose different metrics with different definitions and different reliability. Without a clear framework for what to look at and what to ignore, video analytics produces noise, not signal.
What video analytics actually measures
Most platforms expose a similar set of metrics, even if labels differ:
- Views. The least useful metric on its own. Definitions vary: YouTube counts views at 30 seconds, Facebook at 3 seconds. A “view” comparison across platforms is not like-for-like.
- Watch time / total seconds viewed. Aggregate seconds watched across all viewers. More signal than views.
- Average view duration. Mean watch time per view. Useful but masks distribution.
- Completion rate / percentage viewed. The most actionable retention signal.
- Audience retention curve. Where viewers drop off second-by-second. The single most useful analytic for improving content.
- Click-through rate (CTR). What proportion of impressions convert to views. Reflects thumbnail and title strength.
- Engagement signals. Likes, comments, shares, saves. Quality varies by platform.
- Traffic sources. Where viewers came from. Critical for distribution insight.
- Conversions / outcomes. Sign-ups, purchases, or other goals tied to the video.
The metrics that actually matter
For most B2B and corporate video work, the practical hierarchy:
| Tier | Metric | What it tells you |
|---|---|---|
| Tier 1 (Always look) | Audience retention curve | Where viewers drop off, what to fix |
| Tier 1 | Completion rate | How well the content holds attention |
| Tier 1 | Click-through rate | How well thumbnail/title earn the click |
| Tier 2 (Look weekly) | Watch time growth | Aggregate momentum |
| Tier 2 | Traffic sources | What is driving discovery |
| Tier 2 | Conversions / outcomes | Whether the video is doing its job |
| Tier 3 (Look monthly) | Audience demographics | Who is watching |
| Tier 3 | Engagement signals | Qualitative read on response |
| Tier 3 | Subscriber/follower growth | Long-term audience build |
Reading the audience retention curve
The retention curve is the most actionable analytic available. It shows what percentage of viewers are still watching at each second of the video. Three patterns to recognise:
Sharp drop in the first 15 seconds
Hook is weak. The opening did not earn the next 30 seconds. Fix: rewrite the open to lead with the most compelling element (problem, outcome, hook), not with introductions or context.
Steady decline
The content is fine but unengaging. Pacing too slow, dialogue too dense, visuals too static. Fix: tighten the edit, vary the pacing, add B-roll cuts where the picture is static.
Sudden drop mid-video
Something specific lost viewers at that timestamp. Often: a tangent, an over-long demo section, or an unclear transition. Fix: identify the specific moment and tighten or remove it.
Spike (rewatch)
A specific section is being rewatched. Useful information for tutorial content. Indicates the content is valuable enough that viewers want to revisit specific parts.
Platform-specific reliability
Not all video analytics are equal:
- YouTube Analytics. Most mature, most detailed, most reliable. The retention curve is uniquely available here.
- Vimeo Analytics. Good for embedded video on owned sites. Detailed engagement and CTA tracking.
- LinkedIn Analytics. Limited compared to YouTube. View counts are inflated by the 3-second threshold. Engagement signals are useful.
- Facebook / Instagram Insights. Similar limitations to LinkedIn. Use directional, not absolute.
- Native website embedded video. Depends on the player. Vimeo Pro, Wistia, and JW Player produce serious analytics. Self-hosted HTML5 video produces almost none without third-party tooling.
What video analytics will not tell you
Three blind spots to be aware of:
- Quality of the audience. 1,000 views from the right buyer persona is worth more than 100,000 views from off-ICP audiences. Demographics help but rarely tell the full story.
- Influence on long-cycle buying decisions. The video that influenced a £200k contract may not show up clearly in analytics. Sales-team conversations capture this signal.
- Brand impact. Long-term shifts in brand perception, search volume, and consideration are real but hard to attribute to specific videos.
Video analytics is necessary but not sufficient. The qualitative signal from sales conversations, customer feedback, and inbound enquiry quality is often more important than the numbers.
Common video analytics mistakes
- Comparing views across platforms. Different definitions, different thresholds, not like-for-like.
- Ignoring the retention curve. The most useful analytic, often skipped.
- Reading short-term data on long-cycle content. Brand-building video needs 90-day windows, not 7-day reports.
- Optimising for engagement at the cost of message. Engagement-bait reduces signal in the long run.
- Chasing the algorithm. Trying to game platform rules produces fragile results. Optimising for the actual viewer compounds.
Frequently asked questions
What is video analytics?
Video analytics is the use of platform data (YouTube, LinkedIn, Vimeo, Wistia, etc.) to measure how video content performs. Common metrics include views, watch time, completion rate, audience retention curves, click-through rate, traffic sources, and conversions.
What is the most important video analytics metric?
The audience retention curve, where available. It shows where viewers drop off second-by-second, which makes it the most actionable signal for improving content. Completion rate and click-through rate are next in priority.
How often should I review video analytics?
Tier 1 metrics (retention curve, completion rate, CTR) on every published video and weekly across the catalogue. Tier 2 (watch time growth, traffic sources, conversions) weekly. Tier 3 (demographics, engagement, subscriber growth) monthly. Daily review on long-cycle content produces noise rather than signal.
Are video views a good metric?
Not on their own. Views are the least useful metric in isolation because definitions vary by platform (YouTube at 30 seconds, Facebook at 3 seconds), audience quality varies, and views without context tell you nothing about whether the video did its job. Pair views with retention and conversion data.
What is a good completion rate for video?
Varies by length and format. Short-form (under 60 seconds): 50-70% is competitive. Mid-form (60s-3min): 30-50% is competitive. Long-form (3-10min): 20-35% is competitive. Beyond 10 minutes, 15-25% is the typical competitive range. Lower numbers do not necessarily mean failure if the right people watched the right parts.
How does Lumira Studio approach video analytics?
We focus on the audience retention curve as the primary improvement signal, completion rate as the engagement signal, and outcome metrics tied to the video’s specific job (qualified leads, deal velocity, conversions). Views are tracked but rarely treated as a primary success measure.
