Platform Benchmarks That Actually Matter for Video Performance

Generic benchmarks are seductive because they're easy. "Average CTR for video ads is 0.9%. Average watch time for Reels is 11 seconds." These numbers feel like answers. They're usually distractions.

Why benchmarks lie

The average includes everyone. The DTC brand with a $2M monthly spend and the local business testing ads for the first time. The accounts with 50 followers and the accounts with 5 million. Lump them together and you get a number that applies to exactly nobody.

The benchmarks that matter are specific to your category, your audience size, your spend level, and — most importantly — your own historical performance. If your CTR last month was 0.6% and this month it's 0.8%, that's meaningful even if "the benchmark" says 0.9%. You're not competing against an industry average. You're competing against your own baseline.

What to actually track, by platform

Instagram / Facebook (Meta)

The only metric Meta's algorithm meaningfully optimizes for depends on your campaign objective. If you're running conversion campaigns, it's cost per conversion — not CTR, not CPM, not engagement rate. Everything else is a leading indicator at best.

That said, for creative testing specifically, two numbers are worth watching:

Hook rate (3-second video plays / impressions): tells you if the creative gets past the scroll. Below 25% on cold audiences, the hook needs work.
Hold rate (15-second video plays / 3-second video plays): tells you if the creative holds attention. Below 50%, the body of the ad isn't delivering on the hook's promise.

TikTok

TikTok's algorithm cares about watch time and completion rate more than anything else. Engagement (likes, comments, shares) signals quality, but if people don't watch, the algorithm won't distribute.

Average watch percentage is the north star. Above 50% for videos under 30 seconds is solid. Above 35% for longer videos.
Shares matter more than likes for distribution. A share signals "this was worth sending to someone." TikTok weights it accordingly.

YouTube

YouTube is the only platform where total watch time genuinely determines success, because the algorithm optimizes for session duration.

CTR (click-through rate from impression) matters for discovery but is inflated by suggested/recommended traffic. Your CTR from browse features is the number that reflects thumbnails and titles.
Average view duration, not percentage. A 20-minute video with 40% retention outperforms a 2-minute video with 80% retention in YouTube's algorithm. Duration is the currency.

LinkedIn video is still under-invested relative to the other platforms, which means good creative can disproportionately win. The metrics are less standardized, but:

Completion rate matters because LinkedIn auto-plays in-feed and counts a view at 2 seconds. A high completion rate signals relevance.
Comments carry more weight than reactions. LinkedIn's feed algorithm rewards discussion.

Set your own baseline

Take your last 10 ads or organic videos. Calculate the median, not the mean, for each metric you care about. That's your baseline. Next month's goal isn't "beat the industry average." It's "beat last month's median by 10%."

The median is important because the mean gets pulled around by outliers — one video that went weirdly viral, one that bombed for no clear reason. The median tells you what your typical content does. That's what you're trying to improve.

When benchmarks are actually useful

They're useful for diagnosis, not for goal-setting. If your 3-second view rate is 8% and the category benchmark is 30%, something is broken — not just below average. That kind of gap is worth investigating.

But if your rate is 25% and the benchmark is 30%, you might be perfectly fine. Maybe your audience is smaller and better-targeted, so you get fewer views but better conversion. Maybe your creative is serving a different objective than most ads in the benchmark set. Context eats benchmarks for breakfast.