The complete guide to engagement rate
Engagement rate is the single number that tells you whether a creator's audience actually pays attention. A 500K-follower account with 0.3% engagement engages fewer real humans than a 50K account with 4% engagement. Here's how to calculate it, what's considered good on each platform, and where the number lies to you.
The formula depends on the platform
On Instagram, engagement rate divides by followers: (likes + comments + saves + shares) ÷ followers × 100. A creator with 100,000 followers averaging 4,000 likes and 200 comments per post has a 4.2% engagement rate.
On TikTok and YouTube, the convention is different. Because video distribution is decoupled from followers (the For You Page on TikTok and YouTube's recommendation engine push videos to non-followers), the standard formula uses views: (likes + comments + saves + shares) ÷ views × 100. A TikTok with 500,000 views and 25,000 likes plus 2,000 comments has a view-based engagement rate of 5.4%.
This means a creator's Instagram engagement rate and TikTok engagement rate are not directly comparable. A 4% Instagram engagement rate is healthy. A 4% TikTok engagement rate is below average. Always check which denominator a quoted rate uses.
Instagram benchmarks (followers-based)
Engagement rates fall as follower counts grow because larger accounts accumulate passive followers, giveaway acquisitions, and lurkers. The benchmarks for Instagram in 2026:
- Nano (under 10K): 4 to 8% is healthy, below 2% is concerning
- Micro (10K to 100K): 3 to 6% is healthy, below 1.5% is concerning
- Mid (100K to 500K): 2 to 4% is healthy, below 1% is concerning
- Macro (500K to 1M): 1.5 to 3% is healthy, below 0.8% is concerning
- Mega (1M+): 1 to 2.5% is healthy, below 0.5% indicates an inflated account
If a 500K-follower creator pitches a 5% Instagram engagement rate, either they're a genuine outlier (rare) or something is being manufactured.
TikTok benchmarks (views-based)
TikTok engagement rates are structurally different because the denominator is views, not followers. Views represent the actual audience that saw a post, which is usually a fraction of followers (or much larger than followers if a video lands on the For You Page). Typical view-based engagement on TikTok in 2026:
- 5 to 10% is healthy for most creators
- 10 to 15% is good, indicating strong viewer resonance
- 15%+ is excellent, typically viral or highly targeted content
- Below 4% suggests the audience watched but didn't connect
One quirk: TikTok counts a view after 3 seconds of watch time, which inflates the view count compared to platforms with longer thresholds. Adjust expectations downward when comparing across platforms.
YouTube benchmarks (views-based)
YouTube view-based engagement rates run lower than TikTok because watching the video is itself the primary engagement. Likes and comments are a secondary signal. Typical YouTube engagement (likes plus comments divided by views):
- 1 to 3% is healthy for most channels
- 3 to 5% is good, usually on niche or community-driven channels
- 5%+ is excellent, common on highly engaged smaller channels
YouTube Shorts engagement runs closer to TikTok numbers (4 to 10%) because the format and audience behavior mirror short-form video.
Why follower count is a worse signal than engagement
A 1M-follower Instagram account at 0.5% engagement reaches about 5,000 active viewers per post. A 50K-follower account at 5% engagement reaches 2,500 active viewers. The first account often costs 20 to 50x more for only 2x the actual reach. This is the underlying math behind why micro-influencer strategies frequently outperform mega-influencer campaigns on cost per real impression.
What the number doesn't tell you
A high engagement rate doesn't mean the engagement is real, relevant, or coming from your target audience. Common manipulations include engagement pods (groups of creators artificially liking each other's posts), bot comments (cheap and easy to detect if you read them), giveaway-inflated rates that collapse after the prize is awarded, and comment-baiting captions ("tag a friend who needs this") that drive low-quality interaction.
Pair engagement rate analysis with audience authenticity scoring to filter out these patterns.
