Engagement rate is the single most reliable metric for judging how active a creator's audience really is. A high follower count means nothing if nobody interacts with the content. This number cuts through vanity metrics and tells you what's actually happening.
The calculation: measure the engagement on recent posts (likes, comments, views) and divide by the total number of followers. The result is a clean, comparable percentage.
The exact signals differ slightly between platforms since each surfaces different types of interactions. The sections below break down how it all works.
Coverage
Engagement is computed for every profile that has at least one recent post with non-zero reach.
Coverage by Platform
How We Calculate It
The approach is the same across all platforms: engagement signals are taken from recent posts, the median per post is computed, and divided by the creator's total reach. The exact signals and reach metric adapt per platform, but the core method is consistent.
The Calculation Pipeline
Content Format Handling
On YouTube and Instagram, short-form videos get their own engagement rate: Shorts are separate from long-form videos on YouTube, and Reels are separate from feed posts on Instagram. They are separate data points, not blended into one number.
A creator's posts can vary wildly in performance. One viral post might get 50x the engagement of a typical one. An average gets skewed by that outlier and stops representing what a brand can realistically expect from a sponsored post. The median ignores outliers and reflects the creator's typical, everyday performance.
Engagement Signals
These are the signals we factor into the engagement rate. Each represents a different type of audience interaction, from passive consumption to active participation.
All Signals We Consider
The most common engagement signal. A quick positive reaction that tells you the content resonated with the viewer.
Higher effort than a like. Comments signal deeper engagement and a willingness to start a conversation with the creator.
One of the strongest intent signals. The viewer found the content valuable enough to pass it along to their own network.
A save means the viewer wants to come back to this content later. High save rates are a strong indicator of genuinely useful or inspiring posts.
Views measure reach and serve as the baseline for normalizing engagement on video-first platforms. A view alone is passive, but it's the denominator that makes the rate meaningful.
Engagement Rate for Two Illustrative Profiles
Strong engagement for this follower tier.
Healthy engagement, views-based formula.
NikolaTravels' 5.3% is calculated against followers (follower-based formula), while BorisLifeLog' 4.5% is calculated against views (views-based formula). Comparing them side by side would be misleading because the denominators measure completely different things. A follower-based rate and a views-based rate are not on the same scale. This is why the platform always shows which formula was used, so the number is read in the right context.
Validation 
Engagement rate is deterministic. The formula itself defines the answer, so there's no model accuracy to measure. What we do validate is the inputs (likes, comments, views, follower counts) using the same two-track process we use for every other data point: at least two independent AI sources cross-checking tens of thousands of records, plus manual review of at least 1,000 records per subset.
What we monitor on top of validation
Engagement is recomputed every time a profile's recent posts are refreshed. Stale inputs would show up as drift between our number and current profile data.
Posts with negative or null counts (a known data quality issue in the source data) are excluded from the median calculation rather than treated as zero.
We watch the platform-wide engagement distribution for sudden shifts. A real shift comes from a platform change (e.g. a platform hiding like counts). A sudden shift not tied to a platform change usually means a data regression.