Gender is one of the most used filters on the platform. The system looks at three signals: the creator's first name, profile picture, and biography, and assigns one of three labels: male, female, or not determined. When the signal is too weak to be confident, the label stays empty.
Coverage & Breakdown
The share of profiles we've labeled, and the gender split inside the labeled set.
Coverage by Platform
How We Derive Gender
Five stages run per creator. If the name and photo agree we publish with highest confidence. If only the name resolves we still publish, at medium confidence. If neither resolves we leave the field empty rather than guess.
Three Signals, Weighted by Reliability
Each signal gets a different weight. When the strongest signals agree, confidence is high. When they conflict or are missing, the system falls back or leaves the field empty.
Business profiles and fan pages don't represent a single person, so the field stays empty for those. Coverage numbers above only count eligible individual creator profiles.
Why Coverage Differs by Platform
All three platforms are covered, with tuned thresholds, but the input quality varies. Instagram exposes the cleanest first names in handles and display names. YouTube channels are more often branded than personal, which lowers the share with a clear name. TikTok handles trend toward stylized usernames (numbers, slang), so the photo signal carries more weight there.
Strongest first-name signal. Display names and bios usually contain a clear personal name.
More branded channels. Photo analysis matters more when the channel name is generic.
Stylized handles are common. The profile picture and bio text carry more weight on this platform.
Illustrative Example
See how the multi-signal approach plays out on two illustrative creator profiles with different signal strengths.
How Gender Is Determined for Two Illustrative Profiles
When all three signals agree (like NikolaTravels), the algorithm reaches its highest confidence level. When one signal is weak or absent (BorisLifeLog has a short bio with fewer gendered cues), the algorithm can still produce a high-confidence result as long as the remaining signals are strong enough. The name database and profile picture together are sufficient for a confident prediction. The system only marks "not determined" when none of the signals resolve clearly.
Cross-Platform Signal
When we match a creator's profiles across platforms, gender data from one platform can fill gaps on another. If there's a confident gender label on two platforms but a third profile lacks enough signal, the matched profile carries the label over.
How a Gender Label Fills Across Platforms
TikTok profile lacked enough signal for detection, now filled from matched Instagram profile
Our profile matching system links accounts that belong to the same creator. When a gender label is available on one platform but the other lacks enough signal for a confident result, the matched value fills the gap. If two platforms produce conflicting labels (for example, male on one and female on the other), neither overrides the other and the profile is flagged for review.
Validation 
Every accuracy number on this page comes out of a two-track validation process. We compare at least two independent AI sources on tens of thousands of records, and we manually validate at least 1,000 records per subset, with subsets defined to cover every meaningful use case.
Measured Accuracy
Random sampling biases the validation toward dominant segments and hides failure modes in smaller ones. Stratified subsets force every meaningful use case to be represented. The cost (more AI compute, more reviewer hours) is real, but this is where the failure modes that affect customers in production tend to occur. The full validation cycle is re-run at least annually, and immediately after any algorithm change.