How we label metrics and estimate authenticity

Our rule: what we can't prove, we don't present as fact. Every number on ProofKol carries a source label and a capture date. Missing values are shown as "—", never as zero.

Source labels

Self-reported Self-reported — entered by the creator. Useful, but not independently checked. Creators can only ever save metrics as Self-reported.

Public Public — taken from publicly visible data or an official public API (for example the YouTube Data API: subscribers and per-video views).

Verified Verified — from the platform's own insights, either via a connected account (OAuth, rolling out per platform) or after our team reviewed native analytics with the creator.

Authenticity estimate (V1)

The authenticity score (0–100) is an estimate. It is a simple, transparent heuristic based on the data available on the profile:

• Engagement rate compared with the typical range for the account's size and platform.
• Unusual follower growth spikes in the last 30 days.
• Average views compared with follower count on the main content format.

A lower score is a hint to look closer — not an accusation, and not legal proof. Full bot detection requires data that only the platforms have; once a creator connects their account, we'll compute a documented, verified score from first-party insights.

Bands: 80+ looks organic · 60–79 mostly organic · 40–59 mixed signals · below 40 review carefully.

Conversion data

Social platforms almost never expose conversion rates. ProofKol will measure them with its own tracking links and promo codes. Until then we don't show conversion numbers we can't back up.

Platform coverage

YouTube offers the strongest public data. Instagram and TikTok insights require the creator to connect a business/creator account. X, LinkedIn and Telegram are currently Self-reported plus a link to the profile, which brands can check themselves.