How your scores are calculated

Your single headline number (0–100): are you actually growing? It is computed from your day-by-day cross-platform stats, so it can be tracked over time and climb as you grow. It blends four parts:

• Growth velocity (35%) — your 7-day follower growth rate. Flat sits at the middle; growing pushes it up, shrinking pulls it down.
• Engagement health (30%) — your weighted average engagement rate, measured against a ~6% “excellent” benchmark.
• Reach efficiency (20%) — how far your content travels relative to your audience size (reach ÷ followers). Reaching twice your follower count scores full marks.
• Consistency (15%) — your posting cadence over the last 7 days, against a healthy rhythm of about five posts a week.

Until there is enough daily history to be meaningful, this shows a “needs more data” state rather than a made-up number.

The score brands see when evaluating you (0–100). It is a blend of Influence and Reliability: 70% Influence + 30% Reliability once you have a collaboration track record, and Influence alone before then.

Influence ranks you against every other creator on the platform:

• Follower count (35%)
• Engagement rate (30%)
• Average views (20%)
• Your Creator DNA viral-potential (15%)

Reliability reflects how you handle brand deals:

• Acceptance rate — how often you accept or complete proposals (40%)
• Completion rate — how often accepted deals get finished (30%)
• Response speed — how quickly you resolve inquiries (30%)

If you have no real signal yet (no synced audience or performance data), you appear as a “New creator” rather than being given a misleading low score.

Before you post, the Predictor estimates a draft’s engagement, virality, retention and audience match (each 0–100); the headline Predicted Score is their average. Predictions are grounded against your own past posts on the topic and real top-performing posts in your niche when that evidence exists.

Confidenceis honest about how much to trust each prediction. Once enough of your past predictions have been checked against what actually happened, confidence is anchored to that real accuracy. Before then it reflects how much comparison data backed the prediction — and it is capped, so a prediction made from the draft text alone can never claim high certainty it hasn’t earned.

Six 0–100 traits describing your style: writing style, emotional range, hook strength, consistency, viral potential, and audience alignment. These are calculated deterministically from your real post history — engagement-rate statistics, post lengths, hook patterns and topic mix — not guessed. They unlock once you have enough analyzed posts.

When you check a draft, the analyzer rates the text itself across dimensions like hook, readability, curiosity, retention, emotion, shareability and call-to-action. This is a quality read on the writing before it goes out — it is separate from predicted results and from actual performance.

On posts you’ve published, we show measured results from real platform metrics:

• Engagement score — your real engagement rate on that post.
• Virality score — a weighted blend of engagement relative to how many people saw it: (likes + comments×2 + shares×3 + saves×1.5) ÷ impressions.

These are kept strictly separate from the predicted and draft-quality scores, so a “before you post” estimate is never confused with what actually happened.

• Grounded in your data. Scores come from your real metrics, not placeholders.
• Honest empty states. When there isn’t enough data, we say so instead of fabricating a number.
• Self-correcting. Predictions are graded against reality, and confidence adjusts to the model’s actual track record for you.