Scoring Methodology
The score
Every keyword gets a Personalized Difficulty Score from 0 to 100, where 0 is easy (you can win) and 100 is tough. It's computed live from the current top-ranking videos and your channel.
| Score | Band |
|---|---|
| 0–39 | Winnable |
| 40–64 | Moderate |
| 65–100 | Tough |
The factors
The score is a weighted blend of four factors. Each is itself a 0–100 difficulty:
| Factor | Weight | What it measures |
|---|---|---|
| Creator Authority | 35% | Your subscriber count vs. the median subscriber count of the top-ranking videos. Personalized. |
| Engagement Gap | 25% | Your channel's engagement rate (likes ÷ views) vs. the median of the top results. Personalized. |
| Content Freshness | 20% | How recent the top-ranking videos are. Fresher results are harder to displace. SERP-only. |
| Format Match | 20% | How tightly the top results cluster on video length. A strict format bar is harder to break into. SERP-only. |
Creator Authority and Engagement Gap are personalized — they compare you to the SERP using serp ÷ (serp + you), which sits at 50 when you're at parity, drops when you dominate, and rises when you're behind. Content Freshness and Format Match are derived from the SERP alone, so they're the same for every channel.
Without a connected channel, the personalized factors fall back to parity and the score is less specific to you.
Confidence
Not every factor always has data (e.g. missing engagement numbers). Each score carries a confidence — low, medium, or high — based on how many of the factors had real data behind them. Treat a low-confidence score as a rougher estimate.
The draft scorecard
Score your draft runs a separate pre-publish checklist against your draft title and caption — no API data involved, it reads only what you paste:
| Check | What it looks for |
|---|---|
| Keyword in caption | The keyword phrase appears naturally in your caption |
| No stuffing | Repeating the phrase over and over lowers the score — over-optimization reads as spam to platforms and people |
| Title states the topic | A searcher (and the algorithm) can tell what the video is about |
| Secondary keyword | Only scored when you provide one — never guessed |
| Hashtags | 3–5 topical hashtags; zero and spam both cost you |
Derived, and labeled as derived
Difficulty scores, engagement rates, intent labels, and scorecard results are computed by tuner.social from YouTube API data — they are not numbers YouTube provides. Everywhere one appears you'll see the badge "Derived by Tuner — not a YouTube metric." If we can't measure something, we don't show it: there is no invented search volume, no sampled intent, no fabricated trend anywhere in the product.
On the roadmap
The current model uses the four factors above. Two more are planned and not yet live: Topical Relevance (does your channel have a track record in the topic?) and Ranking Volatility (how much rankings churn, once enough history accumulates).
Limitations
- Scores reflect current competition and can change as the SERP changes.
- Built for social search; it doesn't model external traffic or feed-based discovery.
- The model is being actively calibrated against expert human judgment — treat scores as directional, not guarantees.