The math behind every score. Shown, not hidden.
Every number Cyte shows has a formula behind it. Nothing is proprietary. Nothing is hidden behind “trust our algorithm.” This is the exact math.
Read it, challenge it, tell us we got it wrong. That's the point.
Maximum signal contributions. Scores are clamped to 0–100.
The scoring formula
Tracked prompts run across 7 AI engines: ChatGPT, Claude, Perplexity, Gemini, DeepSeek, Grok and Google AI Overviews. Every answer gets analyzed by Claude Haiku (structured JSON extraction) to detect, for you and each competitor: whether the brand was mentioned, in what position in any list, how many times, and with what sentiment. Those four signals produce a score from 0 to 100 per brand.
A worked example: named first, mentioned three times, described positively. 35 + 28 + 8 + 12 = 83. Named fourth, mentioned once, described negatively: 35 + 13 + 0 − 8 = 40.
That's the whole formula. We chose simplicity on purpose. Transparent scoring gets challenged, iterated, and improved. Proprietary scoring gets defended. Ours is the former.
The four signals, explained
Either the AI engine mentioned your brand or it didn’t. This is a binary gate. If your brand isn’t in the response, nothing else matters — you score 0. This is why presence is the foundation of every other metric.
When AI engines list options, order matters. The first recommendation gets 3–5× more consideration than the third — measured in our own test-prompt panels, not an external study. First position gets +28, second +22, third +17, fourth +13, fifth +9, and any position after that +4.
If your brand is mentioned multiple times in the same response, that’s a stronger signal than a single passing mention. Formula: (mention_count − 1) × 4, capped at 18. So 2 mentions = +4, 3 mentions = +8, 5 mentions = +16, 6 or more = +18. Capped so a single answer spamming your name doesn’t inflate scores.
How the AI describes you. Positive (+12), mixed (+3), negative (−8). Negation-aware — ‘not the best’ correctly scores negative. Claude Haiku does the sentiment analysis with explicit prompting against bias.
How aggregates work
Each prompt × AI engine produces one score per brand. A brand's overall score for a prompt is the average across all supported AI engines. A brand's overall presence is the percentage of prompts where they're mentioned at all (independent of score).
We never weight AI engines differently by default. ChatGPT doesn't count more than Gemini. This is a design choice: your visibility should be measured against the ecosystem as it actually is, not as it might be weighted.
Time-series scores are raw. We don't smooth, normalize, or anti-alias. If your score jumps from 62 to 78 in one day, that's what happened. Volatility is data.
Why one number isn't enough
Other tools collapse everything into a single proprietary “visibility score.” Every prompt counted equally — the definitional “what is a CRM” weighted the same as the high-commercial “best CRM for B2B SaaS” that your buyers actually ask before opening a tab. Every source counted equally — a Reddit comment carrying the same weight as a Forbes article in the math.
That's the easy version. It produces a number, and the number changes, and you can't act on it because no one can tell you why it changed.
Cyte refuses to flatten. Every score decomposes into the four signals above. Every gap traces back to the exact AI answer that produced it. Every idea names the prompt it's targeting, the competitor page it's up against, and the evidence behind it. Read the math. Argue with it. Teach your team to read it. That's the deal.
Gap detection
After every run, we classify each prompt into one of four states:
Priority score = (competitor_lead) + (20 if absent) + (15 if declining) + (5 if new_opportunity), clamped 0–100. Pure deterministic math. No AI involvement in priority ranking — keeps gap detection fast and zero-cost.
How Cyte treats cause
A competitor changes a page. A score moves. Cyte shows the two together as an observation, with its evidence and a confidence level. It never states that a change caused a ranking.
The confidence level is set by rule, not by judgment. An observation gets the highest level only when all four of these hold:
Miss any one and the level drops. The wording follows: “possibly related”, “worth checking”, “what the evidence shows”. You see the evidence either way and can judge it yourself.
Outcome checks
When you mark an idea done, Cyte re-checks the related prompts at 7, 14, 28 and 42 days and shows whether your visibility moved.
The same rule applies to your own changes as to a competitor's. Cyte reports what happened afterwards and how confident the link is. It does not promise a lift.