Skip to content
Reading your numbers

Creative metrics

Read the creative features behind each video without confusing observations with platform counts.

Steps

  1. Pick the biggest proven gap within one brand and channel.
  2. Open its examples and compare the opening, delivery and brand moments.
  3. Make a brief from the finding, then check a finished cut against its targets.
  4. Attention (start) — Hook type: The device that earns attention at the opening. AI-read; category. Compare examples before choosing the next creative test.
  5. Attention (start) — Hook strength: AI reading of how clear and attention-grabbing the opening is. AI-read; 1–10. Test a clearer opening.
  6. Attention (start) — Face in first second: Whether a visible face appears within the first second. AI-read; yes/no. Compare examples before choosing the next creative test.
  7. Attention (start) — Seconds to product: When the product first becomes visible. AI-read; s. Choose when the product enters the story.
  8. Attention (start) — First logo time: When a recognisable brand logo first appears. AI-read; s. Place the brand moment deliberately.
  9. Attention (start) — Mobile hook strength: AI reading of how well the opening reads on a phone. AI-read; 1–10. Test an opening that reads on a small screen.
  10. Story (middle) — Story arc: The structure connecting the opening to the ending. AI-read; category. Compare examples before choosing the next creative test.
  11. Story (middle) — Story beats: Recorded story moments with start and end timestamps. AI-read; beats. Compare examples before choosing the next creative test.
  12. Story (middle) — Pacing: Observed editing cuts per second. AI-read; cuts/s. Test a different editing rhythm.
  13. Story (middle) — Emotional curve: Recorded emotional intensity and valence through the video. AI-read; timeline. Compare examples before choosing the next creative test.
  14. Story (middle) — Emotional peak position: Where the strongest observed emotion occurs in the video. AI-read; % of length. Choose where the emotional high point lands.
  15. Story (middle) — Payoff time: When the video delivers the hook’s promised reveal, result, punchline or answer. AI-read; s. Test when to deliver the promise.
  16. Story (middle) — Payoff type: The kind of delivery: reveal, result, punchline, answer, transformation, offer or none. AI-read; category. Compare examples before choosing the next creative test.
  17. Story (middle) — Hook-to-payoff gap: Seconds between the hook ending and the promised delivery. AI-read; s. Test a shorter wait after the hook.
  18. Story (middle) — Promise kept: Whether the finished story delivers its opening promise, with a factual reason. AI-read; category. Compare examples before choosing the next creative test.
  19. Story (middle) — Loopable: Whether the end flows naturally back into the start. AI-read; yes/no. Compare examples before choosing the next creative test.
  20. Brand — Seconds to product: When the product first becomes visible. AI-read; s. Set the product reveal.
  21. Brand — Seconds to logo: When the brand logo first appears. AI-read; s. Set the logo reveal.
  22. Brand — Product screen time: Share of the video during which the product is visible. AI-read; %. Choose the amount of visible product.
  23. Brand — Brand cues: Observed recognisable brand elements in the content. AI-read; category. Compare examples before choosing the next creative test.
  24. Brand — Mascot present: Whether an identifiable brand mascot appears. AI-read; yes/no. Compare examples before choosing the next creative test.
  25. Brand — Integration naturalness: AI reading of whether the product fits the story naturally or feels forced. AI-read; 1–10. Fit the product into the story rather than interrupting it.
  26. Brand — Creator on-screen time: Share of video showing an identifiable creator; a face alone is not creator evidence. AI-read; %. Choose the creator’s role in the story.
  27. Mobile fit — 9:16 format: Whether the provider records a vertical 9:16 video. Platform count; ratio. Choose a format suitable for the placement.
  28. Mobile fit — Sound-off ready: Whether the message remains understandable without audio. AI-read; category. Compare examples before choosing the next creative test.
  29. Mobile fit — Subtitles present: Whether readable dialogue captions are visible. AI-read; yes/no. Compare examples before choosing the next creative test.
  30. Mobile fit — Text density: Observed amount of on-screen text. AI-read; category. Compare examples before choosing the next creative test.
  31. Mobile fit — Safe-zone issues: Recorded text or product overlaps with platform interface areas. AI-read; category. Compare examples before choosing the next creative test.
  32. Language and tone — Language mix: Observed proportions of languages; location does not establish language. AI-read; language %. Compare examples before choosing the next creative test.
  33. Language and tone — Tone tags: Observed emotional or writing style. AI-read; category. Compare examples before choosing the next creative test.
  34. Language and tone — Occasion: The evidenced occasion or context used by the content. AI-read; category. Compare examples before choosing the next creative test.
  35. Language and tone — Trending audio: Provider-confirmed trend status attached to a recorded music ID, never AI familiarity. Platform count; yes/no. Check music licensing before use.
  36. Response — Views vs usual: Saved views relative to this brand’s own usual on the same channel. Platform count; × usual. Identify examples above the brand’s own usual.
  37. Response — Engagement per view: Saved known engagements divided by known views. Platform count; %. Compare response, not reach.
  38. Response — Shares: Known platform share count; unavailable is not zero. Platform count; shares. Look for content people pass on.
  39. Response — Saves: Known platform save count; unavailable is not zero. Platform count; saves. Look for useful content people retain.
  40. Response — Comments per 1K views: Known comments relative to 1,000 known views. Platform count; per 1K. Read the comments before assuming buying intent.
  41. Response — Comment intent mix: Recorded comment classifications, not assumed purchase intent. Platform count; mix. Distinguish questions and buying intent from reactions.
  42. Response — Call to action: The observed request to shop, follow, comment, visit or do none. AI-read; category. Compare examples before choosing the next creative test.
  43. Owner-account only — Average watch time: Mean seconds watched, supplied by the connected account. Connected account; s. Needs connected account; availability depends on platform approval.
  44. Owner-account only — 3-second views: Account-reported views reaching three seconds. Connected account; views. Needs connected account; availability depends on platform approval.
  45. Owner-account only — Completion rate: Account-reported share reaching the end. Connected account; %. Needs connected account; availability depends on platform approval.
  46. Owner-account only — Retention curve: Account-reported audience remaining through the video. Connected account; timeline. Needs connected account; availability depends on platform approval.

What good looks like

Coverage states N of M eligible videos. Below 30% stays under Not enough data yet; proven requires at least eight videos and a consistent separated interval.

Common mistakes

  • Treating AI readings as causal explanations.
  • Guessing payoff or audio trend status.
  • Adding channels together or calling missing account metrics zero.

Related articles