Why SparkVox Curates Fewer Posts Per Project

A 45-minute recording can contain dozens of interesting sentences. It cannot contain dozens of standalone LinkedIn posts worth publishing. SparkVox used to err on the side of volume. It now curates harder so each draft earns its place in your review queue.

Two numbers on the project ready screen

When processing finishes, you may see something like 14 moments found, 8 posts curated. Found is how many insight-rich candidates scored 70+ on relevance. Curated is how many Sparky actually drafted - the highest-scoring subset for that recording length (roughly eight posts per 30 minutes of content, with a hard cap of 15).

Weak moments are dropped even when there is room under the cap. A sparse pool cannot sneak a low-scoring moment through just because the quantity target was not met.

Why fewer posts is the point

Publishing twelve mediocre posts from one call trains your audience to ignore you. Publishing six sharp posts trains them to read. Curation is the difference between a recap factory and a content system.

Each surviving post still maps to a funnel stage (Awareness, Authority, or Conversion) and gets a topic label tied to subject matter - not post format. You review every draft before anything schedules.

What to do when you want more volume

Longer recordings naturally yield more curated posts. So does splitting very different conversations into separate projects instead of one mega-upload. If a meeting was genuinely thin - a status check-in with no standalone insights - zero strong posts is a valid outcome.

Curation also learns from what you discard and heavily edit on past projects, so weak themes are less likely to refill the queue. How discard signals feed the next project.

Related: full draft pipeline (two candidates, distinctiveness), why multiple themes beat one generic summary, how project processing works, and Voice Confidence explained.