> For the complete documentation index, see [llms.txt](https://oodlz-ai-studio.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://oodlz-ai-studio.gitbook.io/docs/outbound-agents/awareness/1-topics/score-filter-create.md).

# Score, Filter, Create

Every topic scored on Appeal first, then a 3-factor composite (Appeal + Rankability + Volume), labelled Hot/Warm/Cold, deduplicated against history, and written to the approval queue.

> **Volume is the wrong metric. Composite scoring on appeal, rankability, and demand picks the topics that actually move the needle.**

Raw search volume ranks the wrong topics. A 50,000-volume keyword with difficulty 95 is a worse opportunity than an 8,000-volume keyword with difficulty 15 — and both can be worse than a 3,000-volume topic the brand's audience would actually click on. The Oodlz Topics producer scores every candidate on a 3-factor composite — Appeal first, Rankability second, Volume and Trend last — before writing the calendar.

## How It Works

Every topic passes through Appeal scoring before any composite is calculated. An LLM rates each topic on consumer appeal — would a real person click this, share it, bookmark it — without seeing any SEO data. Scores run 1 to 100. The scoring uses 2 modes: informational topics are judged on utility, actionability, and shareability; emotional topics are judged on psychological accuracy and rawness.

Real search volume is fetched per market using the correct geographic location codes. Keyword difficulty merges through a priority chain — real keyword difficulty values from competitor ranked keywords first, neutral defaults only when real values are unavailable.

The composite formula:

| Factor                    | Weight |
| ------------------------- | ------ |
| Appeal                    | 50%    |
| Rankability (inverted KD) | 30%    |
| Volume and Trend          | 20%    |

Topics are then labelled Hot, Warm, or Cold based on composite score. A final AI semantic deduplication pass compares each candidate against the brand's existing topic library — exact-match dedup is not enough, semantic similarity catches the same topic phrased differently. Surviving topics are written to the [approval queue](/docs/outbound-agents/awareness/approval-gates/gate-1-topics.md) with correct market tags, and a run log records discovery counts, dedup counts, and write counts on completion.

## What You Get

* Appeal scored before composite — click-worthiness is the dominant signal, not volume.
* 3-factor composite scoring — the right balance of human pull, SEO rankability, and demand.
* Dual scoring modes — informational topics judged on utility, emotional topics judged on resonance.
* Hot / Warm / Cold labels — quick scan during approval, no need to interpret raw scores.
* Semantic dedup against history — the same topic never reappears in the calendar even when phrased differently.
* Per-market tagging — every topic carries its destination market, scoring is comparable within and across markets.