Skip to main content
Advertising

PPC Relevance Check

mx-ppc-relevance-check

Semantic relevance classification for search terms and ASIN targets: separate from threshold logic, with confidence scores and reasoning.

Run it in chat
/mixshift-ai:mx-ppc-relevance-check <brand>

or just say run relevance check for <brand>

Comes with the plugin, one install activates every skill. See the install flow. Needs brand setup first.

Why it exists

A search term or ASIN can be cheap and still be exactly the right customer, or expensive and still be the wrong one entirely, and pure spend or ACoS thresholds cannot tell the difference between the two.

What you get

A verdict table classifying each term or ASIN as RELEVANT, BORDERLINE, or IRRELEVANT with a confidence score and a one-sentence reason, so negation decisions are made on customer intent, not on data alone.

What your agent can do

  • 01

    Classify terms as relevant or irrelevant independent of spend thresholds

  • 02

    Return confidence scores and human-readable reasoning per call

  • 03

    Requires brand-specific training and account-manager calibration on edge cases

  • 04

    No action recommendations: this skill classifies, it does not decide

StatusLive
CategoryAdvertising
Brand contextRequired
Hosts
Claude CoworkClaude CodeMixShift CLI
Get the plugin

New or existing MixShift customer, both paths start here.

Talk to us

Before you run

Requires brand setup

This skill reads a brand’s context to calibrate its output. Build it once per brand, then every teammate runs on the same shared context.

/mixshift-ai:mx-brand-context <brand>How brand setup works →
What it reads from your brand context
  • Brand terms and sub-brand product descriptions: the classifier’s baseline for this brand’s customer
  • Negation lane rules and protected terms: the brand-specific dictionaries every verdict is checked against
  • ASIN lifetime-orders threshold: gates which ASIN targets enter the relevance phase
  • Narrative positioning and borderline-theme notes: the judgment calls a pure keyword match cannot capture
Gets sharper over time

Every time an account manager overrides a verdict, that override becomes a candidate lane rule or protected term for the brand context, so the classifier’s borderline calls shrink and its confidence rises on that brand’s specific edge cases run over run.