> ## Documentation Index
> Fetch the complete documentation index at: https://docs.qwairy.co/llms.txt
> Use this file to discover all available pages before exploring further.

# How do I improve my brand's positioning in B2B evaluation prompts?

> Compare brand narratives in selected B2B evaluation answers, verify the evidence, and test positioning or content changes.

## Direct answer

Monitor configured prompts that reflect your buyers' evaluation questions, then use **Cockpit > Compare**, stored answers, and Source Explorer to identify where SELF and DIRECT brands are described differently. Verify the pattern before testing a positioning, content, or source-distribution change.

## Data required

Use configured evaluation prompts, topics and tags, provider and model, period, stored answers, SELF/DIRECT classifications, sentiment or perception themes, citation URLs, and approved product positioning and evidence.

## Workflow

<Steps>
  <Step title="Define the evaluation cohort">
    Choose prompts tied to real B2B decisions, such as requirements, alternatives, trade-offs, implementation, and pricing. Record audience and stage as metadata.
  </Step>

  <Step title="Compare observed narratives">
    Review SELF and DIRECT occurrences in Compare and read the associated answers. Separate presence, rank order, claims, sentiment, and citations.
  </Step>

  <Step title="Verify important differences">
    Check claims against approved product facts and inspect recorded sources. Do not infer why a provider produced the difference from the output alone.
  </Step>

  <Step title="Select one test">
    Choose a material, supportable gap and define the content, positioning, or distribution hypothesis plus the cohort you will monitor.
  </Step>
</Steps>

## Interpretation

Product Share of Voice is SELF mention occurrences divided by SELF plus DIRECT occurrences. It measures competitive occurrence in the selected answers, not preference, pipeline, or purchase intent. Narrative analysis must stay linked to its prompts, providers, models, and period.

## Possible next actions

* If SELF is absent from a relevant comparison, test an evidence-led page addressing the buyer's decision criteria.
* If an inaccurate weakness recurs, verify its sources and test a correction workflow.
* If the gap appears only at one evaluation stage, focus the hypothesis there rather than changing all messaging.

## Limitations

Configured prompts approximate selected buyer questions; they do not represent every procurement process. Provider outputs can vary and do not predict purchase decisions. A later change in answers does not prove that your intervention caused it.

## Related pages

<CardGroup cols={2}>
  <Card title="Exports" href="/documentation/workspace/exports">
    Export the evaluation cohort.
  </Card>

  <Card title="Competitor Metrics" href="/looker-studio/sources/competitor-metrics">
    Build documented competitor reporting.
  </Card>

  <Card title="Shared Links" href="/documentation/workspace/shared-links">
    Share a filtered comparison.
  </Card>

  <Card title="Consideration-stage presence" href="/use-cases/funnel/brand-in-consideration-phase">
    Review brand presence during consideration.
  </Card>

  <Card title="SaaS best-tools answers" href="/use-cases/industry/saas-in-best-tools-responses">
    Analyze recommendation-list prompts.
  </Card>

  <Card title="Pages cited by AI" href="/use-cases/content/pages-most-cited-by-ai">
    Inspect citation URLs supporting answers.
  </Card>
</CardGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.