Using Volume Data to build Real Personas

Key Concepts

Persona Loop: A four-step process for creating and testing personas: Observe (analyze demographic signals), Build (create a persona matching those signals), Inject (attach the persona to prompts), Compare (filter results by persona to see differences).

Test Harness Mindset: Personas are not biographies or customer profiles. They are controlled variables that let you test how answer engines respond to different contexts. Build them to be useful for measurement, not comprehensive.

Quick Reference

Step 1: Observe - Extract Demographic Signals

1. Go to Prompt Volumes
2. Run a single keyword analysis (select Exact Match, click Analyze)
3. Review the Demographics tab for:
   • Regions - Where prompts originate (informs language/locale decisions)
   • Age - Age range distribution
   • Income - Income bracket signals

Form a hypothesis, not a conclusion. These are signals to test, not facts to assume.

Step 2: Build - Create the Persona

1. Go to Answer Engine Insights > Personas tab
2. Click Create New Persona
3. Fill in fields that match your demographic signals:
   • Name, Gender, Age Range
   • Company Size, Industry, Seniority, Job Title
   • Behavior (what they're interested in or trying to accomplish)

Don't over-specify. Include only what changes how the engine interprets the prompt.

Step 3: Inject - Attach Persona to Prompts

1. Go to Prompts Table > Modified Prompts tab
2. Select one or more prompts
3. Click Edit Personas
4. Select your persona and add it
5. Click Save

The persona information is appended to the prompt to simulate memory or user context.

Step 4: Compare - Analyze by Persona

Once data is collected, filter your topics or prompts by persona to compare visibility scores across different audience segments.

Rules of the Road

• Don't overfit - demographics are signal, not truth
• Personas are test harnesses, not biographies
• Start broad, then tighten based on results
• Use personas where the difference changes decisions