Audiences

Who is on the panel. Each simulated reader is drawn from an audience's traits, so a clothing shop for ravers and an accounting firm should not share one. Start from the built-in audience, draft one from a description, or import a file. Audiences are saved in this browser.

Draft a set from a description

Who visits the site, and why

Ask a model

Checking for a connected model…

Use any AI chat

Copy the prompt into ChatGPT, Claude or any other chat, then paste its answer here. Without a description, the prompt leaves a gap for you to fill in.

The chat's answer

The prompt, exactly as the model gets it

You write configuration files for Jevsona, a tool in which simulated readers judge a web page. You answer with one JSON object and nothing else.

Design the simulated reader panel for this site:

[Describe your site here: what it offers, and who visits it]

Every reader on the panel is drawn at random from the traits you list, then asked to judge a page as that person.
Return an audience set in exactly this JSON shape (this example is for a different site; do not copy its content):

{
  "format": "jevsona/audience-set",
  "version": 1,
  "name": "Home cooks buying knives",
  "description": "People who cook at home most days and are choosing a kitchen knife.",
  "traits": [
    {
      "id": "role",
      "label": "who they are",
      "values": [
        "beginner_cook",
        "keen_home_cook",
        "gift_buyer",
        "ex_professional"
      ],
      "segment": true
    },
    {
      "id": "cookingFrequency",
      "label": "how often they cook",
      "values": [
        "rarely",
        "weekly",
        "daily"
      ],
      "segment": true
    },
    {
      "id": "knifeKnowledge",
      "label": "knowledge of knives",
      "values": [
        "none",
        "basic",
        "enthusiast"
      ],
      "segment": true
    },
    {
      "id": "purchaseStage",
      "label": "where they are in buying",
      "values": [
        "browsing",
        "comparing",
        "ready_to_buy"
      ],
      "segment": true
    },
    {
      "id": "budget",
      "label": "budget",
      "values": [
        "under_50",
        "50_150",
        "over_150"
      ],
      "segment": false
    }
  ],
  "concerns": [
    "sharpness_lasting",
    "safety",
    "price",
    "maintenance",
    "returns",
    "authenticity"
  ],
  "marks": {
    "fill": "knifeKnowledge",
    "ring": {
      "trait": "purchaseStage",
      "value": "ready_to_buy"
    }
  }
}

Rules:
- "format" is "jevsona/audience-set" and "version" is 1.
- 8 to 14 traits. The first trait has the id "role": 4 to 8 kinds of reader this site gets (at most 12).
- Trait ids are lowerCamelCase. Values are lower_case_with_underscores, 3 to 6 per trait, ordered low to high when they have an order.
- Labels are short lowercase phrases, as they would read in a profile: "budget", "how often they go out".
- Cover: who they are, their situation, what they already know about the topic, where they are in deciding, budget,
  what convinces them, what worries them, how carefully they read, and anything specific to this audience.
- Mark "segment": true on the 4 to 7 traits most likely to change how someone reads the page. Always include role.
- "concerns": 8 to 14 worries this audience has while deciding, lower_case_with_underscores.
- "marks.fill": a trait whose values run low to high. "marks.ring": the one value of one trait that marks readers closest to acting.
- Never use traits that should not change how a page is judged, such as ethnicity, religion or health.
- Answer with the JSON object only: no code fence, no comments, no text before or after.

Import an audience file

Your audiences

Small-business buyers built in6 kinds of reader · 15 traits · personas.v1

A run records the audience it was drawn from and its version. Two reports compare directly only when the audience and the seed match.