Skip to main content

​
Overview

Keyword Explorer uses machine learning to offer AI-powered keyword recommendations for building interest-based cohorts. By analyzing the similarities and connections between keywords, it suggests relevant terms that maximize both reach and relevance.
Prerequisites:
  • Access to the Permutive Dashboard with cohort creation permissions
  • Keyword Explorer enabled for your workspace
  • Contextual API implemented on your website OR Permutive SDK deployed (the SDK can segment contextual cohorts without direct API implementation)

​
Steps

1

Select Keyword Explorer

Select Keyword Explorer as your predicate type. A modal will open with the Keyword Explorer interface.
2

Enter topic

Enter a Topic in the search bar and press Enter. Topics can be single words (e.g., “Fashion”) or multiple words (e.g., “Tour de France”). Add multiple topics to narrow down to a specific concept (e.g., “cocktail” + “drink” vs “cocktail” + “dress”). Alternatively, you can input a prompt in the form of a targeting requirement (e.g., “Users interested in luxury travel”). Keyword Explorer will first transform the prompt into relevant topics and then show the results.
3

Review recommendations

Review the Recommendations table, which includes:
  • Name: Similar or adjacent keywords to your topic
  • Property: The Pageview properties where each keyword appears
  • Relevance: How closely related the keyword is to your topic (based on semantic similarity)
  • Reach: Incremental reach (accounts for overlap between already-selected keywords)
4

Use filters

Use Filters to refine recommendations:
  • Properties: Limit to specific Pageview properties
  • Reach: Set a minimum reach threshold
  • Countries: Filter by geographic location (based on geo_info)
  • Domains: Filter by specific domains (based on client.domain or client.url)
5

Select keywords

Select the keywords you want to include by checking the boxes next to them.
6

Review total reach

Review the Total Reach (deduplicated user count for your selection).
7

Add keywords

Click Add to apply your keyword selections to the cohort.
Using keyword explorer

​
Example

Build a Fashion cohort using Keyword Explorer:
  • Topic: “Fashion”
  • Select relevant recommendations: “style”, “designer”, “clothing”, “trends”, “runway”
  • Filter by Property: client.title, client.tags
  • Filter by Reach: Minimum 1,000 users
  • Total Reach: 45,000 users (deduplicated)
Keyword Explorer SettingsKeyword Explorer settings are configured at the organization level under Settings > Integrations > Keyword Explorer. You must be at the root workspace to access these settings.
  • Minimum Reach: Set a global minimum reach threshold for all recommendations
  • Property Selection: Choose which Pageview properties to include in the model training and recommendations
  • Tokenize: Enable for properties that require keyword extraction (like article titles), disable for properties that already contain individual keywords (like tags)

​
Next Steps

Creating Contextual Cohorts

Learn the full cohort creation process

Back to Contextual Cohorts

Return to product overview