Skip to main content
Version: 2025-12-18

Understanding blended profiles

This article explains the concept of blended profiles (previously known as interlocked or nested quotas) in Cint’s profiling system. It covers what blended profiles are, how they function, why researchers use them, and the operational considerations when applying them to target groups.

What are blended profiles?​

Blended profiles allow you to combine options from two or more profiling questions into a single, multi-dimensional targeting segment. Instead of setting quotas for each question independently (e.g., gender alone, age alone), a blended profile defines combined segments, such as:

  • Females aged 18–24
  • Males aged 35–44
  • Non-Hispanic Females aged 25–34

In a blended profile:

  • Respondents must satisfy all conditions simultaneously.
  • Each segment is represented as a target within the blended profile.
  • Each target has its own quota definition (e.g., completes_goal_percentage, completes_goal).

Blended profiles effectively create a matrix of characteristics, where each matrix cell represents one blended target segment.

Why researchers use blended profiles​

Researchers use blended profiles to achieve precise audience control beyond simple demographic quotas.

Break down the audience into smaller, more specific groups​

Blended profiles allow you to combine characteristics such as:

  • Age × Gender
  • Gender × Ethnicity
  • Household Income × Parental Status

This provides deeper segmentation and more meaningful analysis.

Ensure balanced distribution​

Blended profiles help you maintain balanced representation across combined characteristics. For example, ensuring equal male/female distribution within every age band, not just overall.

Prevent skews​

Without blending, a survey might fill large quotas with the easiest-to-find respondents. Blended profiles prevent this by explicitly defining each allowed combination.

How blended profiles work in the latest API contract​

In the new profiling model:

  • A blended profile is identified using:
{
  "object": "blended"
}
  • Each blended segment is represented as a target inside the profile.
  • Each target contains:
    • A list of conditions (selection or range)
    • A quota object

Example blended target object:

{
  "text": "Female, Yes, Mexican, Mexican American, Chicano",
  "conditions": [
    { "object": "selection", "question_id": 43, "option": "2" },
    { "object": "selection", "question_id": 47, "option": "2" }
  ],
  "quota": {
    "completes_goal_percentage": 100
  }
}

Key points:​

  • Blended profiles do not support open-ended questions.
  • You may create blended profiles using:
    • The generate blended profile API (recommended)
    • In the output from this endpoint, quotas are defaulted based on historical supplier distribution. Review and update before you save if desired.
    • Only one blended profile is allowed per target group.
    • Manually constructing your own blended profile in the draft target group request.
note
  • A given question_id can appear in one and only one target.
  • A given question_id+option combination can appear only once.

Why blending may affect feasibility, cost, and field time​

Blended profiles, while powerful, come with operational considerations.

⚠ Limited feasibility​

The more characteristics you combine, the smaller each segment becomes. In extreme cases, some blended segments may have very low or zero respondent availability. This will be especially true for multi-punch profiles; you may get timeouts or errors since calculating feasibility on multi-punch profiles can be difficult.

⚠ Higher costs​

Hard-to-reach segments (e.g., “Males aged 55+ with high income and multiple children”) may increase CPI due to scarcity.

⚠ Slower fieldwork​

More precise segments take longer to fill, especially when multiple blended targets must reach their completes goal.

Blending should be used when necessary, but avoided for overly granular or unnecessarily complex segmentation.

Blended profiles vs. previous interlocked profiles (quick comparison)​

Previous contractNew contract
Called interlocked profilesNow called blended profiles
Required full expansion of all conditionsManual or system expanded conditions
Used depends_on_questionsNow inferred from conditions in each target
Required ungrouped/grouped dictsNow uses targets[] with explicit conditions
Quotas defined by quota_percentageNow uses completes_goal_percentage or completes_goal
Could only be created via API logicUsers may also manually construct blended profiles

Example of a blended profile​

You can pass two or more regular profiles into generate blended profile endpoint to generate a blended profile payload.

note

Generating blended profiles does not change any data on your target group. This is a helper endpoint that generates a complex JSON payload that you can plug into your draft target group in order to save and launch it.

The output of generate blended profile will discard any quota values passed into it; we will apply our prediction of the right distribution. You can review and edit these values before you save your target group.

Request​

Example request payload to generate a blended profile:

{
  "locale": "eng_us",
  "completes_goal": 1000,
  "start_at": "2026-10-09T12:00:00.000Z",
  "end_at": "2026-10-23T12:00:00.000Z",
  "collects_pii": false,
  "profiling": {
    "profile_adjustment_type": "percentage",
    "profiles": [
      {
        "object": "regular",
        "quotas_enabled": true,
        "targets": [
          {
            "conditions": [
              { "object": "selection", "question_id": 43, "option": "2" }
            ],
            "quota": { "completes_goal_percentage": 40 }
          }
        ]
      },
      {
        "object": "regular",
        "quotas_enabled": true,
        "targets": [
          {
            "conditions": [
              { "object": "selection", "question_id": 47, "option": "2" }
            ],
            "quota": { "completes_goal_percentage": 40 }
          }
        ]
      }
    ]
  }
}

Response​

Example response payload with a blended profile structure:

{
  "locale": "eng_us",
  "profile_adjustment_type": "percentage",
  "profiles": [
    {
      "object": "blended",
      "quotas_enabled": true,
      "name": "GENDER, HISPANIC",
      "targets": [
        {
          "text": "Female, Yes, Mexican, Mexican American, Chicano",
          "conditions": [
            { "object": "selection", "question_id": 43, "option": "2" },
            { "object": "selection", "question_id": 47, "option": "2" }
          ],
          "quota": { "completes_goal_percentage": 100.0 }
        }
      ]
    }
  ]
}

Best practices​

  • Keep the number of blended dimensions reasonable
    • Two-way blends (Age × Gender) are generally feasible.
    • Three or more blends significantly reduce respondent availability.
    note

    Our system will restrict the number of profiles that can be blended, and total number of quotas that can be defined. Check any error messages carefully to see if you have hit system limits.

  • Validate feasibility before applying very strict blends
    • Check historical incidence to provide feasibility information.
  • Use the generate blended profile endpoint whenever possible
    • It ensures:
      • Accurate supplier distribution
      • Correct expansion logic
      • Valid quota calculations
  • Do not use open-ended questions in blended profiles
  • Multi punch questions are generally not recommended in blended profiles, as one respondent may qualify for more than one quota.

Summary​

Blended profiles provide:

  • Precision in targeting
  • Balanced representation across combinations
  • Controlled sampling for multi-dimensional respondent characteristics But they also require careful planning due to feasibility, cost, and fieldwork timing implications.

By using the new blended profile structure and the generation endpoint, you can implement powerful, multi-dimensional quotas in draft target groups with improved clarity and flexibility.