Understanding blended profiles

This guide explains the concept of blended profiles in Cint’s profiling system, including the interlock and control types. 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.

---

Blended types: interlock vs. control

The blended object now includes an optional blended_type field (enum) that determines how the profile behaves:

Blended types	Description	Default behavior	Use case
interlock	Used for full matrix expansion. Ensures balanced representation across all combined characteristics.	This is the default if the blended_type field is not populated.	Standard, balanced segmentation (e.g., Age x Gender).
control	Used for targeted management control. Allows researchers to cap or pause specific high-incidence groups without expanding the full matrix.	Must be explicitly set to control.	Capping or monitoring specific, high-incidence groups.

---

What is an interlock blended type?

An interlock blended type expands every possible combination of conditions from two or more profiling questions into individual targets. Each target receives its own quota, ensuring balanced representation across all combined characteristics.

This is the default behavior when blended_type is not specified.

Example of a blended profile with interlock as blended type

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 the generate blended profile endpoint will discard any quota values you provide in the request. Instead, it applies our prediction for the optimal distribution based on historical data. 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": "1" }
            ],
            "quota": { "completes_goal_percentage": 50 }
          },
          {
            "conditions": [
              { "object": "selection", "question_id": 43, "option": "2" }
            ],
            "quota": { "completes_goal_percentage": 50 }
          }
        ]
      },
      {
        "object": "regular",
        "quotas_enabled": true,
        "targets": [
          {
            "conditions": [
              { "object": "selection", "question_id": 47, "option": "1" }
            ],
            "quota": { "completes_goal_percentage": 60 }
          },
          {
            "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,
      "blended_type": "interlock",
      "name": "GENDER, HISPANIC",
      "targets": [
        {
          "text": "Male, No, not of Hispanic, Latino, or Spanish origin",
          "text_translated": "Male, No, not of Hispanic, Latino, or Spanish origin",
          "conditions": [
            { "object": "selection", "question_id": 43, "option": "1" },
            { "object": "selection", "question_id": 47, "option": "1" }
          ],
          "quota": {
            "name": "Male, No, not of Hispanic, Latino, or Spanish origin",
            "completes_goal_percentage": 42.0,
            "completes_goal": 420
          }
        },
        {
          "text": "Male, Yes, Mexican, Mexican American, Chicano",
          "text_translated": "Male, Yes, Mexican, Mexican American, Chicano",
          "conditions": [
            { "object": "selection", "question_id": 43, "option": "1" },
            { "object": "selection", "question_id": 47, "option": "2" }
          ],
          "quota": {
            "name": "Male, Yes, Mexican, Mexican American, Chicano",
            "completes_goal_percentage": 3.6,
            "completes_goal": 36
          }
        },
        {
          "text": "Female, No, not of Hispanic, Latino, or Spanish origin",
          "text_translated": "Female, No, not of Hispanic, Latino, or Spanish origin",
          "conditions": [
            { "object": "selection", "question_id": 43, "option": "2" },
            { "object": "selection", "question_id": 47, "option": "1" }
          ],
          "quota": {
            "name": "Female, No, not of Hispanic, Latino, or Spanish origin",
            "completes_goal_percentage": 50.2,
            "completes_goal": 502
          }
        },
        {
          "text": "Female, Yes, Mexican, Mexican American, Chicano",
          "text_translated": "Female, Yes, Mexican, Mexican American, Chicano",
          "conditions": [
            { "object": "selection", "question_id": 43, "option": "2" },
            { "object": "selection", "question_id": 47, "option": "2" }
          ],
          "quota": {
            "name": "Female, Yes, Mexican, Mexican American, Chicano",
            "completes_goal_percentage": 4.2,
            "completes_goal": 42
          }
        }
      ]
    }
  ]
}

---

What is a control blended type?

A control blended type allows users to select specific conditions from two or more existing profiles and group them under a single blended profile, with per-target quotas. It acts as a hybrid between interlock blended type and standard quota grouping.

Unlike a default blended profile (interlock blended type), which generates every possible combination of selected attributes, a control blended type allows you to pick specific combinations—such as "White Males aged 18–24"—to monitor or cap individually.

---

Why researchers use blended profiles

• Granular Control (Interlock): To break down the audience into smaller, specific groups (e.g., Age × Gender) and ensure equal distribution within every age band.
• Targeted Management (Control): To prevent specific demographic groups from filling too quickly and consuming the larger pool (e.g., capping a specific high-incidence group mid-project).

Break down an 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 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": "Male, No, not of Hispanic, Latino, or Spanish origin",
  "text_translated": "Male, No, not of Hispanic, Latino, or Spanish origin",
  "conditions": [
    { "object": "selection", "question_id": 43, "option": "1" },
    { "object": "selection", "question_id": 47, "option": "1" }
  ],
  "quota": {
    "name": "Male, No, not of Hispanic, Latino, or Spanish origin",
    "completes_goal_percentage": 42.0,
    "completes_goal": 420
  }
}

If the blended type is not provided, the system will follow default behavior (interlock).

Example: Blended Profile with control blended_type In this example, the blended type is set to control to target a specific combination of gender, age, and state.

{
  "object": "blended",
  "name": "Control quota 1",
  "quotas_enabled": true,
  "blended_type": "control",
  "targets": [
    {
      "name": "Quota 1",
      "text": "Male, 20 to 29, ALABAMA",
      "conditions": [
        {
          "object": "selection",
          "question_id": 43,
          "option": "1"
        },
        {
          "object": "range",
          "question_id": 42,
          "min": 20,
          "max": 29
        },
        {
          "object": "selection",
          "question_id": 96,
          "option": "1"
        }
      ],
      "quota": {
        "completes_goal_percentage": 0.0,
        "completes_goal": 0
      }
    }
  ]
}

Key points:

• Blended profiles do not support open-ended questions.
• To create blended profiles with interlock blended_type, use:
  • 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 interlock blended profile is allowed per target group.
  • Manually constructing your own blended profile in the draft target group request.
    • You can check the structure via the validate profiles API.
• To create blended profiles with control blended_type, use:
  • Manually constructing your own blended profile in the draft target group request or manage profiles request.
    • You can check the structure via the validate profiles API.
  • A maximum of one control blended type blended profile with up to five targets is allowed per target group.
  • The deletion of a blended profile with control blended type will not impact standalone profile conditions.
  • Targets within a control blended profile that reference a removed standalone profile condition are deleted. The control blended profile itself is deleted only when all of its targets are removed this way.
  • Targets within a control blended profile cannot have completes_goal greater than the target group's completes_goal.

Why blending may affect feasibility, cost, and field time

Blended profiles, while powerful, come with operational considerations.

:::caution 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. :::

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

:::caution Slower fieldwork More 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 contract	New contract
Called interlocked profiles	Now called blended profiles with enhancements
Required full expansion of all conditions	Manual or system expanded conditions
Used depends_on_questions	Now inferred from conditions in each target
Required ungrouped/grouped dicts	Now uses targets[] with explicit conditions
Quotas defined by quota_percentage	Now uses completes_goal_percentage or completes_goal
Could only be created via API logic	Users may also manually construct blended profiles

---

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 for interlock blended type
  • 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.