Understanding the fielding assistant

The fielding assistant is an intelligent tool designed to optimize target group performance during fieldwork by leveraging machine learning and aggregate data from the Cint Exchange.

What is the fielding assistant?

The fielding assistant is a set of smart, automated tools offered by the Cint Exchange to streamline and optimize the process of collecting survey completes. Instead of requiring you to develop complex fielding algorithms or manual optimizations, the fielding assistant utilizes advanced machine learning on aggregate data across the entire Cint Exchange to intelligently manage your target groups.

By enabling fielding assistant features, you can enhance the efficiency, cost-effectiveness, and quality of your survey fieldwork. These features are not mutually exclusive and can be combined to achieve tailored optimization strategies.

Key features of the fielding assistant

The fielding assistant offers several powerful capabilities to control your survey's fieldwork:

1. Prevent overfilling (prevent_overfill, Boolean)

• Description: This feature helps ensure your quotas are filled precisely without exceeding your desired limit. It automatically converts your complete target into an equivalent prescreen quota, allowing for rapid filling while preventing unintended over-collection.
• Recommendation: This option is enabled by default in the CintX UI, and Cint strongly recommends enabling it on all API projects for optimal quota management.
• Example JSON:

"quota_overlay": {
  "prevent_overfill": true
}

2. Balanced fill (balance_fill, Boolean)

• Description: When enabled, the system intelligently paces your quotas across different demographic segments. This helps to prevent situations where you might quickly fill easy-to-reach segments, leaving you with only difficult or expensive demographic alignments to complete at the end of fieldwork. Since this slows down your total fill by the availability of your hardest to find demographic it is best used in attractively priced target groups in high volume markets where this can improve response quality quickly and easily without having to invest in more complicated interlocking profiling.
• Example JSON:

"quota_overlay": {
  "balance_fill": true
}

3. Optimize cost per interview (CPI) (pricing, Object)

• Availability: This feature is only available for dynamic pricing customers.
• Description: The CPI optimization feature uses your overall target group budget (or a defined maximum CPI) and starts fielding at a price that our algorithms believe will fill your survey by the specified end date. If traffic to your target group speeds up or slows down, the system will gradually and intelligently modify your CPI to attract the right number of respondents. It will never exceed your configured total budget or maximum CPI, whichever is operating. We recommend using a total budget as that means the system can "invest" the savings from cheaper completes in incentivizing hard to reach audiences, maximizing your fill.
• Configuration: You must define either your total_budget (for the entire survey) OR your maximum_cpi. A minimum_cpi can also be specified which will prevent the system bidding lower than that value.
• Example JSON (using total_budget):

"pricing": {
  "type": "dynamic",
  "total_budget": {
    "value": "5.00",
    "currency_code": "USD"
  },
  "minimum_cpi": {
    "value": "2.00",
    "currency_code": "USD"
  }
}

• Example JSON (using maximum_cpi):

"pricing": {
  "type": "dynamic",
  "maximum_cpi": {
    "value": "5.00",
    "currency_code": "USD"
  },
  "minimum_cpi": {
    "value": "2.00",
    "currency_code": "USD"
  }
}

4. Pace your completes (pacing, Object)

• Description: This feature allows you to schedule incremental quota increases for your target group over time. It is commonly used for long-running projects (trackers) or any survey that requires a steady, controlled rate of completes rather than a rapid fill.

note

While flexible, this feature counts respondents as whole numbers (integers). Extremely small increments (e.g., 0.5 respondents per hour) cannot be processed effectively by suppliers.

• Pacing options: These options determine how the system behaves if you change the completes goal while in field, or if quotas do not fill fully during a specific time slot (e.g., during slower traffic periods like weekends).
  • Linear: The system increments quotas evenly at regular intervals between the start and end dates. This distributes the available completes uniformly over time, regardless of actual fill rates.
  • Adaptive: The system recalculates increment amounts at the start of each interval. It adjusts dynamically based on current performance and the remaining goal, helping you "catch up" if fielding falls behind.
• increment_interval format (ISO 8601 Duration):
  • The duration must always start with P.
  • Days: add a number and the letter D (e.g., P1D for 1 day).
  • Time: start with PT, then add a number and unit (H for hours, M for minutes).
  • Multiple units can be combined (e.g., PT1H30M for 90 minutes, or simply PT90M).
  • Note: decimals are not permitted (e.g., use PT30M, not PT0.5H).
• Example JSON (adaptive pacing, 30-minute interval):

"pacing": {
  "type": "adaptive",
  "increment_interval": "PT30M"
}

• Example JSON (linear pacing, once a day):

"pacing": {
  "type": "linear",
  "increment_interval": "P1D"
}

5. Soft launch (soft_launch, Object)

• Description: This feature manages a controlled initial launch of a target group, allowing you to test survey performance with a smaller audience before a full rollout.

note

During a soft launch, any settings configured in the pacing module are ignored.

• Configuration:
  • end_at (string, required): The date and time when the soft launch ends. At this time, the target group will pause automatically, regardless of whether the filling goal was met.
  • filling_goal_percentage (integer, required): The percentage of your total completes goal you want to collect during soft launch (Value between 1% and 100%).
  • demographics_strictness_percentage (integer, required): This controls how strictly the soft launch demographics distribution should follow the distribution in the full launch.
    • 100: (the maximum) aims to fully match the demographics
    • 0: allows any distribution.
• Example JSON:

 "soft_launch": {
   "end_at": "2026-08-01T15:00:00.000Z",
   "filling_goal_percentage": 10,
   "demographics_strictness_percentage": 80
}