Every scorecard in FiveLumens requires a scoring model. Scoring models determine how results are calculated and displayed across both manual and AI evaluations. There are four available models: Weighted, Equal, Selective, and Audit. Weighted and Equal models assign points based on Yes, No, or N/A responses, with Weighted allowing custom point values by criterion and Equal giving each item the same weight. Selective uses a “good, better, best” approach where evaluators choose N/A, 0, 1, or 2 points, and Audit removes scoring entirely, allowing form completion for data purposes only.
Both manual and AI evaluations use the same scorecard structure. Manual evaluations are completed by analysts, while AI evaluations apply AI instructions to automatically grade transcripts. AI scorecards support the Weighted, Equal, and Audit models. The Selective model is only available for manual evaluations.
Accounts can create and use an unlimited number of scorecards and can mix manual and AI types as needed. The following articles explain each scoring model in more detail.
The Selective Scoring Model is available only for manual scorecards and is used when a simple Yes or No answer does not fully capture performance. This model allows evaluators to choose between N/A, 0, 1, or 2 points, providing a more flexible range of assessment for qualitative behaviors.
Many organizations use this model to represent levels of performance, such as None (not good), Partial (good), and Full (best). It is especially useful for criteria where effectiveness varies by degree or effort rather than a strict presence or absence.
Example
Criteria: Did the agent use multiple open-ended probing questions to understand the customer’s needs?
| Score | Description |
|---|---|
| Full (2) | The agent used multiple open-ended questions to discover the customer’s needs. |
| Partial (1) | The agent asked superficial questions and did not probe enough to fully understand the customer’s needs. |
| None (0) | The agent did not attempt to find out the customer’s needs. |
This model provides greater nuance in evaluations and helps analysts identify areas of partial success or progress, not just compliance or noncompliance.
Both the Weighted and Equal scoring models use a Yes, No, or N/A format to assess criteria. The difference between them is how points are distributed across the scorecard.
Weighted Scoring allows each question to carry a different level of importance. Administrators can assign specific point values to each criterion so higher-impact behaviors or compliance items have more influence on the final score. This model is best used when some actions are more critical to the customer experience, compliance, or brand standards than others.
Equal Scoring gives every question the same value. Each criterion contributes equally to the total score, regardless of its topic or category. This model provides a simple, consistent approach that works well for general evaluations or when all standards are considered equally important.
Example
If a scorecard includes five questions and one is deemed more important, the Weighted model might assign it 30% of the total score while the remaining four share the other 70%. In an Equal model, each question would simply count for 20% of the total.
Use Weighted when you want to emphasize key criteria and Equal when you prefer a straightforward, balanced scoring approach.
The Audit Scoring Model functions like the Equal model in structure, using Yes, No, or N/A responses for each criterion. However, the key difference is that Audit evaluations do not produce a score and are not attributed to an agent’s performance.
This model is designed for gathering insights rather than measuring individual results. It allows you to use a scorecard format to collect consistent data across specific call types or scenarios without affecting performance metrics.
For example, an organization might run an audit on closed sales calls to identify which sales techniques are most often used in successful conversions. Conversely, another audit could focus on non-sales calls to uncover missed opportunities or areas for skill development.
Audits are ideal for research, process analysis, and identifying patterns or training opportunities across large sets of interactions without influencing QA scores or coaching outcomes.
Choosing the right scoring model helps ensure your evaluations align with your organization’s goals. Whether you need detailed behavioral insights, consistent compliance checks, or data-only analysis, each model offers a specific purpose. Weighted and Equal provide flexibility in measurement, Selective allows greater nuance for manual evaluations, and Audit supports broader data gathering without impacting scores.
By selecting the right model for each scorecard, you create a balanced approach to quality assurance—one that supports accuracy, clarity, and meaningful coaching outcomes.