Create Custom Scorecard Criteria

This guide provides a framework for writing clear, well-structured criteria that ensure accurate grading of your Revenue AI Scorecards. Following these best practices helps you avoid vague, unverifiable, or overly broad criteria that can lead to inconsistent scoring results.

Guiding Principles for Writing Effective Criteria

Revenue Scorecards evaluate each criterion with a simple Yes or No response. To ensure accurate and consistent scoring, each criterion must be:

1. Objective and Measurable – Focus on observable behavior or language in the transcript.
2. Specific and Clear – Use precise language; avoid generalizations.
3. Single-Focus – Test for one behavior at a time. Do not bundle multiple concepts into one criterion.
4. AI-Friendly – Structure the criterion so the AI can easily detect it in conversation
5. Free of Subjectivity – Avoid fuzzy language like “Did the rep build rapport?”. These phrases are too open-ended and should be restructured into specific actions a rep might take to build rapport.

Common Pitfalls to Avoid

• Break down broad behaviors: Instead of “Did the rep handle objections well?”, ask “Did the rep acknowledge and restate the customer’s objection before responding?”
• Focus on observable signals: Instead of “Did the rep prepare for the call?”, look for evidence like “Did the rep reference pre-call research?”
• Replace vague verbs: Words like “engage” or “connect” should be translated into specific actions (e.g., “Did the rep ask open-ended questions?”).

Examples of Well-Written vs. Poorly-Written Criteria

Identify a Category for Each Criterion

To ensure scorecard scores can be rolled up and aggregated via our scorecard analytics, assign a category to each criterion.

• Categories should group similar criteria to allow for structured reporting and analysis.
• Common categories might include Introduction, Qualification, Objection Handling, Call Control, Solution Pitch, Closing, etc.
• If a new category is needed, ensure it aligns with broader analytical goals and is not too niche.
• Poorly chosen or inconsistent categories can negatively impact the usability of analytics and reporting.

Formatting Guidelines for Custom Criteria

When submitting your criteria for use in Revenue AI Scorecards, you’ll need to format them as a CSV file with the following three columns: Category, Criterion Name, and Criterion.

Field Definitions

• Category: The group this criterion belongs to (e.g., Introduction, Qualification, Objection Handling). Categories help organize your data and enable roll-up reporting in analytics.
• Criterion Name: A short, human-readable label that summarizes the criterion’s purpose. It should be unique and clear enough to identify at a glance.
• Criterion: The full, closed-ended question being scored. It should be answerable with a simple “Yes” or “No” and follow the best practices outlined above. It should also be no more than 512 characters long.

How to Format Your CSV

Each row should represent one scoring criterion. Make sure your file includes a header row with the correct column titles: Category, Criterion Name, and Criterion.

Example CSV Format

Submit Your Criteria

Once your CSV file is finalized, send it to our Support team for review and upload. If you need help structuring your file or refining your criteria, we’re here to help.