Sentinel
      Back to home
      1. Help Center
      2. Sentinel

      Effective Queries in Sentinel: Best Practices for Clarity and Precision

      This article is relevant for: FD Live Professional

      Sentinel is already very capable of translating general questions into specific data queries and producing the sought lists of customers, tickets, or events. Working with Sentinel is like working with someone very skilled at querying data, but with only an average level of domain knowledge. In other words, precision matters: the more clearly you phrase your questions, the better Sentinel’s results.

      Ask Your Questions Clearly and Precisely

      If your question is ambiguous, Sentinel may return incomplete, incorrect, or no results at all. Based on our observations, here are some key recommendations:

      1. Leave No Room for Interpretation

      When forming a question, be as explicit as possible. For instance, if you want a list of all customers who purchased tickets between March and June 2024 but did not purchase any tickets afterward, a vague query like:

      “Give me a list of all customers who bought tickets in Spring 2024 but didn’t buy a ticket afterwards.”

      can lead to confusion because:

      • “Tickets” could mean any number of tickets (1, 2, 3, …).
      • A single ticket could also be interpreted differently than multiple tickets.
      • “Spring 2024” is not a clearly defined timeframe (it varies geographically and contextually).

      A better approach is:

      “Give me a list of all customers who bought one or more tickets between March and June 2024 but didn’t buy one or more tickets afterwards.”

      2. Be Specific with Data References

      • Name events or timeframes directly. Instead of using terms like “festival” or “season,” specify the exact event name or the explicit dates, since a “season” may differ widely across organizations.
      • Start with the data category. For example:
        • “Give me all customers…”
        • “Give me all events…”
        • “Give me all transactions…”
      • Use the exact names of data points as defined in Future Demand’s database (e.g., venue, artist name, ticket type). See Usable data points for Sentinel for a full list.
      • When querying custom properties, refer to Custom properties queries in Sentinel.

      3. Verify Correct Naming Conventions

      • Ensure the values you’re using match what is actually recorded in Future Demand (e.g., venue is “Carnegie Hall,” artist name is “Maria Callas,” composer is “Mozart,” etc.).
      • If you’re unsure, check by clicking on an event in Lookout and selecting “edit event details” or using the filtering function to confirm the correct spelling.
      • Another good strategy is to start with a simple question to see which data points appear in the results, then refine your query accordingly.

      4. Pay Attention to Ticket Categories

      • If your question involves discounts (e.g., who did or did not receive discounted tickets), always reference the exact ticket categories. See Which ticket categories does Future Demand use? for details.
      • For example:

        “Give me all clients who purchased more than 4 tickets with ticket type regular in 2024.”
        implies you’re looking for clients who bought more than 4 non-discounted (regular) tickets, potentially indicating they might be interested in a subscription offer.

      5. Double-Check Spelling

      • Even small typos can yield no results or incorrect results. Always review your queries for spelling errors before sending them to Sentinel.