Effective prompting is key to getting the most out of Buster. By understanding how Buster interprets requests, you can ensure you receive accurate, relevant, and helpful responses.

​
Be Direct and Clear for Ad-Hoc Questions

Buster excels at answering specific, unambiguous questions about your data. When you have a clear question in mind, state it directly.

Good Examples:

  • “How many customers do we have?”
  • “What was the total revenue last month?”
  • “Show me monthly social media reactions for the last 12 months.”

These types of questions are targeted, allowing Buster to understand the exact information you need.

Avoid Vague Requests:

Try to avoid overly broad or ambiguous terms.

  • Instead of: “Sales” Try: “What were the total sales figures for Q2?” or “Show me the sales trend over the last year.”
  • Instead of: “Who’s the best?” Try: “Which salesperson had the highest revenue last quarter?” or “Which product category generated the most profit?”
  • Instead of: “Find order 12345” Try: “Find the order with internal order ID 12345” or “Find the order with Shopify order ID 12345” (Especially important if multiple ID types exist).

While Buster can sometimes handle nuanced requests if the nuances are well-defined within your data or previous interactions, directness and clarity generally yield the best results.

​
Provide Context for Exploratory Analysis

When using Buster for more investigative or exploratory tasks, providing context is crucial. Buster needs direction to perform meaningful analysis aligned with your business goals.

Why Context Matters:

Without context about your business objectives, Key Performance Indicators (KPIs), relevant timelines, specific goals, or topics of interest, Buster might generate analyses based on metrics that aren’t relevant to you.

Example Scenario: Investigating a Sales Drop

  • Less Effective: “Figure out why sales dropped.” (Buster will make assumptions about what metrics to investigate)
  • More Effective: “Figure out why sales dropped last quarter. Did it correlate with our recent website changes? Check performance in the Northeast region compared to others. Our main KPI is conversion rate.”

The more context you provide (hypotheses, relevant events, business specifics), the better Buster can tailor its analysis to your needs.

​
Specify Visualization Styling

You can instruct Buster on how to style the visualizations it creates. If you have specific preferences for colors, chart types, or other visual elements, include those details in your prompt.

Example:

  • “Show monthly revenue for the past year as a line chart, and make the line green.”
  • “Update the ‘Sales Performance’ dashboard to use a dark theme.”

Being specific about styling ensures the output matches your requirements.

​
Understand Buster’s Limitations

Currently, Buster’s primary capabilities involve:

  • Building data visualizations
  • Creating and modifying dashboards
  • Answering questions based on the connected data sources

Buster is not currently capable of:

  • Performing web searches
  • Conducting advanced statistical analysis (e.g., correlation, regression)
  • Forecasting future trends

Keeping these limitations in mind will help you set appropriate expectations for your requests.