Zoë analyzes your data using four building blocks: metrics, slices, filters, and time periods. Understanding these will help you ask better questions and get more precise answers.
You can think of it this way:
Metrics are the numbers you want to measure.
Slices are the ways you break those numbers down.
Filters limit which rows are included.
Time periods control the date range or time grouping.
Ask Zoë for what you want using this vocabulary and she'll build the right query — no need to click through menus by hand.
Metrics
Metrics are business measures such as revenue, orders, customers, conversion rate, or average order value.
Use metrics when you want to answer "how much," "how many," or "what rate."
Examples:
Total revenue
Number of orders
Active customers
Conversion rate
Slices
Slices break a metric into categories.
Use slices when you want to answer "by what."
Examples:
Revenue by product
Orders by region
Customers by acquisition channel
Conversion rate by campaign
Filters
Filters limit the data included in the result.
Examples:
Only orders from the last 30 days
Only customers in the United States
Only paid campaigns
Only enterprise accounts
Time periods
Time periods control the dates included in an analysis and how time-based results are grouped.
Examples:
Last 7 days
This quarter
Month over month
Weekly trend
Putting it together
Combine metrics, slices, filters, and time periods in one question and Zoë will build the right query. For example: "Show me revenue by product, filtered to paid campaigns, for this quarter."
Best practices
Start with one metric, then add slices and filters only as needed.
Use business terms rather than database terms when possible — Zoë understands your language, not your schema.
If a result looks wrong, check the metric definition, date range, filters, and slice before assuming the data is incorrect. Ask Zoë to explain how she calculated it.
If you are not sure which metric or slice to use, ask Zoë to explain the available options, or ask a workspace admin/data owner to confirm the semantic layer setup.