Cubes

A cube is a pre-modeled, multi-dimensional dataset designed for fast analytics. Cubes are the layer underneath dashboards and reports — they're what makes analytics fast, consistent, and self-service.

Why cubes (instead of querying raw tables)

Cubes solve three problems:

1. Consistency — "bookings" means the same thing in every dashboard.
2. Speed — pre-aggregated, sub-second response on millions of records.
3. Self-service — business users can slice and dice without writing queries.

The four built-in cubes

💰 Sales Cube

The cube that powers all "how much did we book?" questions.

Measures:

• Bookings (closed-won $)
• Deal count
• Average deal size
• Win rate
• Sales cycle (days)
• Average discount %

Dimensions:

• Account (and account tier, industry, size)
• Owner (and team, region)
• Product (and category, family)
• Period (day, week, month, quarter, year)
• Won/Lost reason
• Lead source

Sample questions:

• "Bookings by region by quarter."
• "Win rate by lead source over the last year."
• "Average deal size by product family."

📊 Pipeline Cube

For "what's in the funnel?" questions. A snapshot view + a time-series view.

Measures:

• Pipeline value ($)
• Pipeline count
• Weighted pipeline (value × probability)
• Days in stage
• Days in pipeline

Dimensions:

• Stage
• Owner (and team)
• Source
• Product
• Probability bucket (commit / best case / pipeline)
• Snapshot date (for trend analysis)

Sample questions:

• "Pipeline by stage today vs 30 days ago."
• "Average days-in-stage for the Proposal stage."
• "Coverage ratio (pipeline ÷ quota) by team."

🎧 Service Cube

The cube behind every case-related dashboard and SLA report.

Measures:

• Open case count
• New case count
• Resolved case count
• First-response time (minutes)
• Resolution time (hours)
• SLA met % (first response and resolution)
• CSAT score

Dimensions:

• Status, priority, category, origin
• Agent (and team)
• Account (and tier)
• Product
• Period

Sample questions:

• "SLA met % by priority by month."
• "Cases by category by product — which products generate the most support load?"
• "Reopened cases by agent — quality signal."

📣 Marketing Cube

Powers campaign and lead-source ROI analysis.

Measures:

• Members enrolled
• Members responded
• Conversion count (member → opportunity)
• Sourced revenue (closed-won where campaign was first touch)
• Influenced revenue (closed-won where campaign was any touch)
• Campaign spend
• Cost per lead, cost per opportunity, ROI %

Dimensions:

• Campaign (and type)
• Channel (email, webinar, event, etc.)
• Period
• Lead source
• Persona (job role)

Sample questions:

• "ROI by campaign type."
• "Cost per opportunity by channel."
• "Conversion rate by persona."

How users interact with cubes

The cube UI lets anyone:

1. Pick a cube.
2. Drag dimensions to rows and columns.
3. Pick measures to display.
4. Filter to narrow the view.
5. Pivot instantly to see the data differently.
6. Chart the result as bar, line, area, heatmap, or table.
7. Save the view as a report.
8. Pin the view to a dashboard.

No SQL, no formulas — just drag and drop.

How cubes stay fresh

• Incremental refresh every few minutes for hot data (today's activity).
• Full refresh nightly.
• Snapshots for time-series analysis (e.g., daily pipeline snapshots that power trend charts).

Refresh status is visible in the admin console.

AI Copilot and cubes

The AI Copilot reads directly from cubes — when you ask "What's our win rate by region?", it's the Sales Cube responding. This means:

• AI answers are consistent with dashboards and reports (same source of truth).
• AI can explain the answer by referencing the cube measure and dimension.
• Admins can audit AI queries through the cube access log.

Custom cubes

Admins and developers can build new cubes for domain-specific analysis:

• A Subscriptions Cube for SaaS revenue (ARR, NRR, churn).
• A Renewal Cube for renewal pipeline visibility.
• A Partner Cube if you sell through channels.

See Customization › Extending Objects for how to define new cubes, dimensions, and measures.

Tips for analysts

• ✅ Start every new analysis from a cube — don't build dashboards on raw queries.
• ✅ Use snapshot dimensions for trend questions ("How did this metric change?").
• ✅ Save shared views that the team can subscribe to.

Tips for admins

• ✅ When a measure is computed inconsistently across reports, promote it to a cube measure — single source of truth.
• ✅ Monitor cube refresh times — if they slow down, prune unused dimensions.
• ✅ Document each cube's measures and dimensions for end users — discoverability is the #1 adoption blocker.