name: BigML FinOps
description: Financial operations guidance for managing BigML API costs
url: https://bigml.com/pricing

cost_drivers:
  - name: Parallel Task Capacity
    description: >
      The primary cost driver is the number of parallel tasks your subscription allows.
      Free tier allows 2 parallel tasks; Prime tier scales based on configured capacity.
    optimization: >
      Batch workloads during off-peak hours and use sequential task scheduling
      to minimize the number of parallel tasks required.

  - name: Dataset Size
    description: >
      Dataset size affects processing time and resource consumption. Free tier
      caps at 16 MB per task.
    optimization: >
      Pre-process and filter data before uploading to minimize dataset sizes.
      Use BigML's Flatline transformation language for server-side data reduction.

  - name: Private Deployment vs Cloud
    description: >
      Private deployments have fixed annual costs starting at $10,000/year (Lite).
      Cloud (Prime) pricing scales with usage.
    optimization: >
      Evaluate break-even point: if cloud costs exceed $10,000/year, consider
      Lite private deployment.

cost_saving_strategies:
  - name: Annual Billing Discount
    description: Save 30% by paying annually instead of monthly for Prime subscriptions
    savings: 30%

  - name: Quarterly Billing Discount
    description: Save 15% by paying quarterly for Prime subscriptions
    savings: 15%

  - name: Academic / NGO Discount
    description: Students, researchers, and NGO employees receive 50% discount on Prime
    savings: 50%
    eligibility: Students, academic researchers, NGO employees

  - name: Batch Predictions
    description: >
      Use batch prediction resources instead of individual predictions to process
      large volumes more efficiently.
    recommendation: >
      For bulk scoring, create a batch prediction resource from a model and test
      dataset rather than making individual prediction API calls.

monitoring:
  - method: api.get_tasks_status()
    description: Monitor task consumption to avoid unexpected queuing delays
  - url: https://status.bigml.com/
    description: Monitor platform availability to plan workload scheduling

deployment_options:
  - type: Cloud (SaaS)
    billing: Subscription (monthly/quarterly/yearly)
    best_for: Variable workloads, small teams, getting started

  - type: BigML Lite (Private)
    billing: $10,000/year or $1,000/month
    best_for: Organizations needing data residency or steady high-volume workloads

  - type: Bronze Enterprise (Private)
    billing: $45,000/year + $10,000 setup
    best_for: Large enterprises with dedicated infrastructure requirements