specificationVersion: "1.0"
id: serper-finops
name: Serper FinOps Framework
description: >-
  FinOps analysis for Serper's Google Search API covering cost allocation,
  optimization, and forecasting aligned to the FinOps Foundation FOCUS
  specification. Serper's credit-based, usage-driven model makes it
  straightforward to attribute costs to specific teams or workloads and
  to forecast spend from query volume projections.
url: https://serper.dev

framework:
  standard: FinOps FOCUS
  version: "1.0"

costAllocation:
  model: credit-based
  billingUnit: credit
  description: >-
    Each successful API query deducts one credit from the account balance.
    Requesting 100 results per page deducts two credits. Credits are purchased
    in advance in fixed packages; unused credits expire after 6 months.
  taggingRecommendations:
    - Use separate API keys per team, product, or environment to isolate consumption.
    - Label API keys with cost-center or project metadata for chargebacks.
    - Monitor credit consumption via the Serper dashboard or API usage endpoint.

pricingDimensions:
  - dimension: Query Volume
    description: Primary cost driver; all charges scale linearly with successful query count.
    unitCost:
      freeTrial: 0.00
      starter: 1.00
      standard: 0.75
      scale: 0.50
      ultimate: 0.30
    unit: per 1,000 queries
    currency: USD

  - dimension: Results Per Page
    description: >-
      Requesting 100 results (num=100) doubles the per-query credit cost
      compared to the default 10-result page.
    multiplier: 2.0
    defaultResultsPerPage: 10
    maxResultsPerPage: 100

  - dimension: Credit Package Size
    description: Larger credit bundles yield lower per-unit costs.
    packages:
      - credits: 2500
        price: 0.00
        type: free-trial
      - credits: 50000
        price: 50.00
        costPer1k: 1.00
      - credits: 500000
        price: 375.00
        costPer1k: 0.75
      - credits: 2500000
        price: 1250.00
        costPer1k: 0.50
      - credits: 12500000
        price: 3750.00
        costPer1k: 0.30

optimization:
  recommendations:
    - name: Right-size credit packages
      description: >-
        Purchase credit packages aligned to 6-month projected query volume to
        avoid over-provisioning (wasted expiring credits) or under-provisioning
        (higher per-unit cost from multiple Starter purchases).
      savingsPotential: high

    - name: Cache search results
      description: >-
        Implement a short-lived cache (e.g., 1-5 minutes for news; 24 hours for
        evergreen queries) to eliminate duplicate credits consumed on identical
        queries fired within a burst window.
      savingsPotential: high

    - name: Minimize result page size
      description: >-
        Use the default 10-result page (num=10) unless larger result sets are
        required. Requesting num=100 doubles credit consumption for each query.
      savingsPotential: medium

    - name: Consolidate to higher-volume tiers
      description: >-
        When total monthly consumption consistently exceeds 50,000 queries,
        upgrade to the next credit tier to reduce the effective cost per query.
      savingsPotential: medium

    - name: Monitor and alert on credit burn rate
      description: >-
        Set alerts when remaining credits fall below a threshold (e.g., 10% of
        purchased bundle) to avoid service interruption and plan re-purchases.
      savingsPotential: low

forecasting:
  approach: volume-based
  inputs:
    - averageQueriesPerDay
    - resultsPerPageRatio
    - planTier
  formula: >-
    monthly_cost = (averageQueriesPerDay * 30 * costPer1kQueries / 1000)
    * resultsMultiplier
  notes: >-
    Credits are valid for 6 months. Annual forecasting should account for
    query growth trends and seasonal search volume peaks. Plan re-purchase
    timing should be aligned to credit expiry windows.

reporting:
  keyMetrics:
    - name: Cost Per Query
      formula: "plan_price / credits_purchased"
      unit: USD
    - name: Credit Utilization Rate
      formula: "credits_used / credits_purchased"
      unit: percentage
    - name: Effective Monthly Spend
      formula: "queries_per_month / 1000 * cost_per_1k"
      unit: USD
    - name: Credits Remaining
      description: Monitor via Serper dashboard to prevent service interruption.