Skip to main content
Ambient Systems
Members

Planning and variance lifecycle: core vs platform

Planning and variance asks how actuals compare to the organization’s own plan—budget, forecast, or board case—and what explains the gap. The baseline is internal intent, not a peer (benchmarking-lifecycle.md).

FP&A close on work-cycles.md is the recurring rhythm (close calendar, board packs); this cycle is the plan-vs-actual analytical product built on the same catalog metrics.

Index: work-cycles.md.

End-to-end flow​

paid_platformambient_core_OSS

fpaWorkflow_and_core_metrics

ambient_calc_inputs

plan_versions_and_scenarios

variance_bridges

owners_and_commentary

Phase mapping​

1. Metric and calendar alignment​

  • Core — *.core.* close metrics (margins, growth, working capital, cash flow); segment operational metrics per analysis lens; fpaWorkflow strings describing typical FP&A use on each row in catalog YAML.
  • Platform — Fiscal calendar, plan version locking, mapping actuals period to plan period; metadata plan_version_id, optional scenario_id (base, upside, downside).

2. Plan storage and scenarios​

  • Core — No plan SSOT in ambient-core; plans are tenant data in the platform OLTP or lakehouse.
  • Platform — Versioned budgets and forecasts at catalog metric id granularity (or mapped account trees); scenario comparison UI.

3. Variance computation and decomposition​

  • Core — calc.expr and calc.inputs define how a KPI is built from components; platform uses the same graph for variance bridges (price vs volume vs mix) where inputs exist.
  • Platform — Variance waterfalls, commentary, accountability by department; not peer-ranked.

4. Close integration​

  • Core — Assurance and bridge products may explain why actuals moved before variance is finalized.
  • Platform — Tie variance cycle to close status, reforecast triggers, board deck export.

Distinct from benchmarking​

Benchmarking diagnoses performance versus external pace-setters and improvable structural gaps. Planning variance diagnoses performance versus yesterday’s plan. The same metric (for example operating_margin) can run through both cycles in one quarter with different metadata (peer_group_id vs plan_version_id).

Examples​

  • Any lens: Quarterly actual revenue_growth and ebitda_margin vs annual budget using industry.core.* metrics after catalog expansion.
  • Aviation network carrier org: Actual CASM/RASM vs operating plan using aviation.network_carrier.* definitions; variance on fuel and load factor inputs.
  • Retail org: Same-store sales growth actual vs plan using retail.operations.same_store_sales_growth.

Canonical source: ambient-core at docs/planning-variance-lifecycle.md (sync ref 05a4338).