Churn – Giga Innovate

Official Metric Documentation

Created through

Metric Harmony Alignment Process

1

Metric Name & Summary

Metric Name:
Churn (Customer & Revenue)
Description:
The rate at which paying customers discontinue their subscriptions, and/or the portion of recurring revenue lost from those cancellations or downgrades.
2

Business Context & Purpose

Why We Track It:
  • Giga Innovate is a FinTech scale-up relying heavily on subscription-based revenue for real-estate analytics.
  • CFO (Fiona Caldwell) tracks monthly/quarterly churn to forecast revenue and reassure investors about retention.
  • Head of ML (Aurora Milton) needs consistent churn data to train predictive models, ensuring advanced analytics match the CFO's official perspective.
Primary Users:
Finance ML & Product Executive Team CEO CFO
3

High-Level Definition

3.1 Type of Churn

1
Customer (Logo) Churn

Percent of clients who fully cancel.

2
Revenue Churn

Percent of recurring revenue (MRR) lost due to cancellations or contract non-renewals.

3.2 Time Window

Official board-level churn is published monthly, though some advanced analytics might use rolling 30-day windows or a quarterly approach.

3.3 Inclusions / Exclusions

Inclusions:

Customers who fully cancel any real-estate analytics subscription and non-renewals at the end of an annual contract.

Exclusions:
  • Downgrades (treated as "contraction" not full churn)
  • ML "experimental" pilot users who were never paying
  • Suspended/paused accounts (unless status changes to fully canceled)
4

Detailed Formula & Calculation Steps

4.1 Plain-English Formulas

Customer (Logo) Churn Rate:
( Number of Paying Customers Lost in a Month / Number of Paying Customers at Month Start ) × 100%
Revenue Churn Rate:
( MRR Lost from Cancellations / MRR at Start of Month ) × 100%

4.2 Technical Logic/Examples

SQL Snippet (Giga Innovate's subscription DB): 

WITH start_mrr AS (
SELECT SUM(
CASE
  WHEN billing_cycle = 'annual' THEN (amount / 12)
  ELSE amount
END
) AS mrr_start
FROM sub_analytics
WHERE status = 'active'
AND period_label = 'start_of_month'
),
lost_mrr AS (
SELECT SUM(
CASE
  WHEN billing_cycle = 'annual' THEN (amount / 12)
  ELSE amount
END
) AS mrr_lost
FROM sub_analytics
WHERE status = 'canceled'
AND canceled_in_month = 'October'
)
SELECT
(lost_mrr.mrr_lost / start_mrr.mrr_start) * 100 AS revenue_churn_rate
FROM start_mrr, lost_mrr;

In ML, Aurora Milton's team feeds the same cancellation flags into churn prediction models to ensure consistency between finance and product analytics.

5

Data Sources & Systems

Sub_Analytics

Giga Innovate's main subscription table (fields include status, billing_cycle, amount, cancel_date).

Finance & ML Data Pipelines

Finance transforms data for monthly reports; ML/analytics use near-real-time feeds.

Update Frequency

Official churn rate is reported monthly; ML uses rolling 30-day windows.

6

Exceptions & Special Cases

Edge Cases to Be Aware Of

  • Downgrades vs. Churn:
    Downgrades are treated as "contraction" and are tracked separately.
  • Annual Deals:
    If an annual subscription ends mid-year and is not renewed, it is considered churn effective at the end of that contract month.
  • Paused Accounts:
    Accounts that are paused remain "active" until status changes to fully canceled.
7

Ownership & Maintenance

Owner
Fiona Caldwell (CFO)
Manages the official churn definition
Co-Owner
Aurora Milton (Head of ML)
Ensures ML churn predictions align with finance logic
Update Cadence:

Documentation is reviewed quarterly; official churn is published monthly.

Change Process:

Proposed changes must be approved by both the CFO and ML lead.

8

Revision History

Date Change Approved By
2024-01-05 Created official churn doc—excludes downgrades, includes annual plan non-renewals. Fiona Caldwell (CFO), Aurora Milton
2024-03-15 Added partial-month rule for cancellations and alignment with ML prediction pipeline. Fiona Caldwell, Aurora Milton
9

Disclaimer

This definition is for Giga Innovate's internal and investor-related reporting. It may not mirror GAAP/IFRS revenue recognition. Data relies on our "Sub_Analytics" pipeline. For advanced ML triggers, partial-day or partial-month changes may appear in near-real-time, but official churn stands monthly under CFO approval.

Why This Addresses Giga Innovate's ML-Finance Tug-of-War

By clarifying how mid-month cancellations, annual deals, and downgrades are treated, Giga Innovate's CFO and ML teams no longer produce contradictory churn figures. CFO Fiona Caldwell has final sign-off for board reporting, while Aurora Milton's ML pipeline references the exact same cancellation flags. This alignment ensures product analytics match finance's official churn definition—removing previous confusion over partial cancellations.