Fix bugs, DRY up code, add monthly/daily summaries, and optimize dashboard (~176 → 38 queries)

[rameerez]

Summary

Full review of the profitable gem against the Pay gem v11 source and LicenseSeat production schema. Fixes bugs, consolidates duplicated logic, and reduces dashboard query count from ~176 to 38.

Bug Fixes

• calculate_new_mrr_in_period too permissive: Was using where.not(status: EXCLUDED_STATUSES + CHURNED_STATUSES) which let through incomplete, incomplete_expired, and unpaid subscriptions. Now uses where(status: 'active') to match MrrCalculator's logic — only active subscriptions contribute MRR.
• calculate_new_subscribers inconsistent filtering: The public method wasn't filtering excluded statuses (trialing, paused), but the _in_period version was. Now delegates to calculate_new_subscribers_in_period for consistent behavior.
• Constants inaccessible from MrrCalculator: EXCLUDED_STATUSES and CHURNED_STATUSES were defined inside class << self, making them unreachable as Profitable::EXCLUDED_STATUSES. Moved to module level.
• MrrCalculator hardcoded status list: Replaced ['trialing', 'paused'] with the shared Profitable::EXCLUDED_STATUSES constant.

DRY Consolidation

Five period-based methods now delegate to their _in_period counterparts instead of duplicating the query logic:

• calculate_churn → calculate_churn_rate_for_period
• calculate_churned_customers → calculate_churned_subscribers_in_period
• calculate_churned_mrr → calculate_churned_mrr_in_period
• calculate_new_mrr → calculate_new_mrr_in_period
• calculate_new_subscribers → calculate_new_subscribers_in_period

Added subscriptions_with_processor helper to eliminate repeated .includes(:customer).select('...customer_processor').joins(:customer) chains.

Dashboard Performance (~176 → 38 queries)

Section	Before	After
Monthly summary (12 months)	72 queries (6/month × 12)	5 queries (bulk load + Ruby grouping)
Daily summary (30 days)	60 queries (2/day × 30)	2 queries (bulk load + Ruby grouping)
Period cards (3 periods)	27 queries (9/period × 3)	18 queries (6/period via period_data)
Top cards + header	17 queries	13 queries (+ 2 cached by Rails)
Total	~176	38 (+ 2 cached)

How:

• Batched monthly_summary: 5 bulk queries load all subscription data for the full 12-month range, then groups by month in Ruby. No DB-specific SQL needed.
• Batched daily_summary: Same pattern — 2 bulk queries for the full 30-day range.
• New period_data method: Computes all 7 period metrics in one pass, reusing intermediate values (e.g., churned_count used for both the churned customers display and churn rate calculation; new_mrr/churned_mrr used for display and mrr_growth).
• Controller precomputation: Moved all Profitable.* calls from the view into the controller as instance variables.

Verified on LicenseSeat production:

Completed 200 OK in 71ms (Views: 2.1ms | ActiveRecord: 23.4ms (38 queries, 2 cached))

New Public API

• Profitable.period_data(in_the_last:) — returns a hash of NumericResult values: new_customers, churned_customers, churn, new_mrr, churned_mrr, mrr_growth, revenue

Test plan

• bundle exec rake test — 249 tests, 382 assertions, 0 failures, 0 errors
• Coverage: 94.64% line, 90.77% branch
• Production sync verified: diff between lib/profitable.rb and test/test_helper.rb shows only a comment block difference
• Tested on LicenseSeat with real Stripe data — 38 queries, 71ms

🤖 Generated with Claude Code

[claude]

---

Code review

No issues found. Checked for bugs and CLAUDE.md compliance.

---

[rameerez]

New Features: Monthly & Daily Summary Tables

This PR also adds two new dashboard sections and their corresponding public API methods:

Monthly Summary (last 12 months)

Profitable.monthly_summary(months: 12) returns an array of hashes, one per month (oldest first), each containing:

Key	Type	Description
:month	String	"2026-01" format
:month_date	Time	Beginning of month
:new_subscribers	Integer	Distinct customers with new non-trialing/paused subscriptions
:churned_subscribers	Integer	Distinct customers whose subscriptions ended (canceled/ended)
:net_subscribers	Integer	new - churned
:new_mrr	Integer	MRR from new active subscriptions (cents)
:churned_mrr	Integer	MRR lost from churned subscriptions (cents)
:net_mrr	Integer	new_mrr - churned_mrr
:churn_rate	Float	Churned / subscribers at start of month (%)

Daily Summary (last 30 days)

Profitable.daily_summary(days: 30) returns an array of hashes, one per day (oldest first):

Key	Type	Description
:date	Date	The date
:new_subscribers	Integer	New non-trialing/paused subscribers that day
:churned_subscribers	Integer	Subscribers who churned that day

Dashboard UI

Both summaries render as tables in the dashboard view with:

• Color-coded positive (green) and negative (red) values
• MRR amounts shown alongside subscriber counts
• Newest entries at the top (.reverse in the view)

These are the "Monthly summary" and "Daily summary" sections visible between the top metric cards and the period breakdowns.