[mnapoli]

Summary

Day-N retention doesn't move once day N is over. If we bring back people who signed up weeks ago (an email about what's new, say), the retention charts can't show it.

"Daily active users" now shows where each day's active people come from, and who left:

• Above the axis: regulars (blue), then people back after 7+ days away (light blue), then new sign-ups (amber, as before). A campaign that brings people back shows on the day it lands, whenever those people signed up.
• Below the axis: people lost, counted the day after their last visit when they don't come back in the 7 following days. A Product Hunt spike's departures show the next day, next to their cause, not a week later.

Changes

• BackAndLostQuery counts both from user_active_days, with no new stored data. Bots are left out, like the other metric queries.
  • back: signed up more than 7 days ago and not seen in the 7 days before.
  • lost: seen the day before and not back in the 7 following days.
• ActiveUserCountsQuery splits "returning users" into regulars and back, so bar totals still match active_user_counts. It also returns lost.
• Lost isn't final until its 7 days are over. Until then it counts the people not seen since, a number that can only go down. Those bars are striped and the tooltip says "so far". Today has no lost yet.
• "Back" and "lost" are approximate until Nov 6: before Oct 30, active days were backfilled from actions, which miss people who only read. The tooltip says so. The bars aren't marked.
• DailyBars can hang a value below the axis, on the same scale. The other charts are unchanged.

Visual evidence

Local app with synthetic data: about 4,800 sign-ups since Oct 15, a launch spike on Oct 20, and a pretend "we're back" email on Oct 28 that brings some dormant people back. Today is Nov 1, range 30d.

	Before (production)	After
Light, 1440
Dark, 1440
Light, 390
Dark, 390
Whole page, light, 1440

Tooltip on Oct 28:

Device verification

There's no new mobile interaction, only a chart. I checked 390px in Chromium emulation, not on a device: the legend wraps and the bars below the axis stay inside the panel.

Limitations

• "Back" doesn't say why people came back. To measure a campaign, send it to most people and hold back a random share, then compare how many of each group come back within 7 days.
• Someone with the Mac app running counts as active every day, so they never show as lost.
• The shadow from the day before is still drawn, but it's no longer in the legend.

Summary by CodeRabbit

• New Features
  • Daily active-user metrics now distinguish regular users from users returning after at least seven days away, and show users lost.
  • Charts display losses below the axis, with striped bars indicating counts that are not yet complete.
  • Updated chart legends and tooltips explain these categories and whether loss counts are settled.

[coderabbitai]

No actionable comments were generated in the recent review.

Recent review info

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: 1f2b844f-cc4e-4842-be45-a816396863a3

Commits

Reviewing files that changed from the base of the PR and between 7f9b1db and 805df03.

Files selected for processing: 7

• app/Queries/ActiveUserCountsQuery.php
• app/Queries/BackAndLostQuery.php
• resources/js/pages/operations/metrics.tsx
• resources/js/types/operations.ts
• tests/Feature/Controllers/ProductMetricsControllerTest.php
• tests/Unit/Queries/ActiveUserCountsQueryTest.php
• tests/Unit/Queries/BackAndLostQueryTest.php

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 9 remain after this review.

Walkthrough

Active-user metrics now separate regular users from users returning after seven days away. They add daily lost-user counts, completeness and backfill indicators, and chart segments for users lost below the axis.

Changes: Active-user metrics

Layer / File(s)	Summary
Compute seven-day returns and losses app/Queries/BackAndLostQuery.php, tests/Unit/Queries/BackAndLostQueryTest.php	BackAndLostQuery calculates daily returns and losses using a seven-day gap. It marks loss counts as complete when the seven-day period has elapsed and flags dates whose lookback predates recorded activity. Tests cover exclusions, date boundaries, backfill status, and ranges after today.
Add return and loss fields to active-user results app/Queries/ActiveUserCountsQuery.php, resources/js/types/operations.ts, tests/Unit/Queries/ActiveUserCountsQueryTest.php, tests/Feature/Controllers/ProductMetricsControllerTest.php	ActiveUserCountsQuery replaces returning with regular and adds back, lost, lost_complete, and backfilled. The result type and test expectations use the updated fields.
Display return and loss categories resources/js/pages/operations/metrics.tsx	Daily bars support below-axis values. The active-user chart shows regular, back, and lost users, and reports whether loss counts are complete or approximate.

Estimated code review effort: 3 (Moderate) | ~25 minutes

Change: Feature

Sequence Diagram(s)

sequenceDiagram
  participant ActiveUserCountsQuery
  participant ActivationByDayQuery
  participant BackAndLostQuery
  participant OperationsMetricsChart
  ActiveUserCountsQuery->>ActivationByDayQuery: retrieve daily activation counts
  ActiveUserCountsQuery->>BackAndLostQuery: retrieve daily back and lost counts
  BackAndLostQuery-->>ActiveUserCountsQuery: return counts and status flags
  ActiveUserCountsQuery-->>OperationsMetricsChart: provide active-user results
  OperationsMetricsChart->>OperationsMetricsChart: render back and lost segments

Loading

Possibly related PRs

• dev-null-sh/dev-null.sh#1324: Adds and records user_active_days, which BackAndLostQuery reads for the seven-day metrics.
• dev-null-sh/dev-null.sh#1337: Calculates signup-cohort retention from user_active_days and uses the same historical completeness assumption.

Suggested reviewers: xheaven, nunomaduro

Merge Risk: Minimal · up to 805df

The reviewed change is mergeable after normal checks; no concrete issue requiring a fix was established.

Security architecture risk: Low · up to 805df

The change remains within the administrator-only metrics dashboard and returns aggregate counts rather than individual activity records. No new access-control bypass or persistent state change was identified. The main remaining uncertainty is the additional database workload, especially for historical ranges.

Retained concerns
No architecture-level concerns identified.

Security review details

Security Blast Radius

• inferred — The inspected change affects administrator-visible aggregate metrics and database read workload. It does not add an observed path for ordinary or unauthenticated users to retrieve individual activity data.

Trust Boundaries and Controls

• observed — The metrics route remains inside authentication middleware, and its dedicated request authorizes administrators only. The boundary comparison found these controls unchanged from the specified PR base. The new SQL uses bound parameters for dates, gap values, and bot exclusions.

Pre-merge checks: 5

Check name	Status	Explanation
Description Check		Check skipped - CodeRabbit’s high-level summary is enabled.
Title check		The title clearly and concisely describes the primary change: splitting daily active users into regular users and users returning after at least seven days away.
Docstring Coverage		Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentat…
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Docstring Coverage

Explanation

Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentation-only edits and deletions within surviving functions are included; unchanged diff context and fully deleted functions are excluded. Analyzed 2 functions across 7 files.

Finishing Touches

Generate docstrings

• Commit to this branch
• Create a new PR

Generate unit tests (beta)

• Commit to this branch
• Create a new PR

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts