→Navigara · Product·Engineering Performance

Engineering performance, measured and scaled.

We map your repository the way a senior engineer does, and grade every change against it. Human or agent, the same way.

What is ETV?

Trusted by engineering teams

Kiwi.comFinshapePartners BankaPurple TechnologyGreysonItrinityESETFTMOKeggSecond Foundation

Repository map

Modules, services and stores, and how they connect. Every diff is graded against this map.

Rebuilt on every push
Repository knowledge graphIngestEngineDataAppWebhooks · collector/webhooks · 1.4 ETV last quarterWebhookscollector/webhooksCollector · internal/collector · 14.3 ETV last quarterCollectorinternal/collectorEvent queue · internal/queue · 2.9 ETV last quarterEvent queueinternal/queueKnowledge agent · agents/knowledge · 24.8 ETV last quarterKnowledge agentagents/knowledgeScoring engine · internal/scoring · 58.4 ETV last quarterScoring engineinternal/scoringSummary agent · agents/summary · 19.6 ETV last quarterSummary agentagents/summaryMigrations · store/migrations · 3.6 ETV last quarterMigrationsstore/migrationsStore · backend/store · 31.2 ETV last quarterStorebackend/storework_mix · store/work_mix · 4.2 ETV last quarterwork_mixstore/work_mixReports UI · components/reports · 17.1 ETV last quarterReports UIcomponents/reportsAPI server · internal/server · 27.5 ETV last quarterAPI serverinternal/serverDashboard · frontend/app · 11.9 ETV last quarterDashboardfrontend/app
Touched by the commit aboveDirectly downstreamNot affectedAsync, through the queueNode size: ETV graded last quarter

Imports and call graph

Every push re-reads what depends on what.

Ownership from history

Who touched each module, and when.

Docs, PRs and issues

Linked to the module they discuss.

1,284Files96Modules412EdgesThis commit reached 4 modules across four layers of the app.
01The unit

It starts at the file change.

A model reads each diff against the repository and grades what it was worth in ETV, not how much was typed.

feat(metrics): add five-way work categorization (#1482)

View on GitHub
MaintenanceAIMarcus Doyle·acme/checkout·4 hours ago·e855b191
Summary
1.24ETV0.98ETV0.36ETV0.22ETV

This commit introduces a new capability that elevates tests and documentation into first-class, independent buckets, spanning the scoring engine, the schema and the reporting UI.

To support this, CategorizeFileChange now returns a five-way split, and work_mix stores the per-category totals so a report can aggregate them without re-reading the diff.

Obsolete AST-based pairing routines were removed in the same change, which is why part of it grades as maintenance rather than as a feature.

Full commit message

Related entities

Knowledge graph showing relationships to other commits, PRs, and issues.

  • Pull request#1482Merged

    Five-way work categorization

  • Issue#1435Closed

    Tests and docs disappear into maintenance

  • Commita41c7f0Follow-up

    Backfill work_mix for Q1 commits

The same unit, all the way up.

One unit, five altitudesAddition, not weighting
  • 01File changeGraded from the diff
    A model reads it in the context of the repository.
    0.07 – 0.14ETV
  • 02CommitSum of its file changes
    150 files here, most worth little on their own.
    2.80ETV
  • 03Pull requestSum of its commits
    What reviewers already argue about, now with a number.
    8.40ETV
  • 04Developer, 12 monthsSum of their merged pull requests
    A quiet month of hard work beats a loud month of renames.
    354.2ETV
  • 05Team, 12 monthsSum of its developers
    71.0 a month, the number hiring can’t inflate.
    781.0ETV

Nothing estimated, nothing weighted. Every team number traces back to its file changes.

02What the month was made of

Five kinds of work.

Features, maintenance, tests, docs and fixes. Good AI adoption grows features while fixes stay flat.

How is performance evolving?

Work split by type each month.

ETV
  • 4
  • 31
  • 41
  • 33
  • 78
  • 67
  • 103
  • 98
  • 91
  • 147
  • 88
SepOctNovDecJanFebMarAprMayJunJulAug

The month, written up.

What shipped and what got in the way, each claim linked to its commit.

ReportsWritten summary of the monthJul

The team delivered the foundations of the checks service and quota management, alongside a redesign of the cost insights dashboard. Volume dropped, and the month went into security hardening and architectural refactoring instead.

Highlights

  • Shipped the checks read path and wired PR checks end to end, backend to UI 1a84b86b (Paige)

  • Introduced a company context agent with background research and onboarding setup f642dca5 (Paige)

  • Established the data model, migrations and store layer for the quota system 3487b86a (Marcus)

Observations

  • The reporting module needed repeated fixes to correct metric aggregation bbe7bb6f (Sam)

  • Security hardening took a batch of authorization fixes after a pre-release pen test d8f09e07 (Paige)

  • Refactoring decoupled the reconcile package from the collector service 4a30bb48 (Marcus)

Based on 126 commits59.8 ETVEvery claim links to the commit behind it
03Benchmarking

Benchmark your team against itself.

Other teams and other companies add context. Last year’s team is the fair comparison.

How to read thisBig number is the gap to the baseline in percent. Under it, the level itself in ETV per contributor, per quarter.Every bar shares one scale, 0 to 75 ETV. Green is above the baseline, red is below.
Year over yearThe one that mattersETV per contributor, per quarter

Your team against your team, four quarters ago

+138%per contributor, per quarter, in four quarters
  • Q2 2026 (now)You58.9 ETV
  • Q2 2025Baseline24.8 ETV

Same code, same team, same review bar. When this number moves, you moved it.

InternalUseful, not decisiveEach percentage is the gap to your team, same quarter, each team on its own codebase.ETV per contributor, per quarter

Against the other teams in your org

2nd of 48% behind the Platform team
  • Platform team+9%64.1 ETV
  • Checkout teamYou58.9 ETV
  • Growth team−30%41.2 ETV
  • Billing team−51%28.7 ETV

The team on a decade-old payments service loses this chart and can still be your best.

ExternalContext, not a targetPositions in the peer set, plus the 500 OSS Index, on the same unit.ETV per contributor, per quarter

Against other companies on the same unit

Above median40% above the cohort median
  • Peer cohort, top quartile+21%71.4 ETV
  • Checkout teamYou58.9 ETV
  • Peer cohort, median−29%42.0 ETV
  • 500 OSS Index−40%35.6 ETV

Good for a board meeting, bad as a goal. Their number carries their stack.

The same comparison, in public.

The same engine runs live on the top 500 open source projects. Read the direction, not the level.

+124% mean perf / dev since June+757% OpenAI developers601 engineers · 7 big-tech orgs · 13 months500.navigara.com →
→Start here

See where your teams
stood a year ago.

We backfill from the commits you already have, so you see twelve months of history in the first week.

How ETV is calculated