Skip to content

Now with PR-level cycle time

The engineering KPIs your leadership deck has been missing.

Lattice unifies pull requests, deploys, incidents and on-call signals into a single performance graph — so directors and VPs can see, in one screen, how their teams actually ship.

Trusted by 5,000+ engineering teams

The product

Six things every engineering leader keeps reinventing in spreadsheets.

We pulled them out, instrumented them properly, and wired them into one graph. No more screenshots from six dashboards into your monthly review.

  • Cycle time, broken open

    See exactly where the days go — from first commit, to ready-for-review, to merge, to production. Per repo, per team, per quarter.

  • Deploy frequency, that means it

    We separate real production deploys from pipeline noise. No more counting CI runs and calling it throughput.

  • Change-failure rate, not change-fear

    Tie incidents back to the deploys that caused them. Surface the rollbacks that quietly didn't get logged anywhere.

  • Mean time to recovery, end-to-end

    From PagerDuty alert fired to verified-fix-in-prod. Including the wait for someone to actually open their laptop.

  • Review load, fairly distributed

    Find the senior engineer reviewing 40 PRs a week while their staff peer reviews three. Rebalance before they burn out.

  • Quarterly board view

    One PDF, generated on the fifth of every quarter, that your CEO can hand to the board without a single edit.

Trusted by industry leaders

How it works

From raw signals to one performance graph.

Cycle time over a quarter, broken into PR phases.

One graph, every phase of the SDLC.

We instrument every stage between first commit and production deploy — not just the merge. That means you can finally answer the question every CTO actually has: where does the time go?

  • Time-to-first-review per repo, per author seniority
  • Review-to-merge latency, separated by reviewer role
  • Merge-to-production lag, with the CI failures factored in
Deploy frequency, weighted by service criticality.

Real deploys, not pipeline runs.

Every minute spent fighting CI noise is a minute you're not coaching engineers. Lattice tags every deploy with the artifact that actually shipped — and ignores the failed ones, the dry runs, and the rollbacks that cancelled themselves.

  • Per-service deploy weighting (a payments deploy ≠ a docs deploy)
  • Auto-detection of feature-flagged dark launches
  • Trends grouped by team, not by repository

Customer voices

What engineering leaders say after a quarter on Lattice.

  • "We replaced four spreadsheets and two BI dashboards with Lattice in the first month. By the second board meeting, the CFO stopped asking how I knew the deploy numbers."

    Mira Rao VP Engineering, Northbeam

  • "It found a senior reviewer who was carrying 38% of all PR review load on our platform team. That conversation alone saved us a probable resignation."

    Daniel Kovac Director of Platform, Quanta

  • "The change-failure rate dashboard correlates incidents back to the actual PR. We caught a regression pattern in a payments service that had been bleeding us for three quarters."

    Sara Petrov CTO, Ridgeway

Pricing

Priced per active engineer. No seat audits.

Pay only for engineers who actually pushed code in the last 30 days. We don't bill for the alumni still on your GitHub org. No per-feature gates on the things that matter.

  • Starter

    For teams shipping their first dashboards.

    $ 29 /engineer/mo

    • Up to 25 active engineers
    • GitHub + GitLab integrations
    • 30 days of data retention
    • Slack digests
    • Community support
    Start free
  • Enterprise

    For organisations with regulatory needs.

    Custom

    • Self-hosted or VPC deployment
    • SOC 2 Type II + ISO 27001 controls
    • SAML SSO + SCIM provisioning
    • Indefinite data retention
    • Named CSM, 1h response SLA
    Talk to sales

Questions

Things engineering leaders ask before they buy.

How long does the initial setup actually take?

About fifteen minutes for the GitHub or GitLab connection, plus another ten if you're wiring up PagerDuty and Linear. We don't backfill on import — you'll see meaningful trends inside three working days, and a complete quarterly view inside ninety.

Do you store the contents of our pull requests?

No. We process PR titles, authorship, timestamps, review events and merge metadata. We never read or persist diffs, source code, or comment bodies. The full data inventory lives in our trust portal.

How do you count "active engineers" for billing?

An active engineer is anyone who has merged at least one pull request to a production-tracked repo in the last thirty days. People on parental leave, contractors between engagements, and alumni still on your org chart are not billable.

Can we self-host?

Yes, on the Enterprise plan. We ship a Kubernetes Helm chart and a Terraform module for AWS, GCP and Azure. The whole stack runs inside your VPC; nothing leaves your network unless you opt into our managed model registry.

Is there a free tier?

The Starter plan is free for teams of five or fewer engineers, indefinitely. Above that you're on a fourteen-day trial of the Team plan, no credit card required. We don't degrade the trial — you get every feature.

How do you handle multi-org GitHub setups?

We support unlimited GitHub orgs on every paid plan. We also support hybrid setups — GitHub Enterprise on-prem alongside cloud orgs — without forcing you to consolidate. Map repos to teams in any topology you like.

Stop screenshotting your engineering metrics into Google Slides.

Start a free trial today. Connect your first repo in fifteen minutes. See your team's first cycle-time graph by lunch.

Start free for 14 days