From 22 to 38 Engineers: How One Team 10x'd Output in 6 Months [Case Study]

TL;DR: Active contributors on one engineering org's Warclick leaderboard expanded from 22 to 38 over six months — not through aggressive hiring. Two things happened. First, all-branch analytics surfaced work that was always happening on feature branches but invisible to main-branch tools. Second, and more importantly, developer leaderboards combined with AI tools gave non-engineering team members — PMs, QA engineers, data analysts, and technical operations staff — a reason and a means to start building. When being the top PM contributor on the team leaderboard creates a clear "exceeds expectations" signal at review time, the incentive structure changes. Commits grew by 914%, code reviews by 433%, PR cycle time dropped from 4–5 days to 2–3 days, and zero engineers departed in a six-month window against a baseline of 3–4 expected attritions.

The Problem: Growing Without Getting Faster

Twenty-two engineers across three teams: frontend, backend, and platform. On paper, the organization was healthy. Hiring was on track. Code was shipping. But something wasn't adding up.

New engineers took three or more months to become productive. Code reviews were backed up in a queue that grew longer every sprint. Knowledge was siloed — the platform team didn't know what frontend was building, and nobody could tell which engineers were actually carrying the load.

The metrics told a story that felt wrong:

• 397 commits/month across the organization (about 18 per person)
• 710 code reviews/month (about 32 per person, but heavily concentrated)
• PR cycle time: 4–5 days on average
• Team satisfaction: 6.8 out of 10, trending downward

The engineering director's assessment captured it well: they were growing, but it didn't feel like growth. Something was invisible.

The Diagnosis: Measuring the Wrong Things

The root cause wasn't effort. It was visibility. Like most engineering organizations, this team was measuring commits to the main branch and calling that "productivity." Main branch commits are the end of the pipeline, not the pipeline itself.

When they audited where work was actually happening, the picture shifted. Feature branches contained 3x more activity than main. Junior engineers were shipping meaningful work on branches that never surfaced in their analytics tools. Three senior engineers were shouldering 70% of all code reviews, a bottleneck throttling the entire team's throughput.

The measurement system was lying to them. Not maliciously. By omission.

We've written about why main-branch-only measurement misses 80% of work.

The Intervention: Three Changes, Run in Parallel

The team didn't make one big bet. They made three simultaneous changes, each reinforcing the others.

Change 1: All-Branch Visibility

They switched from main-branch-only analytics to all-branch telemetry using Warclick. Instead of seeing commits that reached production, they could now see commits, PRs, and reviews across every branch in every repository.

The immediate discovery was significant. Work they'd never been able to quantify — feature branch development, experimental prototypes, infrastructure changes, code review mentorship — suddenly had numbers attached to it. Junior engineers who appeared inactive in their old tools were, in reality, some of the most prolific contributors on feature branches.

Immediate actions taken: They redistributed code review load based on actual workload. They onboarded new junior engineers into active feature branches instead of isolated starter projects. They made team specializations explicit and used the data to prevent single points of failure.

Change 2: Developer Leaderboards

With complete data in hand, they turned on Warclick's leaderboard, the Clan Table. Every contributor could see their commits, reviews, PRs merged, and coding days ranked alongside their peers.

The results in the first month came quickly. Commit visibility increased 150%. Code review participation broadened — junior engineers who had never reviewed code before started volunteering because they wanted to climb the board. Knowledge sharing spiked as engineers asked top performers how they worked.

But the more consequential shift was who started showing up on the leaderboard.

Product managers, QA engineers, data analysts, and technical operations team members began contributing code. Not because they were asked to. Because they saw an opening. With Claude and other AI tools removing the technical barrier, a PM could build a meaningful internal automation, a data dashboard, or a lightweight feature without deep engineering expertise. And the career incentive was concrete: be the top PM contributor on the team's leaderboard, or the only data analyst contributing code in your org, and you've given your leader a clear "exceeds expectations" data point at review time.

This is the dynamic that's genuinely new. Before AI tools, the contribution bar was too high for most non-engineers to clear consistently. Before all-branch leaderboards, there was no visible recognition to make it worth trying. Together, they opened the contributor pool to roles that had never appeared in engineering analytics before. The jump from 22 to 38 active contributors didn't come from 16 new engineering hires. It came from making contribution visible, fair, and achievable across every role in the org.

The most telling retention metric: over the six-month observation window, the team lost zero engineers. Historically, they'd expected 3–4 departures in that timeframe.

The psychology behind why leaderboards drive retention and motivation.

Change 3: AI Adoption Tracking

Warclick's AI detection revealed that 88.5% of active engineers on the team were already using AI coding tools, primarily Claude, which accounted for 85.9% of AI-assisted commits. More importantly, 90.1% of new code had some level of AI involvement.

The data shifted the conversation from "should we adopt AI?" to "how do we optimize the AI workflow we already have?" Junior engineers were adopting AI tools 2x faster than seniors, which meant new hires were reaching productivity in 6 weeks instead of 3 months.

We measured AI adoption across hundreds of engineers — here's the complete data.

Want to see numbers like these for your team? Warclick captures every commit, review, and PR across all branches, plus AI adoption data. See your real engineering output in 30 minutes. Start Your Free 7-Day Trial.

The Results: 6-Month Before and After

Here's the full dataset. None of this is cherry-picked — you're looking at the complete picture across the observation window.

Volume and Scale

Two notes on these numbers. First, the commit jump from 397 to 4,028 doesn't mean engineers suddenly started working 10x harder. Much of that increase came from visibility — work that was always happening on feature branches but was never counted before. Real productivity gains sit on top of that measurement correction, but the headline number includes both. Second, the active contributor count going from 22 to 38 doesn't represent 16 new engineering hires. It represents the combination of previously invisible engineers now visible across all branches, plus non-engineering contributors — PMs, QA, data analysts, technical ops — who started building once AI tools removed the technical barrier and the leaderboard gave them a reason to do it. Transparency about what these numbers mean matters.

Velocity and Quality

Team Health

Why It Worked: The Reinforcing Loop

The three changes weren't independent. They created a compounding cycle.

Visibility Reveals Capacity

When you can see work across all branches, you discover how much work was actually being done, who has real capacity versus who's overloaded, and which teams are bottlenecked. The team stopped hiring based on gut feeling and started hiring based on data.

Recognition Drives Motivation

Leaderboards sound like they should create toxic competition. The opposite happened here. Engineers saw their peers' contributions for the first time and responded with curiosity, not jealousy. Recognition led to knowledge sharing. Competitive drive was channeled into helping the team.

Recognition drives retention more than compensation when the recognition is fair. Fairness requires complete data.

AI Accelerates Everything

With 85%+ of the team using Claude, the friction that slows down traditional engineering teams largely disappeared. Junior engineers could ask Claude before asking a senior teammate. Async work became viable. Onboarding compressed from months to weeks.

AI didn't replace engineers. It removed the friction between their intention and their output.

AI Opens Contribution to New Roles

The more surprising effect was outside the engineering org entirely. AI tools dropped the barrier to meaningful code contribution low enough that non-engineers could clear it — consistently, not just occasionally.

A PM who wants to automate their release notes process doesn't need to know the codebase cold. They can describe what they want to Claude, iterate on the output, and ship something real. A data analyst can build a dashboard automation. A technical ops person can write a deployment script. The work is different in scope from what senior engineers do, but it's real work on real branches that shows up in Warclick.

And the leaderboard creates the incentive to start. Being the highest-contributing PM on your team's leaderboard is visible to everyone, including your leader. For a PM whose review is coming up, that's a meaningful signal — one that doesn't exist without all-branch visibility and recognition built into the platform.

This is the combination that unlocked the 22-to-38 contributor expansion: AI as the enabler, leaderboard recognition as the incentive, and all-branch measurement as the proof.

What This Means for Your Team

This case study comes from a team that was already functional. They weren't in crisis. They weren't failing. They were growing, and their tools couldn't keep up with the complexity of that growth.

Most teams try to scale by hiring more people without improving visibility, adding process without adding clarity, and measuring metrics that only capture a fraction of the work. The result is modest growth paired with rising burnout.

The approach here was different: visibility gave them clarity, recognition gave engineers motivation, and AI leverage gave everyone the tools to be effective faster. The result was 10x output growth with a culture that got stronger, not weaker.

How to Run This Playbook Yourself

1. Switch to all-branch analytics. If your current tools only measure main branch, you're making decisions based on incomplete data. Start with a free Warclick trial and compare what you see to what you thought was happening.
2. Introduce leaderboards with context. Share the data with your team before you turn on rankings. Build trust first.
3. Measure AI adoption, then optimize it. Don't guess whether your team is using AI tools — measure it.

We built a detailed 30-day playbook for rolling this out step by step.

See your team's real numbers. Warclick gives you all-branch visibility, developer leaderboards, and AI adoption tracking, starting at $4/warrior/month. Setup takes 30 minutes. Start Your Free Trial.

Frequently Asked Questions

Why did commits jump 914% — was that real growth?

Three things combined, not one. First, a measurement correction: feature branch commits were never counted before, so surfacing that work mechanically increased the number. Second, actual productivity gains from existing engineers: faster cycle times, higher deployment frequency, and redistribution of code review load are real workflow improvements, not artifacts. Third, new contributor types joining: PMs, QA engineers, data analysts, and technical operations staff began committing code once AI tools lowered the barrier and the leaderboard made the recognition worthwhile. Each of these is a distinct driver with a distinct mechanism, and conflating them obscures what actually happened.

How long did it take to see results?

The team saw first-month results immediately after turning on leaderboards: a 150% increase in commit visibility and a measurable broadening of code review participation. PR cycle time improvements (4–5 days to 2–3 days) emerged over the first two months as review load redistribution took effect. The retention signal — zero departures versus 3–4 expected — was visible at the six-month mark. Results compound over time. The data in this case study reflects the full six-month picture, not a cherry-picked window.

How do you attribute improvements to analytics tools vs. team growth?

The active contributor expansion from 22 to 38 wasn't driven primarily by hiring. It reflects two things: all-branch analytics making previously invisible engineering work visible, and non-engineering contributors (PMs, QA, data, technical ops) joining the leaderboard as AI tools made meaningful contribution accessible. The cleanest signals that real workflow change happened — not just more people — are the per-person metrics: PR cycle time fell even as contributor count rose, code review participation rate went from 32% to 68% of the contributing population, and team satisfaction improved. The zero-departure result across six months is the strongest indicator that something cultural changed: attrition typically rises during rapid growth periods, not falls.

What happened to code quality as output increased?

Defect rate held flat across the observation window. Deployment frequency rose from 2x/month to 3x/week without a corresponding increase in incidents. Code review participation more than doubled (32% to 68%), adding a quality gate as volume increased. The team did not trade quality for speed. That's the result of measuring review health alongside commit volume — without code review data, you'd have no signal that quality was holding.

Can other teams replicate these results?

The approach is replicable, but the numbers will vary by starting conditions. Teams that have been underreporting feature branch work will see larger apparent gains from the measurement correction. Teams with a higher existing AI adoption baseline will see smaller incremental gains from AI acceleration. The pattern that holds consistently, based on Warclick platform data (Q1 2026), is that review load redistribution and hidden contributor recognition produce measurable improvements within the first quarter. The cross-functional contributor expansion — PMs, QA, data, ops joining the leaderboard — is replicable by any org where adjacent roles have the motivation to build and AI tools available to do it. If your org has non-engineering staff who want to stand out as top performers, the combination of AI tooling and a visible contribution leaderboard creates exactly the right conditions.