Most B2B sales teams prospect the same way: LinkedIn searches, firmographic databases, conference lists. The result is a pipeline full of companies that every competitor is also targeting.
There's a better signal hiding in plain sight: GitHub activity. When a startup is growing fast, it shows up in their code before it shows up in the press. Hiring spikes, new contributors joining, aggressive commit velocity, new repositories being created — these are leading indicators of growth, not lagging ones.
This is the methodology behind GitLeads. Here's how it works.
Why GitHub Is a Leading Indicator of Startup Growth
Engineering activity on GitHub reflects real business momentum. A startup that just closed a Series A is hiring engineers. New hires mean new contributors on repos. New products mean new repositories. Shipping fast means high commit velocity. These signals appear before the company updates their LinkedIn headcount, before the press release, before your competitors find them on Crunchbase.
The lag between "GitHub signals growth" and "the world notices the growth" is typically 30-90 days. That's your window.
The 6 Growth Signals GitLeads Tracks
1. Commit Velocity
Raw commits per day over a rolling 30-day window, normalized by repository age. A 6-month-old startup shipping 40 commits/day is a more interesting signal than a 5-year-old company doing the same. We weight recency: commits from the last 7 days count 3× more than commits from 21-30 days ago.
Why it matters: Sustained high commit velocity means engineers are actually building, not just maintaining. A startup actively shipping new features is growing into new customer problems — which means new budget.
2. New Contributors
First-time contributors to any repo in the organization over the last 30 days. We specifically look for contributors who have no prior commit history in the org — not just new commits from existing contributors.
Why it matters: New contributors = new hires, almost always. When a startup goes from 4 to 9 engineers in 60 days, something happened — fundraising, a big customer, a product bet paying off. This is the most reliable "something changed" signal.
3. Repository Creation Rate
New public repositories created in the last 60 days, filtered for repos with actual commits (not empty init repos). Weighted by the number of contributors who've committed to those new repos.
Why it matters: New repos mean new products, new services, new technical initiatives. A company that spun up 4 new repos this month is expanding scope — and expansion scope means new vendor relationships.
4. Star Acceleration
The rate of change of star growth, not absolute star count. A repo going from 200 to 800 stars in 30 days is a stronger signal than a repo with 10,000 stars gaining 200/month. We calculate a momentum score: (new stars this month) / (average new stars per month, trailing 6 months).
Why it matters: Viral star growth means the developer community is discovering this company. Developer-led growth companies often have engineering leaders making tooling purchase decisions. And star acceleration often precedes a product launch or press coverage.
5. Recruitment Signals
We scan README files, repo descriptions, and GitHub org profiles for job posting keywords: "hiring", "we're growing", "join our team", "open positions", engineering role titles. We also check if the org's GitHub profile links to a jobs page.
Why it matters: A company actively recruiting on GitHub is signaling growth intent to the developer community. They're comfortable with the visibility. And actively hiring companies have budget — they just unlocked a headcount round.
6. Push Activity
Total push events across all repos in the last 14 days, weighted by the number of unique contributors pushing. This is separate from commit velocity — it captures deploy activity, CI/CD pushes, branch activity that reflects a team actively working across multiple surfaces.
Why it matters: High push activity across multiple contributors means multiple workstreams running in parallel. That's a scaling engineering team, not a solo founder project.
How Scoring Works
Each signal generates a sub-score from 0-100, based on the distribution of that signal across all ~50,000 GitHub organizations we monitor. A score of 80 on commit velocity means that org is in the top 20% of commit velocity across our entire index.
The composite score is a weighted average:
- Commit Velocity: 25%
- New Contributors: 30% (highest weight — most reliable signal)
- Repository Creation: 15%
- Star Acceleration: 10%
- Recruitment Signals: 10%
- Push Activity: 10%
We weight new contributors most heavily because it's the hardest to fake and most directly correlated with business momentum. Commit velocity can be gamed (automated commits, massive refactors). New contributors almost always means new humans on payroll.
What Makes a Good GitLeads Lead?
Not every high-scoring org is a qualified lead. Here's how to filter:
Strong signal combination: New contributors plus new repos means the team is growing AND shipping new products. That's a company in expansion mode — ideal for new vendor relationships. New contributors alone might just be a one-time hire.
Check the org size context: An org going from 2 to 6 contributors (200% growth) is more interesting than one going from 50 to 54 (8% growth), even if the absolute numbers favor the larger org. We normalize for this in scoring, but when you're reading the list, look at the growth rate, not just the current size.
Repository type matters: New repos that are product repos (named after a feature, a product, a service) are better signals than new repos that look like internal tooling. You can often tell from the repo name and description whether this is customer-facing.
Avoid the star trap: High star count alone is a vanity metric for sales purposes. A company with 50k stars on an open-source project might be a tiny team with no budget. Cross-check star acceleration with contributor count — both high means real traction.
Using the Weekly Top 50 Report
Every Monday, GitLeads emails a ranked list of the top 50 GitHub organizations showing the strongest growth signals from the previous week. Here's how to work it:
- First pass: Scan for orgs you've never heard of. Familiar names you already know aren't leads — they're companies you've already missed or already have in pipe.
- Check the signal breakdown: For each interesting org, look at which signals are driving the score. New contributors + new repos = hiring + building. That's your opening.
- Find the ICP match: Does this org match your buyer persona? Check their GitHub org description, website link, and primary repo language. A company writing TypeScript + React is a different buyer than one writing Python + ML notebooks.
- Reach out before the press does: The goal is to contact them this week, not next month. The signal is freshest now. Use the GitHub org link to find their website, then find the right contact.
Putting It Into Practice
A repeatable weekly workflow:
- Monday morning: Read the Top 50 report (arrives in email ~8AM UTC)
- Spend 20 minutes flagging 5-8 orgs that match your ICP
- For each flagged org: find website → find buyer → send a personalized outreach that references what you saw on GitHub ("I noticed your team has been shipping fast — 3 new repos and several new contributors in the past few weeks")
- Track which GitHub signal types most often convert to meetings for your ICP — optimize over time
The reference to their GitHub activity in outreach is the differentiator. It's not a cold email — it's a warm observation. "I saw you're growing fast" backed by specifics is far more compelling than "I wanted to reach out about our solution."
The Compounding Advantage
The teams that win with this methodology are the ones that stay consistent. One week of GitHub prospecting produces leads. Twelve weeks of GitHub prospecting produces a mental model of the market: which sectors are heating up, which companies are accelerating their engineering investment, which GitHub languages and repo patterns predict the buyers who convert for your product.
That market intelligence is worth as much as the individual leads.
GitLeads is free to start — sign up for the weekly Top 50 report and run this methodology yourself.