The Visualization Nobody Wanted to See
Someone built a project that turns GitHub outages into red squares on a contribution graph. It's simple. Brutal, even. When GitHub's servers hiccup, your green streaks vanish, replaced by gaps that look like failure. The project itself is maybe thirty lines of clever code, but it exposes something we've all quietly known: we've gamified our work in ways that don't actually reflect what we do.
I spend most of my time building products that live on GitHub. Pull requests, commits, all of it feels tangible. I'm not sure this is the right move, but I've started wondering if the visibility of my contribution graph matters more than the actual contribution itself. Last month, I worked on a critical backend refactor for a client—three solid weeks of architecture decisions, database optimization, nothing that triggered a single commit some days—and my graph was essentially flat. Meanwhile, someone else shipped fifteen small features and looked twice as productive.
The red squares are honest. They're embarrassing. They're also completely outside your control.
Infrastructure Reliability as a Productivity Metric
- GitHub has 99.9% uptime SLA, which sounds good until you realize that means roughly 43 minutes of downtime per month, and you can't guarantee when those minutes hit
- When Figma went down in November 2024, entire design teams couldn't work—not because they were lazy, but because infrastructure failed them
- We measure developer output. We rarely measure the systems that enable it.
- There's a weird acceptance in our industry that platform reliability is someone else's problem
Building AI tools in Bogotá means I'm often working across multiple time zones with teams in California, Europe, somewhere in Asia. When GitHub goes down at 2 PM Pacific, it's 5 PM for me, and suddenly that's tomorrow's problem because I need to wait for infrastructure to come back. I'm not sure this is actually a productivity problem or just the cost of doing business globally, but it feels like we should be angrier about it than we are.
The red squares force a conversation that metrics dashboards usually hide: your team isn't failing when servers fail. Your infrastructure is.
What We're Actually Measuring
Here's the uncomfortable part. Contribution graphs measure activity. They don't measure thinking. They don't measure the three hours I spent reading through a competitor's API documentation, or the architecture decisions I made in conversations that never touched a keyboard, or the fact that I talked someone out of a bad technical direction that would have cost six weeks to untangle later. None of that shows up anywhere.
I've worked with organizations—startups, mid-market companies, even one enterprise client—that hired based on GitHub profiles. High commit counts, long streaks, perfect contribution records. Some of those people were incredible. Others were committing frequently but building nothing of value. The graph became the goal instead of the signal. I'm not even sure which approach is worse anymore.
Red Squares is doing something useful by making the invisible visible. It's reminding us that your metrics aren't neutral. They're political. They reflect what your organization decided to measure, which usually reflects what's easy to measure rather than what matters.
The Thing We're Not Talking About
In product development and digital transformation work, I'm always pushing clients to instrument their systems better. Measure the right things. Understand your bottlenecks. But then I look at GitHub contributions and I realize we've built an incredibly visible system that measures the wrong things reliably instead of the right things imperfectly.
The red squares keep me up because they're a perfect metaphor for a larger problem: we've outsourced our self-perception to platforms that optimize for engagement and retention, not accuracy. GitHub wants you to keep committing. Slack wants you to keep messaging. Linear wants you to keep tracking. The incentive structure is to make these activities visible, measurable, and addictive.
I don't have a solution here. Maybe that's the point.
