ArchSetu
ArchSetu
Back to blog
Case Study18 June 20269 min read

How a Bengaluru Startup Found 847 Dead Functions Before Their Series A Demo

847 dead functions removed. Build time dropped 34%.
The Situation

A Bengaluru-based developer tooling startup, 3 years old, 11 engineers, preparing for a Series A pitch. The CTO had a nagging feeling the codebase was not as clean as it could be. Three weeks before the investor demo, he decided to find out.

Three Weeks Before the Demo

Arjun Sharma had built the company from a weekend side project. The codebase had grown organically, like most startups: features added fast, refactors deferred, and dead code from abandoned experiments left behind "just in case." The engineering team called it "the archaeology problem." Understanding any part of the system meant digging through layers of decisions made by people who had since left.

On a Tuesday afternoon, Arjun pasted the company GitHub URL into ArchSetu. He expected maybe 50 dead functions. A bit of cleanup work for a sprint.

The report came back in 38 seconds.

847
Dead Functions
41 / 100
Health Score
89
Max Complexity
18.3%
Duplication

What 847 Dead Functions Actually Means

The first reaction was disbelief. Then Arjun started clicking through the dead code list. He recognized names. paymentProcessorV1. legacyOAuthFlow. sendgridEmailFallback. These were functions from integrations they had ripped out 18 months ago when they switched vendors. The code was still there, tangled into the call graph like tree roots, making every new developer afraid to change anything nearby.

The real cost was not storage. It was cognitive load. Every time a new engineer opened a file and saw three different versions of the same utility function, they spent 20 minutes figuring out which one was actually in use. Multiply that by 11 engineers across 200 working days.

I had been telling investors our codebase was production-grade. The ArchSetu score told me we were at 41 out of 100. That was a wake-up call I needed three weeks before our biggest pitch ever.

Arjun Sharma, CTO

The Cleanup Sprint

They ran a focused two-week sprint. The ArchSetu report gave them a ranked list of dead functions sorted by file location, so they could batch-delete entire files instead of hunting function by function. The blast radius analyzer confirmed that none of the dead functions had hidden callers through dynamic patterns.

They deleted 847 functions, 23 entire files, and 12 dependency imports that existed only to support the dead code.

847
Functions Deleted
23
Files Removed
34%
Build Time Reduction
74 / 100
New Health Score

The Investor Meeting

When the Series A investor asked about technical debt during due diligence, Arjun was ready. He pulled up the ArchSetu before-and-after report. Health score from 41 to 74. Dead code from 18% to 1.2%. A measurable, timestamped proof that the team had addressed the debt systematically, not just claimed they had.

Showing the investor a static analysis report with a before/after health score was more convincing than any architecture diagram I could have drawn. It was objective. It was third-party. And it told a clear story about how seriously we take engineering quality.

Arjun Sharma, CTO

The round closed at the target valuation. The lead investor later mentioned the engineering quality conversation as one of the things that gave them confidence.

Run ArchSetu on your own codebase

Paste a GitHub URL and get your full analysis report in under 60 seconds. Sign up first to save the results, or try it free without an account.

Analyze your repository