February 28, 2026 · Zapat Team
The True Cost of a GitHub Issue: Why Per-Issue Pricing Makes Sense
Most engineering teams do not track the cost of resolving a single GitHub issue. They should. The numbers are revealing.
The Hidden Costs
A typical bug fix or small feature involves:
- Triage time: 5-15 minutes to read the issue, reproduce it, understand the codebase context
- Implementation: 30 minutes to 4 hours of focused coding
- Context switching: 10-30 minutes of lost productivity switching into and out of the task
- Code review: 10-30 minutes from a second engineer
- Review revisions: 0-60 minutes addressing feedback
- CI/merge: 5-15 minutes waiting for checks, resolving conflicts, merging
Total: 1-6.5 hours of engineering time per issue.
The range is wide because issues vary. A simple config fix might take an hour end-to-end. A multi-file bug fix with edge cases might take a full day. At a fully loaded cost of $75-125/hour (salary + benefits + equipment + overhead), a single issue costs $50-500 to resolve.
Most teams process 20-50 issues per sprint. That is $1,000-25,000 in engineering time per two-week cycle — just on issue resolution.
Per-Seat vs. Per-Issue
Per-seat pricing (like Copilot at $19-39/seat) charges you regardless of usage. A developer on vacation still costs the same. A week with no issues still costs the same. You are paying for potential, not outcomes.
Per-issue pricing aligns cost with value. You pay when work gets done. No issues labeled? No charge. Busy sprint with 30 issues? You pay for 30 issues. The cost scales with your actual output.
Built-in Cost Controls
Zapat gives you full cost visibility and control. Most issues cost $6-12 based on actual compute usage (LLM tokens, container time). You set a per-issue budget limit so you never get surprised — and you see the cost breakdown on every PR.
Compare this to the human equivalent:
| Human Engineer | Zapat | |
|---|---|---|
| Cost per issue | $50-500 | $6-20 |
| Time to resolve | 1-6.5 hours | 5-30 minutes |
| Available hours | 6-8/day | 24/7 |
| Context switching cost | High | Zero |
| Review included | No (needs second engineer) | Yes (multi-agent review) |
When Per-Issue Breaks Down
Per-issue pricing works best for well-scoped issues: bug fixes, small features, refactors, test additions. It works less well for:
- Large architectural changes spanning many files
- Vaguely described issues ("make it better")
- Tasks requiring human judgment about product direction
These tasks still need human engineers. Zapat is not a replacement for your team — it is a force multiplier for the 60-80% of your backlog that is well-defined engineering work.
The Math for a 5-Person Team
A 5-person team typically resolves 30-60 issues per sprint. At Zapat's average cost of $10/issue, that is $300-600 per sprint for the work that would otherwise consume 30-390 engineering hours.
Even at 50% success rate (Zapat handles half the issues, humans handle the rest), you reclaim 15-195 engineering hours per sprint. That is meaningful time freed up for the work that actually requires human creativity.
Try It
New accounts get $15 in compute credits. Refer friends — you both get $5 each, up to $15 more — bringing your total to $35 in credits. Label an issue, see the cost, judge for yourself whether the math works for your team.