"We'll add tests later." I've heard this at every company I've worked with. The logic seems sound: ship fast now, automate later. But "later" has a compounding cost that most engineering leaders underestimate by 5-10x.
The math nobody does
Let's say your team of 8 engineers ships without automated tests. Here's what I've measured across real engagements:
- Manual regression testing: 2-3 days per release cycle. If you release weekly, that's 8-12 engineer-days per month spent clicking through flows that a script could run in 10 minutes.
- Bug investigation time: Without tests catching regressions early, bugs reach staging or production. Each production bug costs 4-8 hours of investigation, fix, review, and deploy. Teams without automation average 3-5 preventable production bugs per month.
- Context switching: A developer interrupted to fix a bug loses 20-30 minutes of context per interruption. Multiply that across 3-5 incidents per month across the team.
- Deployment fear: When nobody trusts the deploy, releases slow down. Weekly releases become biweekly, then monthly. Feature velocity drops 30-50%.
Conservative estimate: a team of 8 without test automation wastes 40-60 engineer-hours per month on preventable work. At $75-150/hour loaded cost, that's $36,000-$108,000 per year. Every year you wait.
The compounding problem
It gets worse over time. Code written without testability in mind is structurally harder to test later. Functions are too coupled, side effects are everywhere, and there are no seams for mocking. What would take 2 weeks to automate at month 3 takes 2 months at month 18.
I've seen teams spend 6 months retrofitting test automation onto a 3-year-old codebase. The same setup would have taken 2-3 weeks if done from the start.
What "doing it right" actually costs
A focused test automation setup for a typical SaaS product:
- Week 1: Playwright setup, CI integration, 5-8 critical path E2E tests
- Week 2: API test coverage for core endpoints, test data management, reporting
- Ongoing: 2-4 hours per sprint to maintain and expand coverage as features ship
Total investment: 2 weeks upfront + minimal ongoing maintenance. The ROI is measurable within the first month.
What to measure to make the case
If you're an engineering leader trying to justify the investment, track these metrics for 30 days:
- Hours spent on manual regression per release
- Number of bugs that reached production and could have been caught by an E2E test
- Average time from bug report to fix for regression bugs
- Number of delayed releases due to quality concerns
These numbers will make the business case for you. Every team I've done this exercise with has ended up investing in automation within the next quarter.
Want to scope what automation would look like for your team? Book a free intro call. We'll review your current setup and give you a concrete plan with timelines.