Everyone says "maintenance is the real cost" but I wanted receipts.
60-70% of testing budgets go to maintenance, not new coverage (World Quality Report). 55% of teams spend 20+ hours per week maintaining existing e2e tests. Selenium shops report roughly 80/20. 80% maintenance, 20% new tests. One fintech documented 23 hours per week just updating tests for UI changes. Just updating...
Monday.com described the pattern pretty well: "The same pattern kept repeating itself: a harmless-looking UI change, a merge, and then a failing test in CI."
That's the death spiral. Build the suite, UI changes break half of it, spend a sprint fixing selectors, another UI change breaks them again. Eventually we're spending more time babysitting tests than shipping features lol
Hasura got so fed up they deleted 95% of their e2e tests. "The engineering cost of keeping the tests working... it used to take at least two hours, in the best scenario, to get some code into production." Two hours. Best scenario.
For teams here that actually got maintenance under control: was the fix technical (better selectors, fewer tests, different locator strategy) or organizational (dedicated rotation, different ownership model)?
Zerocheck runs E2E tests on every PR with recordings, screenshots, and step traces.
Get a demo