When do you decide to rewrite vs. incrementally refactor?

We have a 5-year-old monolith service handling our core billing logic. It works but the technical debt is compounding — every new feature takes 3x longer than it should. The team is split: half want a clean rewrite in a modern stack, half want incremental strangler-fig refactoring. For those who've been through this: what signal made you choose rewrite over refactor (or vice versa)? Was the decision right in hindsight? Context: ~200K LOC, team of 8, revenue-critical service.