Used to work at a PaaS, which had a 17 year old codebase. It did something that none of our competitors could do, which was really important to our client base.

Used to work at a PaaS, which had a 17 year old codebase. It did something that none of our competitors could do, which was really important to our client base. We were a second-tier player in the space, where the 1st-tier player would be exponentially more expensive to implement and of little benefit to any but the largest clients (think, SAP vs Mysql, for analogy's sake). When courting new customers, we often encountered the question, "When are you going to re-write your product and get off [language platform]?" We wrote a white paper to explain where we were, where we were going, and how [language platform] wasn't impeding our 1) security, 2) performance, 3) extensibility. But we got that question every time. CEO asked tech team how long it would take to re-write the entire 17 year old code base, without losing any client facing features, in [any_new_platform]. They ended up selling to a big financial concern (for analogy's sake, think CA -- Computer Associates). Was probably the right move.