I led two medium to large (18 months, 5-10 engineers) successful rewrites. It's not that we didn't like the old systems, they were simply unmaintainable in the

I led two medium to large (18 months, 5-10 engineers) successful rewrites. It's not that we didn't like the old systems, they were simply unmaintainable in the opinion of every engineer, including the ones that built them. The rewrite was gradual and the systems lived side by side (sometimes handling subsets of the scenarios as features got implemented). We had the tribal knowledge in-house to successfully rewrite them in a better, leaner stack. The end result was a much faster and easier-to-maintain system. Features that took weeks to build could be built in 1 or 2 days. Rewrites are hard, but they can be done if you have a good team that know what they are doing. You should take a really hard look before moving ahead with a rewrite, but you shouldn't fear it when it's necessary. Maybe there is an analogy somewhere with a root canal; they are no walk in the park but when you have to do it, you have to do it - and a good professional will be able to do a root canal safely.