Your premise is wrong. You have not inherited the worst case and tech team you have ever seen. Two reasons for this: (1) You haven't inherited anything, the bu

Your premise is wrong. You have not inherited the worst case and tech team you have ever seen. Two reasons for this: (1) You haven't inherited anything, the business owns the code, not you as an individual. You and the tech team need to work together to to make sure the code keeps generating revenue, and possibly more. No one is owning other people or teams. (2) The code is generating $20m annual revenue. That's pretty cool and not bad at all! I'd follow the following steps: 1. Start by defining responsibility areas: input, code, output (business value). Any codebase can be modelled in this way. Once you have explicitly defined input and output of your code base, you know what your degrees of freedom are as long as you don't mess with input or output of your application. Also a good way to get to know the stakeholder landscape. 2. Introduce version control, move everything to Git. Git enables a nice way-of-working that is recognized industry-wide. Team work is everything. 3. Start writing tests. Preferably E2E tests that will be stable for a long time to come. In all cases, don't disturb the revenue flow with your changes. This will help you to make changes without having angry coworkers in your mailbox when your change caused existing functionality to break. 4. Fix the low-hanging fruit first. Define a list of maximum 5 issues that can easily be fixed in isolation and will improve the code base. Be sure everyone understands why and how things are done. This will boost team ownership. 5. Improve the codebase step-by-step. Be sure for every improvement to explain why it is important in terms of business value. If you can't explain it to yourself, maybe you are just fixing this for esthetics and it's not really important at all. And finally, don't go for a full rewrite. Rewrites always seem easy, until you remember that you forgot to take into account all the edge cases the original code base did take into account and it's not as simple as you've thought after all. Instead move parts of the code to a new codebase and migrate slowly from v1 to v2.