I believe a standardized, "mass producible" probe design where the engineering parts (propulsion, power, communications, etc) are the same from one to the next

I believe a standardized, "mass producible" probe design where the engineering parts (propulsion, power, communications, etc) are the same from one to the next or easily swapped with modules of a common architecture when you need a different option, with several hardpoints that custom scientific instruments can be mounted is really the way to go. You build say 3 flyby probes and a lander every year and the only thing you're scheduling is where they're going. If a particular science instrument isn't ready for a launch, it'll go on the next one to that destination. The design doesn't need to be static, indeed it makes a lot of sense to be continually improving it, but instead of investing all that engineering time into a one off, you get 10 to 15 missions out of each iteration. There might be some missions that really need a completely custom architecture, but these can be limited in scope and highly optimized instead of a $10 billion dollar omnibus. With greater numbers you can both implement process improvements to reduce defects and costs, and also because individual missions are less critical you can lower standards and take greater risks. Further, budgeting becomes so much easier - you don't have things getting cancelled after 80% of the money has already been spent, or repeated delays that triple your program costs, to say nothing of project management not having to reinvent the wheel each time. And from a science perspective you can follow up interesting findings much more quickly.