managers who failed to understand that redoing something in some cases requires starting over from scratch. Not everything can be incremental. I think that's when product needs to rely on engineering to do their own time-boxed discovery work (We've called these technical spikes). This will give your team a better picture of the technical constraints and will allow you to pick and choose what aspects run a higher risk of requiring total re-writes than others. Then maybe you take a bit more time on those high risk items to make sure you future proof it more than you would some other things.