I think that while others are trying to fix the UX/UI, the biggest problem to me seems to be that you’re making changes without enough data to back them up.
I can understand that your PO wants to revert the changes, but I’d hold tight for a bit until you’ve had a chance to do some research. A few routes to consider:
A/B Testing - with users who are familiar with the old design, the new design, and neither design.
Direct User Testing - Get some users to walk you through using the system to perform the date picking task in terms of a real-world scenario, asking them to go through what they like/don’t like about the tool. About 5 people usually gives you around 80% of your data.
Time to task completion. If you have the analytics capabilities to pull out the users who visited the page and actually completed the date picking task both before and after the change, you’ll have a good idea as to whether or not this actually speeding up the process.
Task abandonment (bounce rate). Again, this depends on your analytics capabilities, but if you have the means to look at how many users abandon the page without completing the task, you’ll have a basis for comparison to say whether or not the date picker positively or negatively impacts task completion.