No--the goal was to break the API so users noticed, with an easy fix. A lot of users weren't even checking the HTTP status codes, so it was necessary to not return the data to make sure the API calls broke.

We did roll this out in our test environment a month in advance, so that users using our test environment saw the break before it went to prod, but predictably, none of the users who were ignoring the warnings for the year before were using our test environment (or if they were, they didn't email us about it until our breaking change went to prod).