Learning a new interface isn't especially novel for consumers, most of the interfaces you take for granted were adopted in the last 30-40 years. The necessary

Learning a new interface isn't especially novel for consumers, most of the interfaces you take for granted were adopted in the last 30-40 years. The necessary tipping point is that the reward for the effort of learning a new interface must, in the preference interests of the consumer, exceed the cost of the learning effort. If it does not then you should revert to the principle of least surprise.