An interesting thought exercise, however, in reality, this would create chaos.
While the existing patterns for dialing and their subsequent heuristics may not be ideal, we all have adapted to their nuances.
If anything, we blindly react to a dial-pad and type on them in auto-pilot. This type of refactoring would make the user overthink the input process and unlikely gain little in return.
Such changes would be like coming home late at night intoxicated and someone has moved the doorknob you’re used to reaching for and opening.
Further, there’s the likelihood that such changes would result in many errors in the adoption process.
Many banking applications use the standard dial-pad layout for numerical input.
I could just imagine a user’s rage if the dial-pad was changed, and a user accidentally inputs the wrong numerical values when making a banking transfer.
Or when typing in a pin-number - and then they get locked out of their account with too many failed attempts.
The real question - why bother.
The evidence to change is weak.
Moreso, how many people even make that many phone calls these days?
The people who do are probably older and would be frustrated by such changes.
Read this earlier today and asked myself two questions: Is this the time to start redesigning dialer formats given that dialing is probably at an all time low frequency (and declining) for the average person? And, what's to gain from a little efficiency in dialing if I'm dialing one number per day or less often than that?
what's to gain
Being the one to redesign the dial-pad.
Shocked they seemingly completely ignored just trying standard phone keyboard input (which is what people use to enter phone numbers in WhatsApp etc) in favour of being the one to rethink the dial pad.