Strictly speaking, with the heater on, it acts as a miniature HX removing heat from the coolant. Meaning more energy is leaving the system than with the heater off. OK I get it.
But are we talking 1% cooling or 0.001%??
Just looking at the relative difference in size between the radiator and the heater on page 5-14 of the SM, I estimate the heater heat exchanger is roughly 1/8 of the radiator heat exchanger.
Now, assuming that the internal coolant velocities are similar, and the external air flow on the outside of the fins is similar (obviously not), then I would estimate that the heater would add less than 10% the cooling capacity of the radiator... this would account for the reduced convective heat transfer on the heater with the blower motor instead of the airflow through the radiator at speed.
10% is decent, and probably why this "myth" is all over the place.
Like you asked - Is it worth it? To some, maybe. I completely removed my OEM engine oil cooler from the coolant supply and now have dual dedicated air/oil heat exchangers in the rear of the car with a bypass thermostat (to help quickly warm up the oil when first started), and staggered thermostat controls for fans. I also mounted them in high airflow areas to begin with, but I can't take credit for that idea - that came from studying Mark911's oil cooling setup and then modifying it to suit my needs.
The heat exchangers were sized to keep 600WHP cool with track surface temperatures at 110F (an ambient 90F is the most I would ever track with in my old age), and my oil temps in the correct range to suit my the particular oil and bearing clearances.
Hope that helps.
Dave
Edit:
For you engineers, from the SM, the OEM radiator is sized to remove 46,000 kcal/hr (unknown outside delta-T and airflow rate). We know the OEM water pump flows ~40GPM at 6000RPM, but don't know the coolant delta-T.... But, you can compare the 46,000 to an assumed engine power and efficiency.