Except the more common question is "I have X kwh, how far can I go" rather than "I have to go X distance, how many kwh do I need?"

Well, yes... This is why the Tesla dash prominently and always displays the "how far can I go" figure (in Rated Miles), while the Wh/mi statistic is relegated to the trip meters. But there is very rarely a need to explicitly see the Miles/kWh figure. Even while at a charger, the relevant measure is "Miles per hour of charging", which is shown on the dash, and not "miles / kWh", which is tangential and requires extra calculations to do anything meaningful with.

The main logic behind using Wh/mi (or l/100km) is that it makes gauging relative efficiencies much more intuitive. The number of miles the typical person drives per year is relatively constant; the much more important factor is the cost of the fuel. Suppose you have two cars that you put similar mileage on: a 10mpg Hummer H2 and a 30mpg Civic. And suppose you want to trade up to either a 12mpg Hummer H3, or a 50mpg Prius. Which will save you more money?

Perhaps surprisingly, replacing the 10mpg Hummer H2 with the 12mpg Hummer H3 will save you significantly more money (in fuel costs) than upgrading the 30mpg Civic to the 50mpg Prius.

This is more obvious when you look at it in terms of gallons per 100 miles:

Hummer H2: 10 gallons / 100mi

Hummer H3: 8.33 gallons / 100mi

Civic: 3.33 gallons / 100mi

Prius: 2 gallons / 100mi

So switching from Hummer H2 to H3 will save 1.66 gallons / 100mi, whereas switching from Civic to Prius will only save 1.33 gallons / 100mi.

What I'd really like to see is a car that uses seat sensors to accurately track and display Wh / passenger-mile. (or for an ICE, gallons / passenger-mile.) This gets at the fact that a fully occupied SUV can actually be more efficient (in energy/passenger-mile) than a solo-driven Tesla. It's unlikely that it would make everyone suddenly start carpooling, but it would bring to your awareness just how much difference carpooling can make to overall efficiency.