The other issue is that as fuel efficiency increases, the incremental savings in fuel use actually decrease. This is more apparent if you use the European units of fuel consumption which is measured in L/100km. If consumption is reduced from 12L/100km (19.6mpg) to 6L/100km (39.2mpg) you would save 6L on a 100km trip. Doubling mileage again going to 3L/100km (78.2mpg) only saves an additional 3L. This is apparent when you take an already efficient car like the Prius which starts at 44.6mpg and increase it to the 66.2mpg that Google saw. The result over 12,000 miles of annual driving is 88 gallons of fuel saved. At $3/gal, that's $158/year (after factoring in electricity costs). At that rate the $15,000 conversion would take 95 years to recover the cost. Ouch.
Google points out that most of the driving was on short runs where the engine often runs early in the drive to power certain vehicle subsystems. Longer runs would increase the mileage further, but you still have the issue of diminishing returns. With gas at $5/gal and a conversion cost of $10,000, the payback drops to a mere thirty years. Until vehicles are actually engineered from the ground up as PHEVs and mass produced to bring down cost, the cost benefits simply won't be realized by people doing conversions. Reduced emissions and oil use, though, should still take place.
[Source: Google.org via CNet]