Step 1 (underway) achieve one million cars / year BEV production. Realistic target 2020-2022
Step 2 - Achieve 1% worldwide vehicle fleet BEV (20-30 million cars on the streets). Realistic target 2025-2030
Step 3 - Focus on worldwide mass adoption of BEV for mass transit vehicles (cabs, shared transportation, bus, short range ferry using high efficiency solar-battery, hyperloop), here is one of the big leapfrog moments, for every 1% of very high mileage vehicles gone BEV many % of oil consumption could be reduced, and this is a win-win-win scenario as BEV are the most cost effective the more often they are used. This is an effort that should start today.
The critical aspect is reducing oil consumption rather than number of ICE replaced with BEV. The landmark shouldn't be million BEVs delivered but starting to reduce worldwide oil production.
My Brazil is quite myopic on EVs, but there's an interesting tradition here as far as subsidies, focus subsidies on cabs. When Brazil went all out on Sugar Cane Ethanol in the late 70s, subsidies were offered for cabs, 10 years later most cabs went ethanol. Recently when Brazil started to push CNG vehicles, subsidies pushed most cabs to go natural gas in areas served (95% of Brazil's urban population have access to CNG today and a very high share of cabs run on CNG today). Recently a partnership between Nissan and the largest cities in Brazil brought the LEAF to Brazil (which are currently 3x the price of a CNG car without any subsidies, but have been shown to cost half per mile vs CNG or 1/4 the cost of running on gasoline), around a thousand imported LEAFs are running in Brazil today and Nissan announced it will assemble the LEAF in Brazil starting in 2016.
Most cabs drive a dozen hours a day, so I suggest starting the focus there.