Benchmarking and gas mileage

My sister Rein helped start a company called Drivolution that trains drivers to lower their fuel consumption as well as help reduce the chance of accidents.  Drivolution does this primarily for large corporations with big-ish fleets.  On average I’m told the results in cost savings can be dramatic.

Over the weekend we talked about the most recent news that a consumer organisation in Belgium figured out that the gas mileage of cars was always 20 to 30% more than advertised by their constructors.   My sister then claimed that through training she was now capable of driving around with a fuel consumption as advertised.  (20-30% lower as before)

This conversation came days after the “Top Gear Prius vs BMW M3” test in which they concluded that a BMW M3 can consume less fuel than a Toyota Prius … in certain conditions:

The guys from Top Gear are right, you can make a big difference in both directions depending on the driving style and conditions.  The things I take away from all this :

  1. Lies, damned lies and benchmarks : blanket statements are useless without the proper context.  That’s the case for almost all the numbers that we deal with every day.  People in ICT generally know this, but it’s true outside of computing as well.
  2. Use the best tool for the job : don’t go racing with a Prius and don’t try to save on fuel with an M3.  Again, the same applies in computing.
  3. It *is* very possible to lower your gas mileage dramatically
  4. Drivolution (and e-positief) seems to deserve to be riding on a wave of success 🙂

Until next time,

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.