A news story broke a few weeks ago (courtesy of a study by McAfee, the virus protection software company) pertaining to the global energy costs of email spam. The study outlined a very thorough breakdown of the life-cycle of spam, from creation and dissemination to filtering and viewing. The energy usage per year for each step was analyzed for each major country and also generalized over the global scale. The conclusion: over 33 billion kilowatt-hours (KWh) of electricity are used globally each year as a result of spam. According to McAfee's numbers, this is equivalent to the electricity usage of 2.4 million homes in the U.S., and equals the GHG emissions of 3.1 million passenger cars. Not a small amount.
This story reflects a burgeoning trend by companies to spin their products in a way that shines favorably on the environment. We probably never thought about the impact of spam on CO2 emissions, but thanks to McAfee, we can now feel good about buying their product. This is all well and good, and simply reflects the very positive cultural and societal movement towards cleaner and more efficient energy production and usage in order to reduce our environmental imprint. However, as with all science, it is necessary to analyze studies like this in detail, and to be cognizant of any conflicts of interest that may exist. In the case of corporate advertising, the vested interest of the company in producing data which leads to more sales is glaring.
The most interesting part of the McAfee study is that about 80% of the energy associated with spam comes from the user end: viewing and deleting spam, manually filtering, and searching for false positives (scanning the spam folder for valuable emails accidentally filtered from the inbox). The energy associated with each of these categories is defined as "user hours," calculated by multiplying the time spent for these acts by the average power required by the computer.
It is in the application of these "user hours" that McAfee confuses and distorts the issue, and inflates the environmentally deleterious impact of spam presumably for its own economic benefit. The energy associated with user hours is only a factor if, in lieu of viewing and filtering spam, the computer would have been off. It is the difference in energy between normal non-spam behavior and behavior with spam, the opportunity cost (to use a business term), that should be used in the analysis. The only scenario where the McAfee analysis is correct would be where the computer is on an extra amount of time due to the user time spent dealing with spam. If you spend 15 minutes a day dealing with spam, do you stay an extra 15 minutes at work, or at home on the computer, AND do you turn your computer on and off each time you use it? It is safe to say that most people's computing behavior does not follow this pattern.
Here is another way to look at the issue. Most people take time during the work day to go to the bathroom or to take a coffee break. This on average would be about 15 minutes away from your computer. Do you turn your computer off when you go to the bathroom? I know I don't. Using the same analysis and logic as McAfee, the global energy and CO2 footprint of people simply going to the bathroom would be just as large (maybe I just inadvertently proposed the new advertising campaign for the antidiarrheal industry). The point is, by treating user time as added energy, McAfee has greatly inflated their numbers to their benefit.
I commend McAfee for its efforts to analyze the energy associated with an industry. These types of studies are very useful and often shed light on existing inefficiencies. The references it used and the data it collected were both excellent. However, its representation of the final data fell short. Despite this, they were successful in highlighting a much broader and more important related issue: how much energy is wasted by computers that are on and not being used. I'm turning mine off after I write this.