Choosing the right metric

One of my “hot buttons” is when people choose the wrong measurement for optimization.

I did that with Boomershoot 2005 when I tweaked the explosive mixture such that I got good results with a rim-fire .22. It turned out that the typical center-fire bullets had great difficulty detonating them.

This metric problem was one of my points in this post.

I heard of one software company that gave raises out based on the number of lines of code produced by the software developers. This adversely affected the designs, implementations, and even style of the code produced.

Another well known anti-virus software company paid bonuses for quickly finding solutions to new computer viruses. One enterprising employee became well known for finding figuring them out. Of course he didn’t tell management that he had created these same¬†viruses and released them into the wild.

Gun control advocates rejoice and claim they were right when deaths and injuries due to bullet wounds decrease after a restriction on firearms is passed into law. They do that even if violent crime, injuries, and deaths increase. They ignore that many of those deaths and injuries are justifiable or even praiseworthy shootings that stopped criminal attacks.

Another one is that windmill manufactures work toward greater efficiency. In most cases this is the wrong measurement. The only case that this is important is if land space or wind is in short supply. If you have lots of land that can host windmills then the correct metric is cost per kilowatt-hour over the lifetime of the windmill.

Some of the first mass produced solar cells were for use in spacecraft. Area and weight were at a premium and hence efficiency was one of the proper metrics to use in the design. That is not true for the side of the shed I used for mixing and storing explosives over a quarter mile from the nearest usable power line. I have lots of area and I don’t care if it takes ten times as much area as the more efficient but twice as expensive solar cells to generate the same power. This is now being realized by the manufactures.

Examine every metric carefully. Think about the unintended consequences. Think about what is really important to the target audience. It can make a huge difference.

3 thoughts on “Choosing the right metric

  1. Traffic cops getting paid on a per-ticket basis comes to mind (it happens) and judges who get a percentage of the fines.

    Salespeople working the same department in a retail store, getting paid on commission (you know, as an incentive to sell) will fight over customers. Been there– it can get ugly.

    A child care center I know once rewarded little kids with candy for using the toilet, for obvious reasons. It then became obvious that they would go to the toilet freakishly large numbers of times, ecpelling the tiniest little pieces…

    The study of incentive systems (a subset of your metric concept, admittedly) is fascinating, and important, especially with regard to our Congresswhores.

  2. I’m a technical writer. Our productivity is measured in pages written and pages revised. You think that makes for clear, concise, cheap to translate manuals? Heck no.

  3. I believe it was Col. Jeff Cooper who coined this, “Redoubling your efforts after you have lost sight of your objectives.”

Comments are closed.