Standard Deviation and Goal Setting

It seems to be that time of year, a few of our clients are asking us to help set goals for the upcoming year. Typically my experience with goal setting has always been the edict of the management, and arbitrarily increasing goals by 5% or how ever much they wanted to grow. This has always seemed odd to me and I wanted a way to back up the goals to see if a) they were achievable b) if they made sense for the coming year and c) if they were too achievable and we'll always be going above the goal.

To do this, we've started using Standard Deviation based on prior months data to get this new goal. Ironically, its a pretty simple way to tell where you want to be next year, and back it up.

A little background..
Businesses today are all about six-sigma, a system originally created by Motorola to improve processes by eliminating defects. You'll hear the term mentioned in nearly every corporate boardroom about the nation. Standard Deviation is the core to this whole process because it shows what a 'significant change' actually is, in whatever youre measuring. Pretty cool stuff.

Avinash Kaushik had a great post on using Control Limits to judge swings in traffic. The method he discusses uses Six-sigma and standard deviation to work it all out. Its a great way to show clients if the change is indeed good, bad, or just doing what it normally does.

So how does this apply to goal setting ? Because based on prior performance, you can say with a certain amount of confidence that your can actually get to your goal.

I'll grab a sample on how to do this soon, but you basically find the average of your data points, find the standard devitation of the data points and add them together. This will give you one standard deviation above the mean. Add the average plus twice the standard deviation, you have two... and so on. The more data points you have the better but overall thought is to get an idea of where we want to be going forward. You can then present options of what a 'good', 'better', and 'best' goal is.

To go a little deeper into this, a goal in the 'Good' range will be hit 34.1% f the time, a goal hit in the 'Better' range will be hit 13.6% of the time, and a goal hit in the 'Best' range will be hit 2.1% of the time. This is based on the typical 'bell curve' . Ideally, everything would be 'Best' but 2.1% of the time isn't always realistic (and depending on your data, ever achievable). If the difference between the whole deviations is too large, you could pare it down to half of a deviation.

Overall- its a pretty neat way to show what a good goal for the upcoming year should actually be, and you can have a bigger drive to reach it if you know that it might actually happen.