Know-How Exchange : Research/Metrics

Dear Omar

Using statistics to determine this year’s target would be a misuse of the application of statistical process to historic figures – the figure arrived at would probably have no bearing on the desired level of satisfaction you want to achieve unless you are taking a fatalistic or determinist attitude to what will happen!

For example, the standard treatments give, for example the following figures:

Average 77.3

Forecast 76.64

Standard deviation 1.783255

Trend 76.64

Variance 3.18

Median 77

If you wish to set a target for satisfaction which is achievable based on past performance, you could use the average, forecast trend or median values, depending on what satisfaction parameters you are measuring and depending on how you calculate them. The means of calculation of the satisfaction figure could easily render these basic stats utterly meaningless.

Far more importantly, what are you trying to achieve in measuring these satisfaction figures?

I would guess that it is to obtain a level of satisfaction which is acceptable to your customers and which is also achievable within the constraints imposed on your efforts to improve them. As even the basic measures show a trend which is negative (The gradient of the best fit curve is negative) that would imply that by using stats alone, you would set a lower satisfaction target next year than in the past because unless you change things, that is what the stats say is going to happen!

Surely, your aim must be to improve things, so the target should reflect this. Where the stats come in, is to help you to set a target which is achievable based on the changes you are going to make to improve it.

I don’t know how these changes are likely to affect the satisfaction outcome, but you can estimate this from your analysis of what operational changes should make to your satisfaction outcome.

Where the stats become really useful will be to indicate whether or not the target is achievable. You havn’t managed to get above 80 percent and the situation is declining. Setting a target in excess of 77 implies that you must make changes. Where you set the target will depend on the feasibility and the cost of making the changes needed to meet that target and there is insufficient data for me to tell you what that is.

It is safe to say that were you to set a higher target, then the cost of achieving it and the change needed on the ground would increase. Maybe it would increase exponentially, in which case achieving 100% would require an infinite expenditure or infinite resources.

You need to determine the factors which effect improvement and what the relation of an improvement is with additional cost and additional resource. That will show you what you can afford and that will then determine the target which you might realistically set.

Doing it simply from these figures you have presented implies that you will accept decline, in which case, you enterprise is going to be doomed through a lack of customer satisfaction at some time in the future.

Hope that this helps


Steve Alker

Omar - Setting a goal based only upon these past metrics also ignores the effect of competition. Suppose, for example, the competition was regularly scoring in the 90's - a goal of 80 therefore (the previous high) would still fall short of where the number needs to be.

Therefore, I would try to get a measure of competitive satisfaction so that the goal an be realistic, achievable, and reflective of the competitive environment.

Good luck!