Statistics on Emergency changes

[ITILtak]

Hi,

I am creating reports on change management data. I am showing different types of changes (emergency, normal, pre-approved) as a percentage of the all changes in 2010.

It got me thinking, are there any recommendations or guidelines on what is considered to be a normal or healthy ratio for the chage types? for example...

- what is a healthy or normal % for emergency changes to be of the total changes ?

- what % of changes with change reason as "new functionalty" (as opposed to break fixes, maintenance etc) is considered as providing value to customer?

Any thoughts?

Thanks.

[Diarmid]

There may well be guidelines somewhere, but you would have to check their provenance. There is no practical way of determining what constitutes an acceptable (or good, or bad) ratio of the different kinds of change without:

1. Ensuring that the definition of each change type is the same in all participating organizations and is interpreted in the same way also.

2. There being an understanding of what drives the instances of each change type in this particular (identical) environment and whether these drivers are are inherent by the nature of the environment, or can be influenced by, for example, improved processes, training, equipment inter alia.

So, for example, if emergency changes only occur as a consequence of some external and unpredictable factor, you would be hard pushed to reduce the number of occurrences. Of course if the number of other changes fluctuates, increases or decreases over time, then the percentage of emergency changes will also fluctuate, increase or decrease over time.

The number of changes for the introduction of new functionality is driven by business change and improvement and will be higher in a volatile business than in a more sober environment. Again the percentage will depend on what else is happening.

If your organization matures and develops effective controls for whole new areas of "pre-approved" (not a term I approve of ) changes, then the ratios will change dramatically and suddenly be out of kilter with any external benchmark you were using, thus rendering it even more meaningless.

I am of the opinion that there is little or no value (there may even be detriment) in comparing such ratios with other organizations, because you can never know whether you are comparing like with like. Not even ball-park.

The value in tracking these ratios is that unanticipated fluctuations and trends in the figures can alert you to the possibility that things are nor quite right (or at least that you do not fully understand something). Although, of course, things may actually be hunky-dory and the variations may simply reflect what is happening.

Since, in the real world, your systems can never be perfect, it is always worth reviewing, and, for example, if emergency changes constitute higher risk and/or cost, you can look for ways of reducing their frequency.

However, far and away the more interesting thing to look at is the occurrence of failed changes, including those that actually worked, but cost more time and resource and pain than was anticipated.

One final point (then I'll go and have a coffee), since emergency changes are usually undesirable, it pays to review each occurrence to determine whether it could have been avoided and what steps are worth taking for the future, rather than to merely monitor them statistically.
_________________
"Method goes far to prevent trouble in business: for it makes the task easy, hinders confusion, saves abundance of time, and instructs those that have business depending, both what to do and what to hope."
William Penn 1644-1718

[ITILtak]

Thanks for the reply Diarmid.

I see with your point about comparing the ratios because of the differences in interpretation of types, driving factors etc. But there are some well understood or widely accepted charecterizations about change types. For example:

Emergency/urgent changes are always deployed with very short lead time, due to which it has an increased risk. The reality is these will awlays happen and would never be 0%, but definitly if 50% of your changes are emergencies there got be something wrong. So I was wondering if there is a general understanad on the accepted range for this %.

Similarly, break-fixes are very "reactive" type of changes. They are not contributive to the growth of the organization: which should be resulting in "enhancements" of "new functionality" type of changes. While these areessential to keep the light on operationally, I would think that something is not quite right if 70% of your changes are categorized as such. Again I was trying to tfind out if there are guidelines around that.

There is little information out htere in this. I read some where that 3% to 5% is acceptable for emeregency changes. Just curious to know what are finding.

Thanks.

[Diarmid]