Are you using the right measures to control your operation?

In my last post, I dealt with how you can use service levels to demonstrate the IT Value of your services.
Now we get into the engine room – we deal with how you should use service levels and operational level agreements to measure your own delivery operation. Your “delivery landscape” will be a myriad mixture of the services delivered by your own team, adjacent departments and the providers who handle the scope that you outsourced. I will not dive into how to manage each of these components since this is a subject in itself.
Instead I want to deal with the service levels and measurements that you should have in place for these interfaces.  While designing service levels for such interfaces, I have learnt to appreciate the difference in their nature to the ones that you used towards your business services.

Wired

The questions that you would now ask yourself are very different in nature to those that you use to demonstrate your value. Irrespective of whether you are measuring service levels with your provider, or setting operating level agreements with adjacent departments, the questions that you should ask are:
a) Can you use these measures to control and manage your delivery landscape?
b) Can the levels that you have set for yourself be attained?
c) Are you able to measure these service levels properly?

Can you use these measures to control and manage?

  • Control does not mean measuring each and everything that you can. Seek Emphasis in Service Levels. Sometimes the ease of measuring something serves the propensity to measure and report it.
  • Differentiate between measuring (for the sake of control) and reporting (for the sake of understanding) metrics.
  • Before you delve into finding out what you can measure, rather first concentrate on “what must you measure and why?”
  • previous exercise with your business department would have shown you how the service that you deliver inter-twines and actually affects the business operation
  • This will tell you whether you should look out for critical timelines, large transaction volumes, accuracy,…?
  • Concentrate on the few key measurements that you can actually use to control the key parts of the service components that you are managing. Derive these directly from the understanding of what makes or breaks your business.

Can the levels that you have set for yourself be attained?

  • Defining a Service Level Objective for each of your service components are not enough. You should  set values that are attainable within the costs or boundaries of delivery that you have been given.
  • It is no use to accept a 99.9999% availability target from your end client when you are unable to break this down across your applications and infrastructure or are not able to deliver this to within the budget constraints.
  • This is particularly important when you are getting zealous in managing your provider. In your eagerness to measure and manage your provider, you might be setting target values that are either too expensive.
  • After you have set values that you can attain, find ways to actually control how you attain these values. I have seen some IT managers cleverly build in latency into their system – a latency that can be gradually removed as the load on the system increases. These are many such smart practices one can borrow from system architects.

Are you able to measure these service levels properly?

  • There are miles between the intent to measure something and actually being in the position to do so. And the more holistic the intent (like Business Impact), the more precise you need to be in how you measure it.
  • So as you design your service levels, make sure that you actually know how this can be measured. I once had to convince a client that a particular Service Level Calculation (System Availability) was not fully thought through – it took us two hours to could come up with a formula to measure it.
  • Do the math. Start by very carefully designing the algorithm and formula for the measurement – what goes into the numerator? what goes into the denominator? What is the sample size of measurement? Are you aiming at a %age based measurement, or a number based measurement? What are the implications of both?
  • Then ask yourself: Do you have the service management capacity to measure and follow up all that you have designed?
Best practices from the industry, Service Level Agreement Examples, internet searches, and tool vendors will give you a plethora of choices of what you can measure. Reading about and gathering such measures is the easy part. The tough part is making the choice of what you spend your precious service management capacity on – and the art lies in translating this in summation to your understanding of how you are delivering IT value.

photo credit: Lifelog.it via photopin cc

Wordpress SEO Plugin by Wordpress SEO Plugin