(Originally published on www.servicevirtualization.com)
I've often felt that technologists are very good at thinking in terms of whites and blacks, since problems in the realm of IT are often expressed as one of two states: either the server is responding, or it isn't; etc. So it's unsurprising that executives with strong backgrounds in technology (vs. the CxO who is more business focused) think that success is often a good stopping point. After all, if the server has an uptime of 99.9999% then you really can't do much better. Can you?
The answer depends on how you define "success." It's easy enough to define a threshold and state that crossing that line is what success is. But I would challenge you to justify the particular threshold that you have established. For something like 99.9999% uptime that may be easy to do, but for something where the threshold is much lower (and, consequently, the potential upside is larger) you will have a much more difficult time convincing me that you cannot do better.
This is especially true in process related situations where automation is not being fully utilized. In previous blog entries, I've quoted the following figures, which come from Forrester's Q4 2010 Global Release Management Online Survey:
Items 2 and 3 are the result of item 1, since processes that are primarily manually executed do not typically have rollback capabilities built in nor are they speedy. This may mean that an organization has deluded itself into thinking it is Agile in its ability to respond to ever changing market conditions when, in reality, it is Fragile instead.
To the original point, your company's PMO may examine the portfolio of projects underway and determine that a high percentage of them are operating within time and budget constraints, so they are happy. But how do you reconcile their satisfaction with the three items listed above?
To answer: you don't. Your processes can always improve. And, from the same survey, when 44% of the respondents said that it would take a week or longer to roll out a change to even a single line of code, you can start to appreciate exactly how much improvement is possible, at least in the example of application release stated here.
The detriment to getting an action plan adopted and executed is one of determining the financial impact of not executing such a plan, i.e. your primary competitor is inertia. The plain reality is that many organizations do not measure the adverse results of a lack of automation. For example, do you know how much more revenue your company would generate if it could quadruple the number of releases of your primary revenue generating application per year? Do you even know how you would determine this figure?
The upside potential of integrating automation into your processes (whether Dev or Ops processes) should be compelling enough to warrant investigation into the matter. Take a look at the metrics currently used to demonstrate success. Determine which of those are really metrics of the metrics, and then look at the core metrics to see if they are really in line with current industry expectations. Finally, examine how automation would allow those core metrics to be raised so that your organization can become more Agile and less Fragile.
I've often felt that technologists are very good at thinking in terms of whites and blacks, since problems in the realm of IT are often expressed as one of two states: either the server is responding, or it isn't; etc. So it's unsurprising that executives with strong backgrounds in technology (vs. the CxO who is more business focused) think that success is often a good stopping point. After all, if the server has an uptime of 99.9999% then you really can't do much better. Can you?
The answer depends on how you define "success." It's easy enough to define a threshold and state that crossing that line is what success is. But I would challenge you to justify the particular threshold that you have established. For something like 99.9999% uptime that may be easy to do, but for something where the threshold is much lower (and, consequently, the potential upside is larger) you will have a much more difficult time convincing me that you cannot do better.This is especially true in process related situations where automation is not being fully utilized. In previous blog entries, I've quoted the following figures, which come from Forrester's Q4 2010 Global Release Management Online Survey:
- 64% of the respondents were dissatisfied with the level of automation in their software release processes.
- 54% of the respondents were dissatisfied with their ability to recover in the event of a problem either during release or with the application.
- 50% of the respondents were dissatisfied with the speed of each iteration of the release process.
Items 2 and 3 are the result of item 1, since processes that are primarily manually executed do not typically have rollback capabilities built in nor are they speedy. This may mean that an organization has deluded itself into thinking it is Agile in its ability to respond to ever changing market conditions when, in reality, it is Fragile instead.
To the original point, your company's PMO may examine the portfolio of projects underway and determine that a high percentage of them are operating within time and budget constraints, so they are happy. But how do you reconcile their satisfaction with the three items listed above?
To answer: you don't. Your processes can always improve. And, from the same survey, when 44% of the respondents said that it would take a week or longer to roll out a change to even a single line of code, you can start to appreciate exactly how much improvement is possible, at least in the example of application release stated here.
The detriment to getting an action plan adopted and executed is one of determining the financial impact of not executing such a plan, i.e. your primary competitor is inertia. The plain reality is that many organizations do not measure the adverse results of a lack of automation. For example, do you know how much more revenue your company would generate if it could quadruple the number of releases of your primary revenue generating application per year? Do you even know how you would determine this figure?
The upside potential of integrating automation into your processes (whether Dev or Ops processes) should be compelling enough to warrant investigation into the matter. Take a look at the metrics currently used to demonstrate success. Determine which of those are really metrics of the metrics, and then look at the core metrics to see if they are really in line with current industry expectations. Finally, examine how automation would allow those core metrics to be raised so that your organization can become more Agile and less Fragile.