By David Crane
In an earlier blog, my colleague Andy Troup shared an experience where his customer wanted to embark on a process automation project, which could have had disastrous (and consequently frustrating and costly) results, as the process itself was inherently unsuitable for automation.
Automating processes is one of the first projects that organizations embark on once a cloud infrastructure is in place, but why? The answer lies in legacy IT organizations structures that have typically operated in silos.
Many of the IT organizations that I work with have a number of groups such as service development, sales engineering, infrastructure engineering, and IT security that face similar challenges that can include (among many others):
- Applications provisioned across multiple environments such as development, QA, UAT, sales demonstrations, and production
- Managing deployments of application workloads in a safe and consistent manner
- Balancing the speed and agility of deploying the services required to deliver and improve business results while meeting compliance, security, and resource constraints
With the agility that cloud computing offers, organizations look to the benefits that automating the provisioning processes may bring to overcome the above challenges such as:
- Reduced cycle time and operating costs
- Improved security, compliance, and risk management
- Reduced capital and operating expenditures
- Reduced need for management/human intervention
- Improved deployment/provisioning time
- Consistent quality in delivery of services
The IT organizations I work with are often sold these benefits without consideration of the operational transformation required to achieve them. Consequently, when the IT team kicks off a project to automate business processes, especially service provisioning, their focus is on the potential benefits that may be achieved. The result of this focus is that automation becomes a panacea, and not something that should underpin the IT organization’s overall operational transformation project.
As IT leaders, when considering migrating your provisioning processes to your cloud environment, you need to realize that automation alone will not necessarily provide the cure to problems that exist within a process.
You should not consider the benefits of automation in isolation. For example, too much focus on cost reduction can frequently lead to compromise in other areas leading to objections and resistance from business stakeholders. You should consider other benefits that have non-tangible (or direct) metrics, such as improved staff satisfaction. Automation frees your technical staff from repetitive (and uninteresting) activities, which results in both improved staff retention and indirect cost benefit.
As you select processes to migrate from a physical (or virtual) environment to the cloud, the subsequent automation of those processes should not be an arbitrary decision. Frequently my clients choose processes as candidates for automation for reasons based on personal preferences, internal political pressures, or because some process owners shout louder than others!
Instead the desired business benefits the organization wishes to achieve should be considered in conjunction with the characteristics, attributes, and measurable metrics of a process characteristics and a formal assessment made of its suitability for automation.
Your automation project should also be implemented in conjunction with an organization structure assessment, which may require transformation and the introduction of new roles and responsibilities required to support the delivery of automated and self-optimizing processes.
Important Steps to Your Successful Process Assessment
Based on my experience assisting customers in this exercise, I recommend taking these steps before you embark on a process assessment:
- Understand automation and what it actually means and requires. Many organizations embark on automation without actually understanding what this means and the context of automated processes and their capabilities. Subsequent delivery then either leads to disappointment as automation does not meet expectations, or the process is not truly automated but instead has some automated features that do not deliver all the expected benefits.
- Identify and document the expected business benefits to be achieved through introduction of process automation. This is an important task. Without understanding the benefits automation is expected to achieve, you cannot identify which processes are the correct choices to help you do just that.
- Understand cloud infrastructure system management capabilities required to support process automation (e.g., ability to detect environmental changes, process throughput monitoring capability) and implement if required.
- Identify ALL processes required to support automated provisioning (e.g., instantiation, governance, approval) to create a process portfolio.
- Identify the common process automation characteristics that exist across the process portfolio (e.g., self configuration, self healing, self audit and metric analysis). Note that process characteristics are unique, high-level identifiers of automation across the portfolio.
- Identify the common attributes that the process characteristics share. These are more granular than process characteristics and thus may be common to more than one characteristic in the same process.
- Identify the metrics available for each process in the portfolio, and apply a maturity assessment based on their ability to be measured and utilized. Metric maturity is an essential part of the assessment process as it determines not just the suitability of the process for automation, but also its capability to perform self optimization.
Process Assessment Weighting and Scoring
When undertaking a process assessment program, an organization needs to understand what is important and prioritize accordingly. For example, if we consider the business benefits of automation, a managed service provider would probably prioritize business benefits differently to a motor trade retail customer.
Once you’ve prioritized your processes, they can be assessed more accurately and weighted based on each identified business benefit. Prioritization and weighting is essential, and you need to carefully consider the outcomes of this exercise in order for your process assessment to reflect accurately whether processes are suitable for automation or not.
And remember, as previously mentioned avoid considering each assessment criteria in isolation. Each process characteristic and associated attribute can have a direct impact on the desired business benefit, however if its metric maturity is insufficient to support it, then the business benefit will not be fully achieved.
For example, let’s say that you have identified that a business process you wish to automate has a self-healing characteristic. One of the attributes the characteristic possesses is the ability to perform dynamic adjustment based on real-time process metrics. The characteristic and attribute would lead you to expect the realize benefits such as reduced cycle time, reduced OpEx, consistent quality of service, and improved staff retention.
However, although you’ve identified the metrics required to meet the characteristic and attribute needs, they are neither measured or acted upon. Consequently, because the metric maturity level is low, then the expected business benefit realization capability is also lowered.
Figure 1 below shows a small sample of the assessment of a process in relation to a single process characteristic, common attributes, anticipated business benefits and their weighting and the impact that a poor metric maturity has on their capability to deliver the anticipated business benefit.
Figure 1. Assessment displaying impact of low metric maturity
Contrast this to Figure 2 below, which assesses a process with exactly the same characteristics, attributes, business benefits, but has supporting management capabilities and consequently much improved metric maturity:
Figure 2. Assessment displaying impact of high metric maturity
Based on this small data sample, the process in Figure 2 is a more likely candidate for process automation. The assessment process also then identifies, and allows the IT organization to focus on, areas of remediation needed to optimize processes to enable them to be suitable automation candidates.
The result is the IT organization is able to realize not just the business benefits that have been promised by automation more effectively, but they are also able to set realistic expectations with the business, which brings benefits all of its own.
In summary, automation is not the “silver bullet” for broken or inefficient processes. IT leaders need to consider expected business benefits in conjunction with process characteristics, attributes, and metrics and in the context of what is important to the business. By assessing the suitability of a process for automation, you can save the cost of a failed project and disappointed stakeholders. Finally, you should not undertake any provisioning process project in isolation to other operations transformation projects, such as organization structure and implementation of cloud service management capabilities.
I will discuss the steps to success mentioned above in more detail in my next blog.
David Crane is an operations architect with the VMware Operations Transformation global practice and is based in the U.K.