The principles of DevOps go beyond just improving software development processes. They also help foster stronger, more productive teams
Strategies for eliminating duplication within IT
Whether the enterprise is embarking upon transformation, consolidating after a merger and acquisition, or simply attempting to reduce costs or eliminating duplicate IT systems and services, one of the key goals is to always drive efficiency in IT.
Maintaining multiple systems that perform the same or similar functions will always be more costly than maintaining a single system. Among the places to look for consolidation and efficiencies are: human resources, IT hardware and software, as well as other vendor agreements for professional services, maintenance etc.
I am going to walk you through some common sources of duplication and discuss some strategies for eliminating this type of duplication when it arises. Here are some of the key places where duplication can originate:
Mergers and acquisitions
Source: Where IT duplication is born of M&A, the acquirer typically chooses the target state with the superseding applications and services.
Strategy: Any gaps in functionality that arise out of decommissioning applications are handled by enhancing functionality within the replacement applications and services.
Source: In re-organization where managed services exist for IT, verticals are instructed to consume the managed services and decommission their IT. Duplicate IT services are found in an exercise to identify the current state across all verticals.
Strategy: The target state definition will accommodate the singular workloads. The case where managed services don't already exist demands greater attention. I address this below in the strategy for "Attempting to eliminate duplication."
Source: If the managed services provided by IT are of poor quality, shadow IT will be considered as the only viable option by business units. Invariably the business units will not coordinate IT strategy, resulting in a menagerie of duplicate IT.
Strategy: Ensuring managed services fully cover customer use cases for workloads and SLAs is key to increasing adoption of managed services. This will drive down duplication and obviate the need to pursue shadow IT.
Source: In large organizations with poor IT governance and control, duplicate IT systems will proliferate as a natural result of business units pursuing their strategic interests. Poor governance is a consequence of weak leadership. Duplication in IT can arise from ignorance of managed services or wilful refusal to de-dupe IT, even when managed services are clearly suitable for the use cases being implemented by the business unit.
Strategy: Simple but powerful tools for correcting these behaviors come from user communities and performance objective alignment. User communities with participation from all business units can reduce the lack of communication and awareness that can lead to the creation of duplicate IT through ignorance. You could also connect managers' bonuses (or the lack thereof) to the degree of adoption of managed services within their in business units. This can have a catalytic affect on adoption and potentially drive a cultural shift as managers that are unwilling to de-dupe eventually leave the organization.
Attempting to eliminate duplication
Source: A most unfortunate outcome of some de-duping efforts is the introduction of duplication. This situation can arise when the decision is taken to introduce a new system when upgrading an existing system or as a target state with the intent of decommissioning other applications.
Strategy: The introduction of new systems always carries a risk of duplicating existing functions, application and services and for this reason alone one should consider the alternative approach of picking an existing system for the target state and investing in that system rather than a new one.
Remember, picking existing systems may not always be an option, but in my experience this option is too readily dismissed in favour of introducing new technology from scratch. Selecting and investing in an existing system is also, on whole, a less risky proposition and one that can be return on investment more quickly, especially if duplicate systems can be decommissioned quickly.
Finally selecting a target state system from an existing portfolio can be an effective first step toward the transformation of the particular workload in question.