Is it Time for Decentralized IT?
In today’s fast-changing business landscape, companies need to be agile to remain competitive. The IT department has a crucial role to play in ensuring that businesses can stay ahead of the curve. However, traditional centralized IT departments have been characterized as slow to respond and unable to keep pace with the rapidly changing technology ecosystem. This has led to a growing trend of decentralization, where business departments are building their own IT infrastructures. In this blog post, we will explore the advantages and disadvantages of decentralized IT and whether it is the right time for your company to embrace this approach.
Would your company benefit from decentralizing its IT team? Having spent most of my career in fiercely competitive industries, I've come to realize that speed to market is of utmost importance. The potential losses incurred by being slow to respond, or even better, to leap ahead, are substantial.
Many business departments strive for the first-mover advantage. However, centralizedIT departments, burdened with a wide array of priorities, an extensive backlog, rapid changes in their technology ecosystems, and the need to upskill their employees, have often been perceived as sluggish by the business departments determined to be at the forefront. The challenge of an IT department catering to the priorities of every business department promptly is a long-standing predicament. Too many priorities and too few resources. With the advent of AI and the subsequent business opportunities, I anticipate that the tension between business and IT will escalate. Business departments are now initiating or expanding their own shadow IT and analytics teams, constructing their own data repositories, and leveraging separate technology platforms to enhance their agility.
I consulted Google's Bard and asked, "Which companies are adopting a decentralized approach to IT?" The response included notable examples such as Johnson & Johnson,Spotify, Netflix, Google, Amazon, IBM, Walmart, Dell, Target, and GE Healthcare. These companies organize their teams around specific business products or services, empowering them to make autonomous decisions regarding technology development and deployment. This approach likely grants businesses direct control over IT spending and budget.
Bard also highlighted several advantages of decentralization, including enhanced agility, innovation, employee satisfaction, resilience, and even security. The security advantage can be attributed to the separation of assets, resulting in distinct and varied attack surfaces. However, there are also disadvantages to consider, such as increased costs, reduced control, heightened risk, the proliferation of data silos, and an increase insecurity vulnerabilities. It seems that in the realm of security, there are always two sides to the coin. Nonetheless, my experiences have led me to believe that in a competitive market, the value of speed to market outweighs the costs and risks associated with complexity.
As someone who has worked with data for the past three decades, I consider data strategy a crucial aspect of IT decentralization. An interesting drawback of decentralized IT is the proliferation of data silos. Dealing with legacy environments that contain numerous known and unknown data sets, each with its own metrics and lineage details often buried within a select few individuals convinced me that centralized data management was the only way to fully utilize data. However, after fighting this battle for some time, I realize that alternative approaches must be explored. In a large company, achieving centralized control of data is challenging unless there is an overwhelming external force demanding it, such as government regulations. This brings me to another question within a decentralization strategy: what is your data strategy, a data mesh or a data mess?
Data Mesh promotes assigning data ownership and responsibility to individual business units or departments, but this can still lead to complexity. Multiple departments may vie for ownership of similar data, each with its standards for quality, accessibility, monitoring, and control. Data Mesh offers an answer by establishing data owners and requiring compliance with a set of standards for quality, accessibility, control, monitoring, and more. However, when a business unit prioritizes speed to market, gaining consensus and adhering to standards set by other departments or a centralized governance committee may be seen as hindering agility and incurring additional costs and time not necessary for their specific use case. A business department may question the value of these standards and how they will impact return on investment and time to market.
Within a decentralized data environment, the question of who pays for what arises. If one business department builds a data silo with the capacity to serve its use case, but another department starts accessing it, who is responsible for the increased capacity costs? The answer may seem straightforward, but allocating costs and budgets within a company's internal IT financial system can be complex and frustrating.
During a recent conversation about IT and data decentralization with a colleague, she proposed categorizing IT products and services into three categories. The "Core" category includes essential components that everyone agrees should have only one instance, such as email and network infrastructure, which are managed by a central group. The "Common" category encompasses shared applications used by multiple, but not all, business departments. The "Unique" category consists of IT solutions that are exclusively required by one specific business department. In the case of both the common and unique categories, central IT acts as a consultant or advisor. The challenge lies in reaching a consensus on which category each IT solution falls into. Simple, right?
It is important to note that complete anarchy is not the answer. Decentralization is a complex endeavor that requires extensive collaboration and communication. The effectiveness of communication and collaboration within the IT department and with other departments is a crucial consideration. However, if you are unsatisfied with the agility and quality of your central IT department, it may be time to contemplate change and embrace decentralization on a larger scale within your IT strategy.
In conclusion, decentralized IT can be a viable solution for companies looking to increase agility, innovation, and alignment between business and IT departments.However, it is not without its challenges, and companies need to carefully weigh the advantages and disadvantages before embarking on a decentralized IT model.Decentralization can lead to silos and increased costs, but it can also free up the centralIT department to focus on more strategic technology initiatives. Ultimately, it comes down to your company's specific business needs and priorities, and whether a decentralized IT model aligns with your overall strategic goals.
About the Author:
Dan Whitacre focused on innovation and transforming Kroger’s enterprise approaches to data and analytics architecture, technology strategies and practices as well as its disruption roadmap. He served as the CTO and Vice President of Business Development for a technology services organization and spent six years consulting at IBM on Information and analytics strategy and architecture. Dan recently retired as the Senior Director of R&D and Technology Transformation where he led the internal lab ecosystem as well as University Labs at the University of Cincinnati 1819 Innovation Hub and the Northern Kentucky University's Innovation Gallery.