IT as a Cost Centre

Why does Continuous Delivery encounter resistance from IT executives and managers in so many organisations, and why is it so difficult to implement? Why does IT as a Cost Centre results in long-term Discontinuous Delivery?

Introduction

When an organisation is in a state of Discontinuous Delivery, its organisational model can usually be described as IT as a Cost Centre. There will be a functionally segregated IT department, that is accounted for as a cost centre in which costs can be incurred but profits cannot be generated. The IT department will provide services to product departments, which are accounted for as profit centres responsible for investment decisions and profitability.

IT as a Cost Centre relegates IT to a back office function. Each year, the IT department will be allocated a fixed budget to deliver a set of projects required by the product departments. Each project will represent a large batch of pre-agreed requirements. Short-lived project teams will try to deliver those requirements with scope, resources, and deadlines all fixed in advance.

This dovetails with the traditional view of IT as a universal commodity, popularised by Nicholas Carr in IT Doesn’t Matter in 2003. Carr argued IT was merely an infrastructural technology, and concluded it was merely a cost of doing business that could not provide a sustained competitive advantage.

Cost accounting suzerainty

IT as a Cost Centre is a pre-Internet organisational model from the mid-20th century, that still persists today. The world is now reliant on, and connected by technology, yet in a 2014 survey of 700+ organisations by CIO 48% of organisations still had IT as a cost centre.

In The Cost Centre Trap, Mary Poppendieck traces the popularity of IT cost centres back to the ubiquity of cost accounting. Poppendieck describes how the performance of an IT cost centre is measured solely in terms of cost management. This means accounting metrics percolate into the performance metrics of IT executives and managers, creating a culture of cost control with little regard for product development or organisational performance. These incentives will be markedly different to those of product executives and managers, and usually contribute to an adversarial relationship between product departments and IT.

In cost accounting, inventory is tracked as an asset, maximum resource utilisation is encouraged, and development work is capitalised until production launch. These factors create a hidden bias towards the institutionalisation of large projects, due to their perceived economies of scale. In reality, the project delivery model is an ineffective, inefficient vehicle that impedes value, quality, and flow.

Discontinuous Delivery as the norm

A Continuous Delivery programme is an unending journey of continuous improvement, that requires a substantial investment in order to achieve a time to market that can improve product/market fit and increase revenues.

This is likely to be a hard sell in an organisation with IT as a Cost Centre. IT executives and managers will be incentivised to reduce costs wherever possible, while delivering projects that are supposedly on time, on scope, and on budget. As a result, there will be resistance to the idea of spending money on an internal programme with an explicit goal of improving organisational performance and no fixed end date.

Delivery teams will be short-lived, which encourages people to prioritise short-term feature work over long-term architectural work, which restricts the deployability and testability of different services. The reliability strategy will be to Optimise For Robustness, which increases lead times and failure blast radius. Furthermore, the lack of a mandate beyond development work will make it difficult to work with operations teams to establish consistent toolchains for deployments, logging, monitoring, and alerting.

In short, Dr. Eli Goldratt was right when he said in The Goal “if it comes from cost accounting, it must be wrong”.