IT Service Management | Enterprise Architecture
This is a case study of the Australian Bureau of Statistics’ 2016 online census project as a portrait of complete management failure by government, the ABS, and its principal contractor, IBM. The case study is split over several pages because of its length. The table of contents below offers quick navigation to the various sections. Scroll down to read the contents of this section.
IT Service Management
When problems began to arise for the ABS online census site on 9 August, one of the avenues for troubleshooting should have been an ITSM function. But there is little sign that this function existed or assisted with problem resolution.
What is ITSM?
An increasingly common management method to align IT with business goals is called IT service management, or ITSM. The idea behind the method is service oriented architecture (SOA), to think of IT as a service to business processes. That way IT is not a separate area of operations. Instead each piece of infrastructure, software, and support is factored into other business processes, like accounting, manufacturing, sales, or customer support. By effectively integrating the cost and function of IT into each organisational value chain, ITSM compels managers to document and understand in detail each aspect of their IT assets and functions in terms that make sense from a business perspective.
An increasingly common ITSM framework in Australia is the IT Infrastructure Library (ITIL) method. Although ITIL has grown to include a de-facto strategic management component, at its core it is about designing and implementing business aligned IT services. Using this method as a means of examining the online census project, it becomes possible to look for signs of ITSM in the ABS, which has an IT division.
The ABS is mandated to conduct census surveys into the predictable future. Therefore any investment into creating an IT service around this function should have triggered a strategic decision to capture the intellectual property and design of a re-usable online census IT platform. Figure 5illustrates what an ITSM process for the online census might have looked like.
It would have included solid business reasons for creating the service in the first place, a detailed description of service requirements and components, the development of metrics against which to measure service performance, and integration of the service into a service catalogue and the organisation’s wider service portfolio as a precursor to drawing up a service level agreement (SLA).
Any work towards this goal should have created a detailed architectural overview of components and functions, budgeting details, and support process documentation. The people preparing such documentation would have gained significant insight into the workings of the system built by IBM, thus making them capable of helping with any troubleshooting when difficulties were experienced.
There is no evidence that a longer-term strategy to integrate the online platform into its operations was acted on by ABS senior management, and no evidence that the specialists who might have prepared the documentation were of any help with the census night troubleshooting.
Public sector organisations are not legally required to develop an ITSM capability, but it would make sense from professional and shareholder perspectives for any organisation facing budgetary pressures to look closely at the costs and business justifications of its IT functions, and to identify all available cost savings with nominally best-practice methods like ITSM.
Even without ITSM, however, the ABS might have achieved the same level of oversight and internal expertise through enterprise architecture (the subject of a separate section below). If there was in fact an approach to developing an architectural overview of the online census project, its use in troubleshooting did not gain any public profile.
From a professional perspective this might indicate that the ABS never intended to support the system as an internal service, or that the necessary work to do so was given low priority.
It probably makes little difference from an IC or SIS perspective whether the ABS ignored or planned for ITSM integration of the online census platform, except, perhaps, as a reference point or benchmark for an evaluation of any future online data collection initiative by the ABS.
Working only with publicly revealed details and suppositions based on that information, which has not been confirmed or denied by credible authorities, it appears there were serious deficiencies in the architecture of the online census project.
That probably does not mean much to most disinterested observers of the online census project, but it is fairly near the top of the strategic planning hierarchy.
Enterprise architecture is a discipline related to, but assimilating business process management as a lower-level of strategy and process ‘decomposition’ or breakdown. Enterprise architecture is what gives BPM an architectural context.
In the context of this analysis, it is important to emphasise that enterprise architecture is not defined here, as it is in some organisations, as merely an extension or seniority level of technical architecture, but as the more strategic parent of functional and technical architectures. Important rôles are played by those downstream solution architectures, but they are not considered in this analysis.
In that context, an architectural consideration of the online census project begins with identification of organisational capabilities, as illustrated in Figure 6. These capabilities function to create value streams, which are distinct from traditional conceptions of value chains. Streams are high level internal processes that create value for the organisation or its clients, while chains are articulations of end-to-end linkages between supply networks, organisations, and distribution networks.
Where an organisation does not possess its own capability to deliver on the outputs for a required service, the capability may be sourced from an external supplier, the way the ABS contracted IBM and others, each with their own specialised capability sets, as illustrated in Figure 7. However, standards described by the Open Group Architecture Framework (TOGAF) make it imperative that the contracting organisation has at least the capability to manage or supervise the outsourced capability.
Without knowledge to the contrary, that capability would logically reside in the bureau’s IT division. In its public service environment, the ABS also has to deal with the external imposition of governance capabilities. In this case from the Finance Department, which has responsibility for a whole-of-government ICT policy. While, it is unclear what influence the Finance Department exerted on the online census project, enterprise architecture practice would require the ABS to have an internal capability to manage such a governance oversight.
Figure 8 illustrates how capabilities from up and downstream suppliers, partners, or contractors are supposed to align. In the case of the online census project, outcomes indicate that neither the Finance Department’s influence nor IBM’s contribution were effectively managed by the ABS. From a professional perspective this raises the question whether the ABS management structure is too rigid to accommodate this kind of ‘choreography’, or whether its senior management is not capable of applying enterprise architecture principles.
If this layer of strategic management was outsourced to IBM as part of its contract, the question is then one of a contractor failing to meet due diligence standards, or just being incompetent. It is worth noting that IBM is listed by TOGAF as a ‘Platinum Member’, has contributed to the development of architectural standards, and cannot claim ignorance of these standards and associated practices.
However, even if the architectural function was outsourced, a question remains about how the ABS managed or supervised that function. If it did not, ABS senior management must also expect censure for negligence or incompetence.
Failure to effectively manage enterprise architecture at this level compounded the failure to do so downstream via ITSM or business process management, making it harder to manage associated project risk, and increasing the difficulty in troubleshooting events and components about which there was no depth of organisational knowledge or expertise.
While these considerations focus on a specialised management area and may not be evident or significant to lay observers, taken together with a more general suspicion of management failure, the citizen perspective is likely to be that the ABS deserves critique for poor management of outsourced capabilities.
The shareholder perspective is harder to hypothesise. If the online census were regarded as the ground-breaking efficiency measure it was apparently designed to be, cost savings on manual data collection and opportunities for a potentially wider range of value-added products derived from a single data source might justify the risks. However, the return on investment for accepting that risk is still under a cloud, with census costs yet to be released, and assessment of the online failure’s impact on the overall census unlikely to be forthcoming in the short term.
Another shareholder response, however, might be highly critical of departures from due process, governance, and professional standards at the same time as heightening risk by creating serious public concerns about privacy, and having no public affairs strategy to address those concerns.