William Benn

Last month I wrote about the critical dimension of security when outsourcing IT. Security ensures a safe environment in which the technology can operate.

The next factor to consider for any outsourcing relationship is reliability – the extent to which you can trust your technology to perform as expected.

Reliability is absolutely crucial to user acceptance, and hence to achieving business benefits from tech investments. After all, no one wants to be the next high-profile tech disaster.

A lack of confidence in the IT systems also breeds a lack of trust in the overall management and leadership of your organisation in the same way that an inefficient hospital receptionist will cause anxiety about the quality of the surgery.

It’s not just air traffic control, intelligence or banking where reliability of IT systems is of such vital importance. Most systems exist within a complex web of interdependence – so predictable performance and outputs are essential for the effective behaviour of the IT environment as a whole. This may be less of an issue when you have direct control of every part of the chain – but in today’s complex arena of outsourced services, infrastructure and software, the challenge can become unmanageable.

There are two key dimensions to reliability that you can influence throughout the lifecycle of any IT outsourcing arrangement: consistency and stability. Consistency means that the IT services will deliver the expected results in accordance with the agreed specification under all circumstances. Stability means that the services are predictably available and accessible, and that users can have confidence in the operation of the services.

As always, you get what you pay for in IT outsourcing. If you don’t specify parameters for reliability then don’t be surprised if system behaviour fails to meet user expectations. Equally, you may be tempted to set parameters that look good on paper (i.e. 100 per cent availability) but that are unachievable, immeasurable or prohibitively expensive.

The other important principle is to specify reliability in terms of standards and outputs rather than inputs. For example in an IT infrastructure system, availability is the output but the hours of effort put in to keep the system running is an input. You may be interested in the inputs (to ensure that a supplier can feasibly provide the service) but your key measure is the output.

Consistency

In your outsourcing relationships, it’s essential to define the degree of consistency that’s required in the outputs from the system, and how exceptions should be treated. In complex banking systems, for instance, you will want absolute assurance that the system is giving the same answer under the same conditions every time. Hence you’ll require error checking, validation and audit to be at the heart of your outsourcing arrangements.

Thorough testing at every stage in the outsourcing cycle will be essential, and therefore you’ll want to agree the process of testing with your provider upfront. The acceptable pass rate for these tests should be clear and agreed upon, and the costs of testing (which may be significant) should be fully transparent. The implications of failing a test and the costs of rework are best to be formulated in advance so commercial considerations do not derail progress.

The trend towards multi-vendor systems puts even more emphasis upon effective interoperability – the extent to which individual systems work consistently as a whole. An emerging practice of collaborative ‘plug testing’ helps to ensure that systems work to common standards and principles across the full spectrum of technologies.

Stability

For a system to behave predictably, it must operate effectively in everyday conditions but also perform properly when things go wrong. The ‘blue screen of death’ will be familiar to many end users but would be completely unacceptable behaviour for an enterprise wide system.

IT outsourcers build their reputation upon the security and stability of their services (prevention really is cheaper and more profitable than fault fixing). Therefore, if you are working with an established provider, you can usually have confidence in their potential to provide a stable platform. But it’s still essential to work actively with the supplier to ensure this potential is realised in your own unique circumstances.

Working with your suppliers during the selection phase will give you greater confidence in their approach to ensuring the availability of their services. While sticking to the principle of specifying outputs and standards rather than detailed methods, you should satisfy yourself about the arrangements for fault diagnosis and fixing, planned preventive and pre-emptive maintenance, and arrangements for business continuity (including any planned redundancy in the system).

Monitoring and fine tuning

You should have a number of service-level agreements (SLAs) defined with your supplier that are relevant to the reliability of the system. Ultimately, there is no better measure of reliability than customer feedback – and this should be shared with suppliers in a constructive manner to foster a joint problem-solving approach to any issues that are revealed.

An effective IT quality management system, such as Six Sigma, may ensure that a control process is put in place so the outputs are consistent and reliable – and that any minor issues are quickly and cost-effectively resolved in collaboration between client and supplier.

Looking to the future, the good news is that systems generally become more stable and reliable over time. However, the bad news is that this general trend is matched by increasing complexity and user expectations that all systems will have 100 per cent availability under all conditions. The wave of virtualisation and cloud computing should help to create new levels of reliability – and should be factored into any upcoming outsourced IT arrangements.