System governance: you get what you measure

By Andrew Clifford

There are a lot of problems in IT that we find really hard to fix. You can make a huge improvement just by measuring how well systems fit your standards.

Many of the problems that we find really hard to fix relate to the qualities of systems themselves, not how we run projects. This includes problems with user acceptance, service, risks, development capability, and technology.

System governance defines standard criteria which reflect what you want to achieve in these areas. How does measurement against these criteria help you overcome the problems?

Implementing a measurement programme sends a clear signal that these criteria are important. This has a strong motivational effect. People want to do a good job, and knowing that improvements are valued and will be noticed can spur people to go the extra mile. In projects, measurement against system governance criteria helps people make better rounded decisions, and not focus solely on time and budget. The unsung heroes of IT - those supporting unloved "legacy" systems - have a chance to demonstrate their worth. Priorities are clarified, and people can channel their efforts and ingenuity into fixing what is important.

But there's more to it than just motivation and clarification. There are real barriers to fixing the problems: a lack of time, money, and management attention. A measurement programme can overcome these barriers. It establishes that there is a problem relevant to management, justifies remedial action, and provides accountability because improvements can be measured.

To avoid bias, measurement requires criteria that cover the breadth of system qualities. But do not spend too long trying to get a perfect list. During the measurement programme, people will inevitably raise objections that criteria are "meaningless", or "biased", or that there are "special cases". This is good. It shows that the measurement programme is working. Using this feedback as an input to an annual review of criteria can be a great way of drawing in detractors to the programme.

A measurement programme can encourage people to lie about the systems they are responsible for. You can minimise this by including independent validation. But to an extent, the programme benefits from "bending the truth". Someone who claims their systems are worse than they really are, to get the time and budget to fix them, has a chance to have their case heard fairly, and then be judged on their results. Someone who wants to look good and claims that their systems are better than they really are, then has to work to make reality match their claims, or risk being found out.

A measurement programme also helps the business-IT relationship. A simple set of high-level criteria helps explain the IT function's role as stewards of the company's business systems, and gives the broader business the opportunity to set objectives. Measurement makes the IT function accountable, and balances out the shorter-term measurements of project and service delivery. A measurement programme can help the IT function make the case for budget to manage the IT systems most effectively.

Measurement helps because it creates the environment in which improvement is possible. Measurement is not magic, but it is critical to success.

Subscription