Back to Basics - Utility in IT

The execution of IT in central government or large enterprises. Back to Basics – Utility.

It is now the 21st century and we are still in the dark ages with much of large scale government IT. Not all but often – there are still some great examples of innovation. But so often we hear of delays in this decision, and that policy, and still no system to deliver. Often it is broken, late, overpriced or a complete failure. Predicted benefits or “utility” is rarely delivered. Why is that and is there a solution to it? Many such issues have been brought to light by Brexit and Ukraine needs. The issue fills the media.

Much of the root cause is change and misplaced ambition. By the time one political party or another gets into government, the prior governments plans and applications are part-way through creation by the civil service and its suppliers. Immediate stops are placed on politically sensitive systems and re-work and redefinition starts. Waste and confusion is the outcome. Re-work often involves undoing and backtracking and the poor citizen loses out on progress of any particular hue. Worse still is the addition of new functionality on the old developments which then add up to “bastard children of Satan” hybrids.

There has got to be a better way. Here are ten very simple ideas. These are oft stated but rarely implemented;

1.     Concentrate effort on small modules, ideally self-service that cover the 85% of cases. Leave the 15% to expert “handlers” to slowly add onto core function. Get 85% now, not 100% never. Make small incremental releases into live. The CS thus concentrates on the trickier stuff.

2.     No project should sustain more than 10% gross change of function – it just will not be tolerated. To do so would be to allow sloppy up-front thinking. Not sure how to measure that but it has to be possible.

3.     No implementation project to last more than 12 months including all testing and UAT.

4.     Given government is building simplified c85% cases then put out to fixed price tender the build to software houses. Any overrun is down to them and will also have a deletrious impact on their ability to tender for further work if they fail to deliver. Be firm.

5.     Set up a top quality project management cadre in central government to second top quality PM’s to all application builds.

6.     No projects to be initiated within 12months of any due election date.

7.     Use a core set of technologies across all departments that is reviewed and revised annually by an independent group of expert arbiters. No exceptions.

8.     All Permanent Secretaries to be scored on each project their department initiates or undertakes, and a gross project performance metric calculated. Leadership demands accountability which seems to be the one thing the UK CS avoids. This will place it with the perm sec who actually directs the build team. Make score public by Perm Sec and by Department.

9.     Rate all CS by actual delivery not by “working very hard to achieve” – actual achievement matters, effort put in is just a “qualifier”.

10.  Rate all CIO’s every 6 months by utility delivered. Get a rock-solid definition of utility. It will include measures of savings; revenue; time; resources used; dispersion of function and more.