I.T. projects: systems change

Adam Westbrooke explores the issue of systems change from the management perspective. How do those at the top ensure that projects are being undertaken in the right way and for all the right reasons?

At some point, every Managing Partner or executive board will be approached by their IT director or a (usually support) department head asking for approval to buy a large and expensive new IT system. Sometimes this will be replacing an existing system, and other times it may be completely new. But how are the senior management supposed to assess both the need and the proposed solution, other than implicitly trusting their staff and signing off whatever is proposed?
In fact, this is not normally as difficult as it looks, and a few key steps will give confidence the correct decisions have been taken. These steps are taken from the Firstcourt Model for Change , developed to help create successful business change projects.

Identify the business goal
This may seem obvious, but less effective projects start going wrong at this stage. The goal of the project is not “Replace the xxx system”; there must be some reason why the current combination of staff, systems and processes needs to be changed. Usually, this goal will fall into one of the following areas:

• Reduce the number of staff required
• Increase the amount of work that can be done by the existing staff
• Increase the quality and reduce errors
• Increase the speed or turnaround time
• Take on some new form of work

Note that none of these have anything to do with computers or technology. They are a definition of a business change, and in reality should focus on reducing costs or allowing a growing firm to do more for the same cost.

Design processes to meet that need
The second step is to work out what processes are required to support the business need identified. Again, this is not about defining an IT system; the processes should work just as well with paper and people as they do with technology.

This is the step that defines what it is you actually want to do, and as well as providing critical information for describing what you may want an IT system to do, it will also show you how systems with different features can help you achieve the business goal.
Another common mistake at this point is to define a lot of hypothetical things you may want to do in the future. While care needs to be taken to retain some flexibility and identify changes to working practices that may occur soon, changing the way people work is not something that should be undertaken lightly. Given that the new system is based on a requirement for a change now, it will be some time before it is appropriate to go through that upheaval again, so trying to define those future processes in the present is unlikely to be reliable.

Find technology to support those processes
One you have established what it is you are trying to achieve, and how the business processes will work, it is time to look for IT systems to support them. As is well known, computers are good at storing information and at processing large amounts of data according to well defined rules. These are the parts of your processes that you need to identify first.

The first place to look for a system is at your existing one. Can it be reconfigured to match the new processes you want? In a surprising number of cases, changes to customisable data entry screens and some new reports can completely change existing software. The problem is often that the system was put in many years before to support different procedures, and is causing problems with the current ones leading to it being viewed as a hindrance and in need or replacement.

In many situations of course, new software will be required. In this case, tenders can be drawn up from the processes already designed. The important thing to remember is that each system should be checked against these requirements, and not compared against each other feature for feature. Buying a system with 50 features you do not require is not a better option than buying one with only 10 that you do not need!

Another consideration should be whether a single system or multiple systems are required. The overall size of the project may remove any element of doubt on this, but options should be considered against the requirement and not against each other. There is no harm in looking at various combinations from point solutions for specific parts of the procedures to all encompassing systems that can do everything. What you are trying to achieve is buying the minimum amount of technology (and hence spending the least and taking on the least risk) that is needed to support the processes already defined.

Choose the best combination of technology and process
The key to a successful change programme including the installation of a major IT system is to be able to get the best system supporting the business processes, and taking on as much of the work as is sensible. The corollary is that you also have to know when it is not significant that the computer cannot do something and that a manual process is required. As with most other areas, the 80:20 rule applies. Twenty per cent of the cost of a complete system will handle eighty percent of the requirement, and the remaining 80% of cost will only deliver the final 20% of the processes. Not only that, but the project risks split in much the same way, with the final complex elements being much less likely to succeed.

Implement the technology and the process
If the processes to reach a defined goal have been designed up front, implementation of the IT system should be a simple technical matter. By this stage, there should be no discovery that the system does not do something you need, and you should not be having design meetings with the vendor to decide how the system is going to be used. Unfortunately, this is often the way projects are implemented with the purchased decision made first, and then expensive implementation consultancy to decide what it is supposed to do.

Alongside the IT project, it is essential that the other elements for the new processes are also implemented and the staff trained. The project as presented to the staff is not:
“here is a new system, and this is how to use it”
but
“here are our new procedures, and this new system will help with these parts of them”

Following this, it goes without saying that this training should not be delivered by technical staff from the vendor, but in the same way that any change in process would be, usually by management, possibly supported by specialist trainers for some aspects.

Measure the effectiveness against the business goal
This stage is often omitted entirely, even if there is a defined goal to measure.  A good practice is to establish the measurements you will use as part of the goal setting process. This then makes it easy to measure before and after, and has the added benefit of showing up ‘goals’ that are not quantifiable, as suitable measures will not exist.

Refine the technology and the process
The final stage is key to the overall perceived success, and usually to creating real financial success for a change project. While following the steps outlined above will make sure that you get a good working system and have delivered value, no forward planning will be perfect. Once the system has been in place and used for a reasonable period, and following the measurement stage, minor changes and tweaks to both the procedures and the implementation of those procedures can be implemented; consider this the first service pack for your project.

Following these changes, it is essential that the goals are measured again to monitor the improvement. Ideally, the review/refine – implement – measure cycle should be repeated regularly throughout the life of the system to prevent the business getting stuck with inappropriate processes.

Adam Westbrooke is the managing director of Firstcourt, an IT strategy and management advice company specialising in helping professional services firms. For more information call Adam on 0870 350 3660 or see http://www.firstcourt.co.uk .
The Firstcourt Model for Change can be viewed at http://www.firstcourt.co.uk/index_model.htm

Published on: 30-05-2007