System Design 3 Essential Points

It has to be said that generally I love building business systems.Nothing fulfils me more than when I built a system so that a business process that it now takes a matter of minutes rather than days. It clearly fufills my need to have intellectual stimulation of the highest order. This is especially true when I get stuck on a problem for several days and suddenly have a major breakthrough.

However there are things that I need to understand before actually doing any system design work. This is my thinking time when I get all my “ducks in a row” and it amazes me how many clients clearly value my “brain power” but won’t actually pay for thinking time – but that’s another story.

As part of my thinking time, I try and understand the following before commencing any project:-

System Purpose

This relates to::-

• What is the system trying to achieve ?
• What (business) problems will be solved ?
• What benefits will the system acheive ?

Scope

• A project cannot be an infinite scope - it must have a defined purpose. If the scope is not defined upfront, then scope creep is inevitable.
• One thing I have learned over the years is that most of my clients have very little knowledge of building systems and need a lot of help on defining scope. Indeed most clients go through a steep learning process on what can be achieved. As they see the system being built over a period of weeks, they become more enlightened and ask for an ever number increasing number of bells and whistles.
• However, the greatest danger can be towards the end of the project when there is a request for “one small thing” which can have a massive implication on the database structure. Without it, the system is either useless or has limited value. Hence I spend a considerable amount of time trying to identify these killer questions upfront. This is done in the following ways:-
• Asking the clients a lot of thought provoking questions
• Showing clients other projects that I have built
• Thinking through things myself
• Looking at similar systems that I have built.

How are things being done currently

• I need to understand how the problem is being dealt with now. Just because a process is being automated in some way does not solve everything. Indeed, automation may bring about problems that are not currently being experienced.
• What happens if a few months down the line the process changes, making the automation useless ?
• Accordingly, it is necessary to look at the current processes in depth. It has to be said that very few clients are able to look at their processes to the level of depth that is required. And I do not know the clients business. The solution is a re-iterative process – build a first version and then refine/tweak it. However the true scale of the challenge becomes apparent when one considers the system from end to end. The output from one process feeds as the input into another process. This re-inforces the need for continuous testing throughout the system design – not only to ensure that it is working, but that it is working in the way that it should be.

About the Author

Kapil Kapur is the managing director of Fingertips Intelligence. If you would like to know how we can make your company make better decisions, faster then please get in touch on 0845 163 0149.