Renovate, throw away and buy new or rent when needed

If your business doesn’t keep afloat, will your IT drag you down? 

 Photo from flickr.

Photo from flickr.

Todays IT-landscapes are becoming more and more complex due to changes in organisations, business models, mergers & acquisitions etc. 

Consolidating applications for IT’s sake, in order to lower the cost of IT is not a viable long term solution for todays business. In a new global world we need new best practices as the old one doesn’t fit.      

A classic architecture principle is to re-use before buy and buy before develop something new.  The question is in which context this best practice is valid?

Should we re-use something old, if it doesn’t fulfil the new needs of the business? The more we use old legacy systems, the more we cement the old business model and existing organisations.

A more modern principle could therefore be, “If you are doing something new in your business, rent the needed capabilities” and don’t mix it with old legacy. Then throw away the old legacy systems when they not are needed anymore.

But, it will cost more will somebody say. Yes, it does. Flexibility always costs. If you are buying a piece of land and build a factory or an office for fifty years time, the annual cost, for sure, will be lower than if you rent the properties for a period of five years.  If you rent an office for a day will it be even more expensive then a five year lease. 

So when doing investments in IT, the longevity of the needed capability is one of the most important decision criteria’s. 

But if we rent, we have to go for a standard solution. Yes, if you are using SaaS the new way, your options to change are very limited. The benefit is that the new service can be up and running in a fraction of time compared with development of a new bespoke system and often for a much lower cost.  

Cloud-based gives you much more flexibility from a life-cycle perspective, but SaaS limits your options in functionality compared to software development or customized standard systems. 

So where will this reasoning lead to?

First of all, best practice is very dependent of the business context. 

There are very huge differences for the need of IT between manufacturing companies and those in bank & insurance or telecom. If the former case are the production lines based machinery and labour, and the IT supports the manufacturing process. Industrial automation is very different from automated IT-systems for financial purposes.

Second, when trying to automate office work with computers, you have to think about the objectives. Is it lower IT-cost, improve the daily routines or is it to reduce the needed work and involve customers and other into your process? 

Third, continuous improvements in business without changes to IT are a very hard nut to crack. Therefore do we need flexibility in our IT-systems and more of this if it’s a none mature line of business. 

We must therefor build parts of our IT-support to be able to change rapidly, without interfering to existing systems. If it does, then we are in a situation where changes take very long time and where the cost for these change are very high. 
If you are unsure about the near future, rent what you need. If you are very sure that there will be no changes in life, then you have other options as packaged solutions or bespoke development.

Limit your options and the cost will be lower is a very simple rule.

This article was first published internally within Capgemini in Swedish Architecture Community newsletter, June 2014.