Recently I came across the question on what is the total direct/indirect cost drivers for customization. Here are a few I have realized with my experience:
Cost of upgrade – This is the most obvious. Customizations make straight up upgrade impossible. Companies have to do their internal fit/gap analysis and development work to keep customizations. This can be a lengthy and expensive process. Customizations are sticky as they become part of your business process and interfacing systems and hence it can also amount to a huge cost to get rid of them.
Cost of portability – This did not use to be a big reason before. when there was not a whole lot of competition in the software space but with SAAS this is coming up as a big cost item. Vendors can create ETL solutions to move data from one product to another if the original product is vanilla. This can reduce transformation time and cost to huge extent. Customizations limit any such portability as customizations are very ingrained with how one does his/her own business. Most of the time it’s these large number of customizations that prevent companies from quickly moving on to the next generation platforms.
Training: When people are hired to work in HR department or when they use MSS or ESS then training and using knowledge from prior jobs become easy. People can be quickly brought up to speed and enabled to use system. Delivered UPK learning can be installed and Oracle courses can be utilized to cut training cost.
Development: lot of development $ are needed to maintain and tweak customizations on AMS contracts too. I calculate it roughly 30 items per developer. The higher the number of customization bigger is the maintainance group.
Infrastructure: Custom code is generally done bad and hence cause a lot more infrastructure need. For ex a company created this employee profile page for all hr users as a start page where they search for an employee and then jump into various transactions for him. A good idea but they took requirements from all different department and populated information on that page from 10 different component on load. This caused serious performance issues over time and a bad experience.
Impead integration: Building on the point stated below, it’s not just the porting that is hit due to customization, it’s also interfacing with other best of breed products. I worked on integration with a vendor for benefit management and I asked them why don’t you have plug-ins build to integrate with Peoplesoft as so many of their clients are on Peoplesoft. They replied it’s because everyone has customized it so much that they put the requirements on us to change things on our side with each implementation.
Return on investment: Buying an ERP and then not able to use it fully due to customizations is bad for the investment made in it but customization done right are also very important otherwise you waste a lot of time on front end and back end to process manually or adhocly. I worked on a workday project where manager and employee processes were really improved from PeopleSoft 9.1 but core users had to do a lot of work on spreadsheets and access database at the front end and tail end of comp processes. The IT $ saved soon changed into HRIS and HR department $ spent.