To implement PLM is a process and change. Speak to anyone in engineering and manufacturing community and they will bring you lots of stories about complexity of PLM implementations and associated cost. Also, you can hear lots of stories about complexity of moving from one PLM implementation to another or switching from one PLM system to another. Companies are spending tons of money to align PLM systems to a specific set of requirements to fit company date management and process needs.
Couple of years ago, I posted “Is PLM customization a data management Titanic?“. For many of manufacturing companies, it is a reality these days. Implementations done 10 years ago can be hardly maintained. To update current implementation to a new PLM system version or another PLM systems is mission impossible. Companies are hiring advisers and consulting companies involved in the implementations and development of PLM systems to run migration and adjust PLM system to a new set of requirements.
Vendors have been trying to resolve the complexity of PLM systems by applying out-of-the-box configurations. However, the success of these ready-to-go systems was somewhat mixed. Pre-configured templates worked well during marketing shows, presentations and evaluations. However, in order to bring system to production mode, (still) required customization to be done. Very often, the customization was coming to replace pre-configured template totally. The problem is not very unique in PLM space. I posted – How to de-customize PLM article few weeks ago. I discussed the importance to decrease customization level as well as presented some similar customization complexities coming from SharePoint implementations.
Today I wanted to provide some recommendations you can follow in order to stay away from costly PLM customizations. These recommendations will also help you to avoid some typical PLM implementation pitfalls. Here are 4 steps to follow:
1- Ask yourself what problem you want to solve with PLM for the next 2 years. The term PLM used by many people in a variety of forms and meanings. Going with specific scope (e.g. change management, quality, bill of materials, etc.) will help you to chart functionality you expect PLM system to support.
2- Outline main data elements and structures PLM system needs to support in order to solve list of problems from previous step. Do it with no connection to specific PLM system and vendor. Make an agreement in your extended team about that. It can take you some time to get to the agreement, but in many situations this is one of the best investments you can do in order to eliminate extra customization steps.
3- Pickup few PLM systems and try to map your requirements to what these systems can provide you without customizations. Don’t be afraid to change your terminology alongside of this process. However, insure that whatever name PLM system is using, it will do what you expect from functional standpoint. It will be a good idea to hire consultant during this stage. It is worth to spend some dollars to avoid future budget waste.
4- Last, but very important. You need to test that selected system is flexible enough to apply changes on top of pre-configured parameters/templates. It is not unusual to provide out-of-the-box system configuration that cannot be practically changed. Practically means in this context the ability to add/modify system and data model and (at the same time) keep most of existing functionality in place. Stay away from system configurations with scripts and customized behaviors hard-coded to a particular data models and workflows.
What is my conclusion? The combination of flexibility and preconfigured environment is the key to stay away from costly PLM customizations. However, these two characteristics are very often mutually exclusive. Vendors can show up ready-to-be-used PLM configuration that will be literally destroyed as soon as you will have to change something. To run a test of how flexible is out-of-the-box PLM model is a key thing not to be punched in the face by future PLM system re-configuration and customization cost. Just my thoughts…
Best, Oleg
