The discussion about the demand for PLM services yesterday made me think more about challenges in PLM deployments. Implementation is one of the biggest time consuming component in enterprise software. Engineering, manufacturing and product lifecycle domain represents a significant difficulties for software vendor and service provider. It is very rare to see fast implementations. In today’s dynamic business environment it becomes a serious problem for customers. While people are still a major factors for enterprise software implementation to go slow, I wanted not to stop here by blaming people on both (vendor and customer) side, but trying to make an analysis of biggest reasons for implementations to go slow. Here is my top 3 challenges I want to discuss today.
1. How to start small without full enterprise deployment.
Today, full enterprise deployment is one of the fundamental methodology. This is how most of ERP and PLM products are implemented. That’s why, it is slow and requires lot of synchronization between divisions and departments. Item master is linked to BOM, engineering changes, classification, procurement, drawing databases, etc. To have methodology to start small and evolve in the future can be a step in a right direction.
2. How to capture data and focus on business problem solving.
To load existing data is a complex and slow process. It includes data modeling, data sampling, data cleansing, actual import, validation and other steps. Companies have tons of data. In many situations, PLM deployment cannot be done without capturing of all existing data. It slows all down and not allowing to people to start using system.
3. How to solve a problem of interoperability with existing systems, sensors and processes.
PLM system cannot live without connection to existing business systems. It is not unusual to see another PDM and PLM system that company wants to keep, ERP systems, supply chain systems and others. Homegrown systems can present a significant integration challenge too.
Manufacturing enterprise is complex environment. It is near impossible to make a complete revamp in enterprise systems these days. It seems to me, gradual approach can help. The discussion about that isn’t very new. You can hear about staged implementation and phased deployment. It is easy to say and even easy to agree within a specific company. However, products and technologies to support that aren’t ready. Very basic operations like data import or process integration are time consuming. Data modeling is slow and changes can bring existing implementations to halt.
What is my conclusion? Manufacturing companies should adopt agile methods of deployment and implementations in PLM. Together with change management, agile should become a dominant implementation methodology. Manufacturing companies can learn from web companies how to do so. However, PLM tools and technologies must support agile implementation methods. In my view, not all PLM vendors are ready for that. My hunch- technologies and products that can support it are going to be in high demand coming years. Just my thoughts…