PLM Roundtable, Part 1

By Nancy Rouse-Talley

You’ve probably spent or expect to spend a lot of time and money on software that manages product information. According to consulting and research firm CIMdata, companies spent $6.1 billion on such systems in 2005, an increase of over 14 percent over 2004. CIMdata also projects this area to grow to $6.8 billion in 2006 and reach $11.6 billion by 2010.

But how much of this money is well-spent? And what can you expect in the way of technology improvements that will make it easier to manage product definition data? We gathered a roundtable of experts to shed light on the future of the data management segment of PLM.

Getting a Better ROI

DE: Many manufacturing firms have spent more on PLM data-management systems than they intended, and gotten less than they expected. How can customers keep the implementation costs of product data management reasonable?

Jonathan Gable, Dassault Systèmes

< < The best indicator of a vendor’s ability to support service-oriented architecture technology in the future is how open and flexible the vendor has been historically. —  Jonathan Gable, Dassault Systèmes

 

Gable of Dassault: In order for customers to keep implementation costs down and achieve a high ROI, it is critical to select a PLM system with a broad process footprint that works together seamlessly, but can still be installed modularly. This allows the implementation to proceed in phases so benefits can be easily measured and results assessed before moving on to the next phase. Most companies that have had poor implementation results with PLM have either spent several years trying to implement too much functionality at once, or have pursued niche solutions that require costly integration and still do not work well together.

Saitz of PTC: Fortunately, PTC has learned from both its own experiences and the experiences of the market. For PLM to scale from small businesses to large enterprises the applications must be designed to work together and the system must be able to be deployed in an incremental fashion, delivering value at every stage. PLM must be easy to install and maintain and have options for hosting, so companies can invest in high value-add activities like prioritizing those processes for improvement that will deliver the greatest business value and implementing a plan to ensure widespread adoption. By prioritizing the most important processes, assessing their “as-is” and “to-be” conditions, breaking the system implementation into digestible chunks, and implementing a plan for adoption, companies can experience success more quickly and realize the expected value from PLM.

Robin Saitz, PTC

> > For PLM to scale from small businesses to large enterprises the applications must be designed to work together and the system must be able to be deployed in an incremental fashion,  delivering value at every stage. — Robin Saitz, PTC 

Carrelli of UGS: I think there are a number of considerations that can help significantly in this area. First, a well-defined set of requirements and scope can always keep implementation costs reasonable. Second, a phased approach that lets you pilot new PLM implementations as you get business buy-in is critical for successful implementations. Third, select a PLM/PDM supplier with a proven track record of successful implementations and an open business model designed to ensure compatibility between its PLM systems and your current IT environment. Also, make sure they offer strong after-sale support to help work out any implementation issues that may arise. Finally maturing PLM systems will provide more out-of-box functionality that will also reduce the overall cost of implementation.

Miller of CIMdata: A key driver is fundamental program and project management. Companies really need to think about where they want to go with their business, what is important to them that PLM provides,  what are their priorities, what achievements or metrics will indicate success, and then put together a genuine management program with clear objectives.

One mistake that many companies have made is to define too big a first step in their implementation programs. Although an overall vision is necessary to guide decisions and priorities, it’s important to implement portions of the overall PLM strategy in the form of smaller steps that can deliver visible and early success to the company. These initial successes help build confidence and momentum into the next phases. One of the inherent problems with a large implementation is that if you don’t start delivering value within a reasonable timeframe, people lose confidence in the program and support declines. Then the project starts floundering and before long it is viewed as unsuccessful, regardless of what the end results might be.

Wolfe: Customers must recognize that PDM technology isn’t as robust or capable as the makers of this software would lead them to believe. The most successful PDM applications are those that are limited in scope. For example, most workgroup-level CAD data management systems work well. But when you start sharing CAD files among workgroups scattered around the globe, performance becomes awful.

Chris Farinacci, Agile Software

> > Customers want to buy based on a vision for end-to-end PLM — a vision that is specific to and best practice in their respective industry vertical. — Chris Farinacci, Agile Software

 

Customers should also avoid customizing PDM applications. Such customization — which is often motivated by a desire to do too much — leads to huge cost overruns and ongoing support costs each time a PDM application is upgraded. Customers who stick with the off-the-shelf capabilities of their PDM software are much more likely to keep costs under control. If a PDM system doesn’t perform a particular function that you want, don’t customize it just because it would be cool. Instead use the PDM software for what it can do and handle what PDM can’t do the old fashioned way.

Farinacci of Agile: This ... question is at the heart of Agile’s focus ... on delivering configurable, business-ready applications to address specific PLM business problems that deploy in several months versus traditional technology toolkit approaches that are expensive,  require loads of customization, and most times take years to deploy. Today, customers want to buy based on a vision for end-to-end PLM—a vision that is specific to and best practice in their respective industry vertical. But customers want to deploy digestible pieces of that vision via applications in phases based on their most critical business needs.

Can PDM Be More Effective?

DE: What new technologies would enable PDM to become more effective and flexible? What are the chances that PDM systems will become more open and easier to interconnect than they have to date?

Doxey: You don’t need newer technologies to do simple, robust things. It is all in how you develop the program. For example, the way basic SQL and row level locking versus table locking is implemented definitely affects the robustness and flexibility of a system. A system that allows you to have a particular part number on several engineering change orders simultaneously provides for parallel review and lets the document control analyst decide at the last second which ECO ]engineering change order] will get which part revision. In other words, the system enables concurrent engineering. In other systems, you have to process ECOs serially and that slows down the overall product development process. At this point in time, I don’t really need newer technologies in the basic infrastructure, but I can see add-on products which supplement PDM systems. An example of that is using the Google Search Appliance for a supplemental search tool.

Jim Doxey

< < Some (software vendors) go to extreme measures to make things difficult so they can either charge more consulting dollars to do the integration or prevent users from using their system in a heterogeneous environment. —Jim Doxey

 

As far as interconnecting goes, this depends on the vision of your software vendor. Some want to make things more open and others go to extreme measures to make things difficult so they can either charge more consulting dollars to do the integration or prevent users from using their system in a heterogeneous environment. Over time, more open systems will become available, but this may take years because there really aren’t too many vendor-neutral standards for exchanging data and integration of bills of material.

Miller: The industry has continued to evolve and many technologies have been quite valuable in increasing the flexibility and ability to interconnect PLM-enabling systems. Currently, service-oriented architectures (SOA) are receiving considerable attention as vehicles that can contribute to addressing these issues. The objective is to enable overall systems to be provided via components that can be combined and interconnected appropriately. This movement looks very promising and represents a next stage of evolution and maturity in PLM-enabling technologies.

Ed Miller, CIMdata

< < Expect to see PLM suppliers offering ever more focused and pre-packaged solutions to differentiate themselves from competitors.— Ed Miller, CIMdata


In addition to technology, the PLM suppliers are also making systems more robust and flexible by building more knowledge into the systems through well-developed process models. Industry experience is allowing suppliers to build best-practice models of operations into their offerings, in many cases developing pre-packaged process solutions to support specific industry sectors. In addition to operating processes, the suppliers are also focused on developing more packaged and effective implementation processes—to get the PLM solutions implemented more quickly and successfully. As the industry continues to evolve, expect to see PLM suppliers offering ever more focused and pre-packaged solutions to differentiate themselves from competitors.

Saitz: There are a number of technology innovations that will make PLM more robust and flexible in coming years. At the top of the list are digital rights management (DRM) and SOA.

The strong case for DRM is in the area of distributed collaboration. PLM applications have typically implemented access control mechanisms to ensure that team members get the right level of access to data while maximizing the security of that data. However, a big gap in the security lies in protection of content once it leaves the server. In many cases today, once data leaves the PLM system, it is only protected by the good intentions of the team member that downloaded it. This was sufficient in the past because data was shared only within the four walls of a company,  but today collaboration across company and even continental boundaries is commonplace. DRM promises to close this gap by protecting content wherever it goes. With DRM, the data’s rightful owner can decide exactly who gets what access and when, and can ensure that any content that is inadvertently passed on is not usable by the recipient.

SOA is compelling for PLM because it promises to help companies reduce the total cost of ownership for integrations to the extended enterprise. Product data is among the most widely needed information in the enterprise, and making it readily available with minimal hurdles to other enterprise applications is one key to realizing the value of PLM. Ultimately, the success of a services-oriented architecture depends on the discipline of companies implementing it, but an important enabler is the support by PLM and other enterprise application vendors of the key web services standards. This is no easy task, as currently there are over fifty proposed standards related to web services, many of which will disappear or be subsumed by other standards.

It’s important to note that the best use of an SOA is not in the interconnection of multiple, disparate PLM components but in connecting PLM to the broader enterprise. With many PLM vendors, the SOA focus is on stitching together their own disconnected components. This steals focus away from the higher business value of enabling context-appropriate access to product information within other enterprise applications such as CRM,  ERP, or MES. A better approach is to choose a single, integral PLM solution that needs no internal integration patchwork. This allows companies to focus on driving real business value with their integration efforts.

Carrelli: Service-oriented architectures are the most significant technology enablers to increase the flexibility of product-data management systems. Increased openness is a fundamental requirement for these systems to support the level of enterprise value that companies can achieve from PLM. In addition, developing and furthering more standards that can meet the requirements for connectivity will require better cooperation and agreement between PLM competitors.

Bill Carrelli, UGS Corp.

> > Service-oriented architectures are the most significant technology enablers to increase the flexibility of product-data management systems. — Bill Carrelli, UGS Corp.


 

Gable: Clearly, a dedication to service-oriented architecture technologies will make PLM systems more robust and flexible. However,  customers need to be able to separate the hype versus the reality with SOA technology. The best indicator of a vendor’s ability to support SOA technology in the future is how open and flexible the vendor has been historically. To have a truly modern SOA architecture, a software vendor must expose its application functionality with high-level APIs for building composite applications. If a strong architecture did not previously exist, re-architecting a software product to be more SOA compatible will be extremely costly and the vendor’s goals will like fall short.

Wolfe: I haven’t seen any new technologies that look as if they will help PDM do everything that customers expect from it. Most document-management systems are based on some sort of relational database. The relational model is great for bookkeeping, but it breaks down when people force it to handle the heterogeneous data found in manufacturing or construction operations.

L. Stephen Wolfe

< < Unfortunately, most PDM executives think the only way they can hold on to their customers is by taking the buyers’ data hostage. — L. Stephen Wolfe

 

PDM systems could be made more open tomorrow if software company executives would insist that developers employ unencrypted data, publish their data models, and employ standards such as XML in their APIs. Unfortunately, most PDM executives think the only way they can hold on to their customers is by taking the buyers’ data hostage.

#         #        #

In the continuation of this conversation next month (April 2007), our panelists discuss the economic viability of the PLM firms and make some predictions as to where this technology will lead us in the future.

Nancy Rouse-Talley is a freelance writer who focuses on technology subjects, including product lifecycle management,  computer-aided design, and engineering. You can send Rouse-Talley your thoughts on this article through e-mail by clicking here. Please reference “PLM Rountable, Part 1” in your message.

Roundtable Participants

Our panel of independent consultants and representatives of vendor companies includes:

Ed Miller, president of CIMdata and an internationally recognized authority on PLM. He is a frequent keynote speaker at conferences and seminars around the world on trends, directions,  strategies, methods, and technology issues.

Robin Saitz, vice president, solutions marketing for PTC, provider of Windchill content and process-management software as well as Pro/Engineer MCAD, Arbortext publishing, and MatchCAD engineering-calculation software.

Jonathan Gable, vice president of product management for Dassault Systèmes ENOVIA MatrixOne brand. MatrixOne was purchased by Dassault last year.

L. Stephen Wolfe, PE, an engineering management consultant focusing on CAD/CAM, CAE, and product data management. He formerly was publisher of Computer Aided Design Report and Product Data Management Report.

Chris Farinacci, senior vice president of marketing at Agile Software, a publicly traded PLM vendor with a track record of providing PLM solutions to the high tech & electronics, life sciences, consumer packaged goods, automotive, aerospace and defense, and industrial equipment industries.

Bill Carrelli, vice president of strategic marketing for UGS (recently purchased by Siemens). The company’s data-management approach focuses on capturing product knowledge such as CAD product structures, bills of material, options, and variants.

Jim Doxey has worked for as a consultant for several CAD/PLM vendors and as a user of the technology for 22 years. He’s been involved in PLM implementations with companies including Boeing, Philips Medical, Buell Motorcycles, Extreme Networks, and LeapFrog Toys. In addition, he is the co-inventor of Activault for managing SolidWorks files. He is currently working with a large semiconductor manufacturer in Silicon Valley.

Company Information

Agile Software Corp.
San Jose, CA
 
CIMdata
Ann Arbor, MI
 
Dassault Systèmes
Paris, France
 
L. Stephen Wolfe
San Diego, CA
 
PTC
Needham, MA
 
UGS Corp.
Plano, TX