My attention was caught by Facebook video where Stan Przybylinski, VP at CIMdata, was interviewed by Graeme Noseworthy at #AgileEngineering Summit. Video was shared by IBM Watson IoT (btw kudos Stan for such a great video). Watch the video below:
https://www.facebook.com/IBMIoT/videos/1896366650452592/UzpfSTYwMDg5ODIzMDoxMDE1NTk5MDM5Nzc3MzIzMQ/
Stan speaks about PLM historical perspective and new trend of agile development in PLM. Agile has a great opportunity in PLM development which has a high rate of failures, especially if you wait until the end of big project to deliver result. Therefore, incremental innovation becomes popular. Agile can help you to calibrate your incremental development. CIMdata calls a special class of system – socio-technical systems where people have a big impact and also introduce one of the biggest challenges to adopt new technologies. Communication and building awareness can help. Like Stan said – people always need to know what is down to the pike and get prepared for changes.
Status quo is the biggest painful part of PLM development. If I breakdown the status quo of PLM development, I can see actually companies that stuck in two major situations – (1) stuck with PLM; (2) stuck with no PLM.
The first group is large company already adopted current PLM mantra and most probably spent a small fortune on PLM implementations. These companies believe in PLM and use PLM. Some of them stuck with old PLM systems and some of them are pretty successful. As much as these companies are larges providers of PLM business and revenues, they are also big, fat and lazy. The problems are usually obvious for them, but to make next move is not a simple thing. They easy can spend few years to evaluate next technology or product to move on. People leading these companies usually tend to minimize the risks by introducing new concepts and ideas – lot of internal politics and games. .
Second group is different. These are companies that didn’t buy PLM mantra or failed to wrap their heads around PLM concepts. As much as you can find them open to innovate, existing PLM ideas aren’t very popular among these companies. They believe PLM is fat, ugly, complex and expensive. These companies are running variety of software to manage their product development processes. Excel (or legacy databases) usually play a significant role in everything they do. These companies usually have a person I call “Chief Excel Officer” responsible to maintain the most critical spreadsheets.
The boarder between these groups are blurred and you can find companies doing some sort of PLM development (usually to manage their CAD assets and lifecycle), but then pulling data from multiple data sources into one giant Excel spreadsheet to share information downstream and organize processes.
My favorite thing in Stan’s video is related to 2 ways to apply agile development (MVPs) – 1/ Agile PLM implementation (which becomes more popular these days); and 2/ Agile software development which allows to build a system using small incremental steps. You basically have 2 MVPs (minimum viable products) – Customer implementation MVP and PLM software/ platform MVP. This approach is great and can improve future PLM development tremendously.
However, I suggest to take one more step and share my experience of combining the idea of both MVPs into development of single multi-tenant PLM platform. We can combine the process of new PLM system development and do it right away with customers. The key element that can enable this process is cloud technologies combined with agile customer driven development. Think about usage of cloud technologies to bridge between both MVPs. Lean startup development (or Lean startup) was the idea around for the last few years. We can twist it a bit and say – let’s build PLM system, but use MVP approach. You can build a system in stages and uncovering new steps in requirements, planning small deliveries. Great benefits of this approach are ability to minimize risks, work close to the customer requirements and building trust. Main challenge is to manage expectations. People enjoy the process, but also like the results. Agile development is a key to make it successful.
Some of these methods are not unique and already used in PLM (and not only) development. These are 3 innovative methods:
1- Hosted PLM. Existing PLM systems hosted for customers using cloud IaaS or other cloud/hosted infrastructure.
2- SaaS PLM development. Usually it is cloud based system gradually developed and available from one of cloud infrastructure platforms (eg. AWS).
3- Agile PLM implementations. These are typically on premise systems implemented modern agile methods in sprints with relatively small intervals of deployment and customer delivery validation.
However, combining all of them together can create an new approach to deliver PLM experience within time. In my view, the idea has some legs and was adopted during the last few years of OpenBOM development (reminder and disclaimer – I’m OpenBOM co-founder and CEO). The opportunity is to build a system for great multitude of users which will have a set of aligned and verified customer requirements. Magnitude of customer feedback combined with cloud technologies helps to build a system of common requirements capable to support a wide range of manufacturing companies – from very small shops to large OEMs.
What is my conclusion? Agile combined with cloud can become a powerful method to build a system with customers for customers. Two processes of building technologies and implementing PLM technologies can be combined to deliver a system with validated set of requirements aligned to customer expectations. From my experience, these two things are usually very hard to manage – too often in the history of PLM development and implementations, vendor oversold what is available and under-delivered what was promised. Just my thoughts…
Best,Oleg
Disclaimer: I’m co-founder and CEO of OpenBOM developing cloud based bill of materials and inventory management tool for manufacturing companies, hardware startups and supply chain. My opinion can be unintentionally biased
