How PLM platforms will survive in no-stack era

TechCrunch article Software for the full stack era speaks about brutal reality existing business software platforms are going through. From the beginning it speaks about reasons major business platforms were created – IBM to automate clerks’ work, SAP – to unify enterprise finance and Siebel to computerize sales rep Rolodexes.

Which made me think what problem existing CAD / PLM platform solved? My guess today’s PLM platform were born to solve a problem of complex configurable 3D digital models for aerospace, defense and automotive companies. Look around, each of these OEMs is using PLM platform from one of top 3 PLM leaders – Dassault, Siemens PLM or PTC. To made it happen companies spent a fortune on gigantic enterprise IT effort to install, configure, plan, implement, support and maintain upgrades and future business requests. The follow passage from TechCrunch article is my favorite, because it explains how it usually happens in every company.

For all the innovation these relationships have brought to business, there’s been a price: the custom patchwork of disparate technologies that now power global enterprises. Typical IT lacks flexibility and efficiency, and often fails to fulfill its promise to customers. If you look under the hood at most global businesses, what you find is something akin to a Ford Model-T that has been continuously fixed by an expert mechanic to keep it on the road for a century. It might be a feat of engineering, but there ain’t no way that car will be driverless anytime soon. Unfortunately for leaders of global businesses, these clunky technical landscapes are doing more than simply irritating employees and customers — they’re jeopardizing their long-term survival.

The last point is clearly resonating with voices of large manufacturing companies having concerns about slow ROI and the fact traditional PLM platforms reached their limits. Solving urgent business needs by integrating new technologies with existing core PLM solutions can introduce a significant problem. Sometimes, it cost so much that it drives huge platform migrations. For most of PLM systems, the architecture and technologies of core functions such as CAD data management and BOM management were designed back 10-15 years ago. To connect and interplay between heavily customized core PLM modules and expanded PLM solutions can bring significant service and implementation expenses.

Technology and architecture of existing PLM platforms are going deep back to the time when enterprise IT was a king and to run PLM system atop of proven RDBMS was a key to success. This is probably still a reality for many manufacturing OEMs. But large manufacturing OEMs are starting to think about their long-term survival. That’s why GE has FirstBuild microfactories  and car manufacturers are looking at Local Motors community model.

So, what PLM platform will support future manufacturing environment? TechCrunch brings Uber model as an example of how full-stack startups like Uber built their infrastructure:

With so much outdated infrastructure from a pre-Internet era, it’s inevitable that the way we consume even the most basic services will evolve. Companies need to embrace this fact to participate in the software-defined future. But doing so requires a dramatic change in how software is perceived, developed and consumed.

So the question arises, if the tech solutions of the 1980’s and 1990’s won’t suit the needs of today’s innovators, what will? The technologies that power these businesses, no-stack technologies, tend to be API-based micro-services that package up a lot of underlying capability. Unfortunately, the similarity in terms leads many to think that these two are diametrically opposed. They’re not.

Uber, the prototypical full–stack startup, requires no-stack technologies to do so much. Instead of hiring hundreds of engineers to build out capabilities far from their core business, Uber relies on API-based services to power a lot of their communications. When you locate yourself and request a car, Google Maps helps Uber route drivers to your location (for now). When you receive a text message with a driver en-route, it’s powered by Twilio’s APIs. When your receipt appears in your inbox, it’s SendGrid’s transactional email system.

Discussion about no-stack technologies made me think again about un-bundling PLM services. This trend will be an orthogonal to the PLM verticalization we observed for the last 10-15 years when PLM vendors tried to acquire and combine their platforms into gigantic software stacks.

What is my conclusion? The odds are high that new manufacturing companies will try to buildsolutions atop of no-stack technologies. In that case, sooner than later we are going to see a competition between DIY solutions built by manufacturing companies like Tesla Motors, Local Motors, GE, etc. What will be a future role of existing PLM players in that case? How PLM platfromization trend will go alongside with no-stack reality? Interesting question to ask about from PLM analysts, architects and strategists. Just my thoughts…

Best, Oleg

Image courtesy of iosphere at FreeDigitalPhotos.net