The flames on Cloud PLM van make it go faster.

Cloud PLM Isn’t Cloud PLM

Some might surmise that if software is eating the world, then perhaps cloud software is eating conventional software. But is this true in the enterprise? Judging from the explosion of cloud-centric solutions that have entered the market over the last few years, the answer must be an unequivocal yes. However, when we narrow the scope of the discussion to Product Lifecycle Management (PLM) there’s a perceptible shift in tone, the hand wringing and teeth grinding and grumpy grumbling commences. Maybe there’s some concern over who’s responsible for downtime or performance bubbles. Or what it really means to have your IP slightly out of your control, but not (in theory). Yet every major vendor in the PLM space has some type of cloud-flavored offering on hand. But is it really? Depending on the particular vendor, attitude toward offering cloud PLM ranges from “What, you mean there’s some other way besides cloud?” to “I guess we have to have cloud in there somewhere or people will make fun of us.” But wait, are we talking apples to apples? Is cloud PLM really cloud PLM? Oh hell, no.

The Twittersphere (Twitterverse?) was alive with some interesting comments on this topic from the “What’s Shaking in Product Development” Panel from this week’s Accelerate PLM 2015, which is Autodesk’s second PLM-focused conference. My how times have changed. Of particular interest were reactions to Question 3 (Thanks to Oleg Shilovitsky for keeping up with the convo):

Cloud PLM products are widely seen – how do you differentiate between different platforms?

For which responses were indeed quite interesting:


There’s a consensus here: software with largely conventional architectures that have been adapted to some type of cloud delivery / storage solution must be differentiated from software purposefully architected for the cloud. Architecting a platform in the cloud changes fundamental assumptions about how software should operate, by which means users interact, how data is stored, decimated and mashed up, and how the software interacts with other software. Taking classical on-premise servers and swapping them out with virtualized cloud equivalents certainly reduces your physical footprint, but does little to evolve the product.

Cloud PLM is not repackaging what already exists. It’s a do over. To do otherwise is akin to taking an old minivan suffering from touchy brakes and a leaky oil pump, painting flames on the side, and claiming you have a new sports car.

But wait, we’re not done yet. You can’t go halfsies on adopting the cloud either. Cloud PLM means cloud ERP, cloud ECM, all of it cloud. That means abandoning legacies and substantial portability and/or compatibility with existing PLM solutions. Not to mention being comfortable with the security implications, both good and bad. Otherwise you’re negating the whole point of cloud to begin with, by frankensteining shiny new architecture with the same old stuff. If you’re going to do that, you might as well stay where you’re at and be happy. It would certainly be a lot less trouble. Lukewarm sales of most cloud-flavored variants of existing solutions is evidence enough. The thought also summons a specter that customers don’t like talking about: migration. And we’re not talking geese, either. Some customers already have enough migration fatigue, thank you very much.

Existing vendors have a dilemma however, they know winds are changing but they can’t just recklessly fly up into the cloud, expecting everyone to happily come along. There are plenty of customers who aren’t interested. Good customers. Long time customers. High paying customers, even. Some of those customers, much like SAP, may be hiding under a desk waiting for this whole cloud thing to blow over. Their solution in ten years may be to get a bigger desk.

Which returns us to the basic question about cloud PLM: How do we differentiate between conventional PLM with cloud flavor and real honest-to-monkeys cloud PLM? The former is an attempted appeasement to the existing market. The other is not. And that’s the point, cloud PLM for right now is a new market for new customers. For the most part those making the jump are companies that up until now wouldn’t give PLM the time of day. Which is why Autodesk is doing well with PLM360, their product is a natural growth path for companies just now thinking about the fundamentals of PLM. Of course everyone likely wants a piece of the action – but they’re not going to accomplish that with what they already have. This is PLM’s new disruption.

  • Thanks for the discussion, Ed. The challenge in determining what’s “cloud” versus “not cloud” is that definitions vary. Manufacturers need to understand what they want from the cloud. If they just don’t want to own and operate servers, then hosting accomplishes that. Is that “cloud” or not? I guess it doesn’t matter if it meets their needs, right? But if they want the full benefits of cloud (economies of scale, etc.) then they need to look for more of a “true cloud” approach. Here are some additional thoughts:

    • Thanks for your insight, Jim! While the advantage of moving on-premise hardware to the cloud as is makes total sense in theory, how well does it translate in practice? Adoption rate from established customers has been rather slow from all data I have seen. Most customers that are deep in PLM have spent considerable money on it; the cost of server hardware and maintenance is not the determining factor demanding change. Combine that with the typically long sales cycles typical of PLM, and it’s easier to understand why existing customers don’t seem to moving over in droves. Now new customers may be a different story provided as you rightly mention that the vendor is offering a real cloud solution.

  • Hi Ed,
    I’m wondering if PLM disruption is not meant to kill PLM solutions, replacing them with PLM app stores. I don’t think any of the big ones wants to go in that direction (maybe PLM360 is the closest for now). To me the big issue today in PLM for customers is the data lock-in. It’s too expensive to switch from one solution to another and create a real competition.

    • Agreed on the data lock in, Yoann. I’ve talked a bit about resistance to the marketplace model here: There’s an interesting dynamic with marketplaces, looking at the phone and gaming (Steam) marketplace reveals some difficulties – one being that the marketplace owner does well, sometimes at the expense of the app providers. So I imagine everyone wants to start their own market, and you end up with similar data lock in – just with slightly more flexibility. How would a truly independent and universal market look like, and who would run it? The discussion might be worth another post.

  • Pingback: Accelerating the PLM Cloud Integration Bus | E(E)()