After two reposts, I have finally the ability to write with full speed, and my fingers were aching, having read some postings in the past four weeks.  It started with Verdi Ogewell’ s article on Engineering.com Telecom Giant Ericsson Halts Its PLM Project with Dassault’s 3DEXPERIENCE followed by an Aras blog post Don’t Be a Dinosaur from Mark Reisig, and of course, I would say Oleg Shilovitsky’s post: What to learn from Ericsson PLM failure?

Setting the scene

Verdi’s article is quite tendentious based on outside observations and insinuations. I let you guess who sponsored this article.  If I had to write an article about this situation,

I would state: Ericsson and Dassault failed to migrate the old legacy landscape into a new environment – an end-to-end migration appeared to be impossible.

The other topics mentioned are not relevant to the current situation.

Mark is chiming in on Verdi’s truth and non-relevant points to data migration, suggesting PLM is chosen over dinner. Of course, decisions are not that simple. It is not clear from Mark’s statement, who are the Dinosaurs:

Finally, don’t bet your future on a buzzword. Before making a huge PLM investment, take the time to make sure your PLM vendor has an actual platform. Have them show you their spider chart.  And here’s the hard reality: they won’t do it, because they can’t.

Don’t be a dinosaur—be prepared for the unexpected with a truly resilient digital platform.

I would state, “Don’t bet your future on a spider chart” if you do not know what the real problem is.

 

Oleg’s post finally is more holistic, acknowledging that a full migration might not be the right target, and I like his conclusion:

Flexibility Vs. Out of the box products – which one do you prefer? Over-customize a new PLM to follow old processes? To use a new system as an opportunity to clean existing processes? To move 25,000 people from one database to another is not a simple job. It is time to think about no upgrade PLM systems. While a cloud environment is not an option for mega-size OEMs like Ericsson, there is an opportunity for OEM IT together with the PLM vendor to run a migration path. The last one is a costly step. But… without this step, the current database oriented single-version of truth PLM paradigm is doomed.

The Migration Problem

I believe migration of data – and sometimes the impossibility of data migration – is the biggest elephant in the room when dealing with PLM projects. In 2015 during the PI PLM conference in Dusseldorf, I addressed this topic for the first time: The Challenge of PLM Upgrades.
You can find the presentation on SlideShare here.

I shared a similar example to the Ericsson case from almost 10 years ago. At that time, one of the companies I was working with wanted to replace their mainframe application, which was managing the configuration of certain airplanes. The application managed the aircraft configuration structures in tables and where needed pointing to specifications in a document repository. The two systems were not connected; integrity was guaranteed through manual verification procedures.

The application was considered as the single version of the truth, and has been treated like that for decades. The reason for migration was that all the knowledge of the application disappeared, tables were documented, but the logic was not. And besides this issue, the maintenance costs for the mainframe was also high – also at that time vendor lock-in existed.

The idea was to implement SmarTeam – flexible data model – rapid deployment based on windows technology  -to catch two birds with one stone, i.e., latest microsoft technology and meanwhile direct link to the controlled documents. As they were using CATIA V5, the SmarTeam-integration was a huge potential benefit. For the migration of data, the estimate was two months. What could go wrong?

Well, technically, almost nothing went wrong. The challenge was to map the relational tables to the objects in the SmarTeam data model. And as the relational tables contained a mix of document and item attributes, splitting these tables was not always easy. Sometimes the same properties were with different values in the original table – which one was the truth? The migration took almost two years also due to limited availability of the last knowledgeable resource who could explain the logic.

After the conversion, the question still remained if the migrated data was accurate? Perhaps 99 %?
But what if it was critical? For this company, it was significant, but not mission critical like in Ericsson, where a lot of automation and rules are linked together between loads of systems.

So my point: Dassault has failed at Ericsson and so will Siemens or Aras or any other PLM vendor as the migration issue is not in the technology – we should stop thinking about this kind of migrations.

Who are the dinosaurs?

Mark is in a way suggesting that when you use PLM software from the “old” PLM vendors, you are a dinosaur. Of course, this is a great marketing message, but the truth is that it is not the PLM vendor to blame. Yes, some have more friction than the other in some instances, but in my opinion, there is no ultimate single PLM vendor.

Have a look at the well-known Daimler case from some years ago, which made the news because Daimler decided to replace CATIA by NX. Not because NX was superior – it was about maintaining the PLM backbone Smaragd which would be hard to replace. Even in 2010, there was already the notion that the existing data management infrastructure is hard to replace. See a more neutral article about this topic from Monica Schnitger if you want: Update: Daimler chooses NX for Smaragd.  Also here in the end, it became a complete Siemens account for compatibility reasons.

When you look at the significant wins Aras is mentioning in their customer base, GM, Schaeffler or Airbus, you will probably discover Aras is more the connection layer between legacy systems, old PLM or PDM systems. They are not the new PLM replacing old PLM.  A connection layer creates a digital thread, connecting various data sources for traceability but does not provide digital continuity as the data in the legacy systems is untouched. Still it is an intermediate step towards a hybrid environment.

For me the real dinosaurs are these large enterprises that have been implementing their proprietary PLM environments in the previous century and have built a fully automated infrastructure based on custom data models with a lot of proprietary rules. This was the case in Ericsson, but most traditional automotive and aerospace companies share this problem, as they were the early PLM adopters. And they are not the only ones. Many industrial manufacturing companies suffer from the past, opposite to their Asian competitors who can start with less legacy.

What’s next?

It would be great if the PLM community focused more on the current incompatibility of data between current/past concepts and future digital needs and discuss solution paths (for sure standards will pop-up)

Incompatibility means: Do not talk about migration but probably focus on a hybrid landscape with legacy data, managed in a coordinated manner, and modern, growing digital PLM processes based on a connected approach.

This is the discussion I would like to see, instead of vendors claiming that their technology is the best. None of the vendors will talk about this topic – like the old “Rip-and-Replace” approach is what brings the most software revenue combined with the simplification that there is only OnePLM. It is interesting to see how many companies have a kind of OnePLM or OneXXX statement.

The challenge, of course, is to implement a hybrid approach. To have the two different PLM-concepts work together, there is a need to create a reliable overlap. The reliable overlap can come from an enterprise data governance approach if possible based on a normalized PLM data model. So far all PLM vendors that I know have proprietary data models, only ShareAspace from Eurostep is based on the PLCS standard, but their solutions are most of the time part of a larger PLM-infrastructure (the future !)

To conclude: I look forward to discussing this topic with other PLM peers that are really in the field, discovering and understanding the chasm between the past and the future. Contact me directly or join us as the PLM Roadmap and PDT Europe 13-14 November in Paris. Let’s remain fact-based!
(as a matter of fact you can still contribute – call for papers still open)