Introduction
"The revolution devours its own children," once claimed the French revolutionary leader Danton when he was sent to the guillotine by a 1794 tribunal; an institution he helped to set up. This applies not only to major social upheavals, but can symbolically represent major PLM decisions. If you fail to land a big PLM bet, the response tends to "eat" those who were responsible.
For Swedish telecom giant Ericsson, that image has become a reality today: virtually everyone who had some responsibility for the company’s 2016 decision to invest in the 3DEXPERIENCE (3DX) platform, from French PLM developer Dassault Systèmes, as the replacement for its old mainframe solution, has resigned, been fired or switched jobs. The installation is, globally, one of the largest in the PLM area and impacts around 25,000 users.
Today, engineering.com can reveal that these staff changes are not just "normal coincidences.” The indication is instead that Ericsson has halted the PLM project and the introduction of 3DEXPERIENCE (3DX) throughout the company. Sources within Ericsson also state that a group of four people has been appointed with the task of evaluating the investment to determine what went wrong and what can be done to fix these issues.
The global Swedish telecom player has also renewed or extended the license agreements for its old IBM mainframe systems, PRIM/GASK. In addition, the previous V5 version of Dassault's cPDm solution, ENOVIA/Matrix 10, has also been retained.
The plans for a big bang installation that they initially decided upon in connection with the Dassault deal have thus been put on the shelf. According to the first statements back in 2016, the new system was planned to be up and running after 14 months. This has since been postponed and today, more than three years later, resulted in a total stop for the entire project.
Those who have read our articles on Jaguar Land Rover's (JLR) iPLM venture, also based on Dassault’s 3DX, can recognize some of the problems, which in the JLR case led to the system not being fully installed even after almost ten years. Instead, the company has relied on updates based on its previous product data backbone, Siemens PLM's Teamcenter. Paradoxically, JLR is one of the reference cases Ericsson's PLM managers visited before making the decision to invest in 3DX.
Those Behind the Deal
When Ericsson's CIO, Johan Torstensson, announced a month ago that he quit his job to become vice president (VP) and the head of a department within the big Nordic consultant EVRY starting in June, it was the nail in the coffin for the group of people who took the decision to invest in Dassault Systèmes’ 3DX platform in 2016. He was the leading and driving force behind the Dassault deal, but far from alone. Here is a list of the most important people in the group who made the decision:
• Ericsson's CEO, Hans Vestberg, had the ultimate responsibility. Vestberg was fired in July 2016 (engineering.com revealed the 3DX investment in April 2016), but lost his job more out of financial, business-related than PLM-related reasons.
• Then CIO, and Vestberg's heaviest strategist, Anders Thulin, was moved from his position in April 2016 and left the company in July the same year. Johan Torstensson succeeded him as CIO.
• Ulf Ewaldsson, Chief Technology Officer(CTO) was the executive who took on the task of leading the way to big economic and business-related savings ("R&D Efficiency") that could be gained through the investment in 3DX. As these potential savings did not materialize, he was kicked aside, appointed to a new advisory role within Ericsson, to later completely leave the company. He is now senior vice president (SVP) for technology transformation in T-Mobile.
• Joakim Cerwall, VP of Group Function Technology and, together with Johan Torstensson, responsible for the specification, purchase and implementation of the 3DX platform, ended his job in April 2017. Today, he is head of research and innovation at Traton Group (Volkswagen Group's daughter company for Scania and MAN truck manufacturers).
• Torbjörn Sonning, "Business Integration Architect" and "PLM evangelist" at Ericsson until December 2016, when he switched over to his own consulting business. He supported a different solution than Dassault’s.
• Fred Skogli, "Project Office Manager", who held the evaluation project at Ericsson, remains in the company with the title "manager".
What Were the Problems behind Ericsson’s PLM Decision?
The question is of course extremely complex and cannot be answered in a few short sentences. There are several reasons but first, let’s make it clear that it is not easy to swap PLM systems in an organization with over 25,000 users (as in Ericsson's R&D organization) working with complex, high-tech challenges in product development. The difference between this venture and others is the unprecedented scale of change within an originally short-term timeframe. The big bang implementation that Ericsson’s CIO, Johan Torstensson, and Joakim Cerwall aimed for was to have the platform form up and running in 14 months. This didn’t happen, however, and, since then, the project has been dogged by continuous delays. Activation of the platform has been postponed several times.
For obvious reasons, neither Ericsson nor Dassault want to have a public discussion about these issues. The matter is very sensitive, not least due to the fact that both companies are publicly traded on the stock market and leading commercial players in their industry segments.
I have been in contact with Ericssons’s press department and asked them six questions to verify the information in this article.
Ericsson’s Eva Andersson, in the company’s media relations department, responded: “Ericsson does not comment on speculation and rumors or internal projects. However, we can confirm that Dassault is the supplier to Ericsson.”
The thinness of this answer is not hard to understand, given the sensitive nature of the issue. The telecom industry is a highly competitive business, and negative publicity related to failures in advanced technology support systems is not something you want associated with your brand. This is especially true if it leaves even a small stain on “the shining high-tech armor” surrounding advanced product development and complex products, like Ericsson’s mobile and fixed networks. Let me also add that Ericsson has plus 95,000 employees and revenues landed at $24.2 billion (for the full year of 2018). So, the company neither confirms nor denies.
Corrupted and Distorted Data
In general, it can be stated that the toughest problems associated with such a system swap are about migrating legacy data; information that already is in existing systems. This is a delicate and difficult task that partly concerns the quality of existing data and partly requires extensive translation and consulting efforts in connection with the migration of the legacy data.
According to my sources, this data may have become corrupt and distorted in part.
“Every time we have tested the migrated data, all use cases failed,” says my source within Ericsson. “Not once has it been possible to migrate old data. This, the CIO [had] been fully aware of when he repeatedly confirmed that ’the timetable holds, we will deliver ...’ This is an extremely important piece [of the project] and taking care of [data] history has always been high on the agenda.”
Admittedly, code has probably been entered into the database to "clean" it before migration, but there are a huge number of error sources that can arise during the cleaning process. Or old errors can be transferred to the new database during migration. The end result is a difficult-to-read mess, containing so many inaccuracies that the system becomes sluggish, inefficient and, in large parts, it does not work at all when items start "pointing/linking" in the wrong direction.
Schisms between IBM and Dassault
What may be even worse is that corrupted data may have disrupted the functionality of Ericsson's Corporate Basic Standard (CBST), which is the telecom player's description of products. If this has happened, it is serious, but we have not been able to verify that this actually happened.
Other problems have involved the integration of downstream systems, sub-systems, which have been found to be considerably more than Dassault was aware of or failed to account for–we’re talking about as many as 400-500 such systems. This in turn has created schisms between the integrator, IBM, and Dassault, where they blamed each other. Who is right or wrong, or what Ericsson’s people missed in their specifications isn’t possible to answer at this stage.
Another important weak spot that doesn’t make things easier is that Ericsson chose to invest in a solution, 3DX, which was not fully developed for product development in telecom. On the contrary, one of the basic concepts in the deal was that the companies would jointly develop this piece within the framework of the 3DX platform. There was simply no ready-made solution, just a "general" set of tools for creating a PLM infrastructure and software platform for the development of mobile telephony communications and network solutions.
No Ready-to-Use Solution, Just Dassault’s Vision and Roadmap
In fact, Johan Torstensson told me in an interview conducted in April that Dassault's roadmap and vision weighed very heavily when the decision was made. It was definitely decisive:
“This is a business change program,” Torstensson said and continued, “One of the most important things in light of this is the road map of Dassault. It looks very promising, specifically regarding the digital transformation we will carry through at Ericsson. In addition to that, as we conduct a change program for 25,000 people in the company, the GUI—the interface—is crucial. To drive the change, people need to feel that they are coming into a modern kind of system. We need to have one common interface, and that’s what Dassault provides.”
That said, he added that Ericsson was looking for a partner. “We want the software supplier—in this case, Dassault—to be our technology partner. Dassault Systèmes has shown that intention from the top leadership, president and CEO Bernard Charles, all the way down to the project leaders and sales people.”
Furthermore, he concluded that Dassault had committed to being part of the transformation journey, which he calculated would take three to four years.
Bottom line: there was no PLM solution specifically adapted for telecom; the big bang installation was supposed to be carried through across 25,000 workers without a supportive, ready-to-use software in place. With these facts on hand, the whole venture appears to be beyond reason optimistic. It’s no surprise that it failed.
The Risks with a Big Bang Strategy
What is a big bang installation about? I spoke with CIMdata president and analyst Peter Bilello about Jaguar Land Rover’s 3DX-based iPLM and he pointed out two major paths that apply to large-scale cases like this:
“One strategy is the ‘big bang’ strategy,” he said. To deploy a big bang strategy,all PLM capabilities need be purchased and deployed with full functionality at the same time. The second strategy is an incremental deployment strategy. With this approach, the capabilities of the PLM solution are deployed in stages or work streams.”
Twowords are key here: ”full functionality.” That seems to be what is needed for a possibly successful large-scale implementation, which was ignored by the PLM responsible in the case of Ericsson.
Bilello went on and stated that CIMdata generally recommends an incremental deployment strategy. This approach allows for a more manageable roll out of capabilities. It can also be executed with a lower level of risk to the business. But there are negative aspects, too:
“Yes,” said Bilello, “It requires a longer-term commitment from management to execute. Companies that don’t follow through (i.e., don’t implement the entire PLM solution), generally fall far short of the quantity of benefits possible if the entire solution set was implemented.” He added, “Sometimes it isn't the technology, nor the design of the solution, but just how fast the organization can accept the change.”
Ericsson’s CIO Johan Torstenson’s approach can definitely be characterized as the more aggressive of the options pointed out by Bilello, rather than a long-term program by program approach.
Finally, there is a critical success factor that needs to be fully considered in this context: Many benefits can be lost if too little attention and resources are assigned to the supplier and partner chain. As Ericsson faced tough internal problems in the implementation process, the necessary levels of attention to this important part of the product realization chain naturally wasn’t possible to achieve. In the JLR case (DS’ V5 platform vs V6/3DX) this was one of the decisive reasons for the extremely prolonged implementation process. What worked internally in JLR with Dassault’s solutions didn’t work in the supplychain without passing a long line of handling exceptions.
Siemens Teamcenter vs 3DX in the Final Round in a Multi-Million Dollar Battle
The acquisition battle that Dassault won at Ericsson took place in tough competition with, in particular, Siemens PLM, whose Teamcenter suite reached the final round, but lost to 3DX. As always when it comes to large orders, the entire battery of PLM developers was present. PTC was there, as well as SAP PLM, Oracle and Aras, but for various reasons they fell off as the evaluation process progressed.
A general observation in this context is that the PLM system vendors originally developed their platforms for electromechanical purposes. This fact applies not only for Dassault but also for the others. This background still produces effects really for all PLM developers, claims Gartner’s analyst Marc Halpern, as a general observation regarding the industry's ability to develop platforms for today’s changed product landscape: “Current PLM software has been cultivated over decades for electromechanical design. Providers of all of the major offerings are on a steep learning curve to support industries where software in products is the major contributor to perceived customer value and product differentiation.”
This is of utmost importance in the telecom industry. In fact between 65 to 70 percent of Ericsson’s revenues comes from the software side of the business.
Furthermore, on the topic of competition: Ericsson’s PLM bet was extremely attractive from the financial perspective. An installation with 25,000 users is worth a lot of money—not including the prestige of having its exclusive logo on this customer list. In the case of the Ericsson and Dassault deal, the amounts were speculated to land in the area of "several tens of millions" of dollars. However, the contract with Dassault was just one part of this amount, probably a bit lower than normal, since it contained common software development ventures. On the other hand, integrators and implementors, like IBM in this case, don’t work for free.
Dinners with The-Chairman of the Board for Ericsson
As concluded above, Dassault won the Ericsson deal. The French PLM developer’s vision and roadmap played an important role. But there were other, more non-technical reasons: "It didn't hurt either that our chairman of the board (it was Leif Johansson at this time) knows Bernard Charles well and that they have eaten several dinners together,” one of the members in the PLM group told me.
In itself, no one can accuse a PLM player with high sales ambitions of targeting a potential customer company at several levels. On the contrary, it can be a winning strategy to process the customer potential in parallel at several levels; from the chairman of the board, via the executive managers to other influential people at C-level and lower in the organization.
A problem encountered in such a broad attack strategy may be, for example, that operations managers, who are satisfied with things as they are, neither have the time nor the desire to change a functioning model, despite what long-term needs may present. No doubt, by communicating with, for example, CEOs and important board members over time, the organization can be pushed in a more favorable direction toward new and increasingly profitable systems and models. It may not pay off in the short-term, when new systems and models are implemented, but in the long run it can be paramount to cope with challenges associated with a PLM system swap to secure the future. Clearly this applies to an industry with often rapid development cycles, such as telecom.
The big challenge in this context is the balance between preserving what works and the importance of meeting and taking in new technologies and systems. It requires a delicate touch and a sophisticated ability to navigate between the pitfalls.
In the case of Ericsson, one can generally say that the visionary pieces should not have been given as much importance as they ultimately got. Many players on the vendor arena can paint a “nice visionary portrait”, but few can build up technology that meets the vision and creates deliverability around given promises and capabilities on this scale.
Furthermore, the migration problems surrounding legacy data also seem to have been underestimated.
Likewise, the parameters Ericsson used in connection with the decision—in which the DS roadmap and vision had a very important significance, perhaps as much as 50 percent—entailed risks such that the specification set up contained nontechnical and imprecise requirements and where much was left to be resolved in coming developmental situations: "Such things can be solved later.”
What Happened with Promised Cost Savings?
It is also clear that the leading ”PLM-evangelists” in Ericsson promised large cost benefits in terms of PLM. This is in itself nothing strange. Lower costs are always good. There is also nothing that sounds more attractive to the ears of a CEO, board member or owner than promises of radical cost gains. Measures that bring lower expenditures to the table tend to become primary drivers, rather than the technological achievements that can be made; although everyone realizes that the latter are important.
In parallel the CEO often does not have the deep knowledge and insight needed to evaluate the effects of a new PLM system. The same applies to the chairman and board members. They are in the hands of the responsible officer or committee and are forced to rely on the accuracy of the evaluations and recommendations presented by this executive or committee. If one can get a system that reduces the product development work, while it sharpens the company's capabilities and the quality of the products to be developed, the case is planted in nutrient-rich soil. Especially for a company like Ericsson, which was hit by tough financial difficulties with declining sales and profits that had to be improved to create a good mood among the shareholders and on the investment markets during the years around the big PLM investment decision. However, it should be pointed out that the company, under new president and CEO, Börje Ekholm, has shown a positive development in recent quarters.
A State-of-the-Art "Beyond-PLM System" Worth Every Penny for Ericsson
In summary, according to my sources, development ended with the 3DX project stopped. But measures have been taken to resolve the issues, among other things through an evaluation group of four individuals: Jonas Wigander (BDGS), Mikael Andersson (GFTL), Mats Eklind (BNEW) and Manuel Carnedo (DTO).
It remains to be seen where it all ends. Clearly, however, the planned big bang has failed.
First of all, it is now a matter of giving the organization clear structure over who should do what and what should be done.
It is also obvious that the savings that were anticipated as a result of the new system will not be achieved. On the contrary, Ericsson can expect that the resources needed to get the solution in order will land at considerably higher levels than what was originally planned for.
In this context, it is also relevant to recall Albert Einstein's old words of wisdom: “We can't solve problems by using the same kind of thinking we used when we created them.”
According to this thesis, it would thus be a good idea to make a whole new take. An optional path is to replace the main supplier and choose a more industrially proven solution, in which case Siemens’ recently-acquired Mendix software is an interesting option that could provide a non-political way out.
Regardless of which route they chose to go, it is important to anticipate higher costs than expected. On the other hand, the investment in a PLM system is small compared to Ericsson's revenues. And given that the system is one of the cornerstones of product realization, it is a small investment that can strengthen Ericsson's role as one of the world's largest and most important providers of mobile traffic systems—not least important in light of the next generation of high-speed 5G solutions has gained considerable momentum.
According to CEO Börje Ekholm, Ericsson has so far been the market's most successful player and has brought home more than half of the contracts signed globally so far.
In the short term, one can continue to use all or parts of the old mainframe solution and produce potent products. But over time, Ericsson must invest in a "beyond PLM system,” a functioning Product Innovation Platform, as defined by analysts Gartner Group and CIMdata.
It is a tough but feasible task.