A Conversation With David Levin
At the recently concluded Bricsys World Conference 2011 in Brussels, Belgium, I had an opportunity to sit down with David Levin, the Chairman of Russian CAD software powerhouse LEDAS. We discussed quite a few things including his company’s 12 year relationship with Dassault Systemes.
Deelip: I have been following LEDAS for quite a while now. It appears that 2011 has been quite a year for the company. In January you released RhinoWorks, a Rhino plug-in that contained your direct modeling technology. A month later you stepped down from the position of CEO and Dmitry Ushakov took over. In August your 12 year long contract with Dassault ended. In September you sold your IP to Bricsys and “packaged” Dmitry Ushakov along with it. And now in October you appointed 32 year old Alexey Ershov as the CEO of the company. Are these events related in some way? Do they reflect a new direction that Ledas is taking? Are you happy with all these happennings?
David: I can answer in a very radical manner. LEDAS was founded in 1999 but it’s only this autumn we began to really make business.
Deelip: So you mean that earning millions of dollars since 1999 was not business?
David: Of course, my statement is partially a metaphor but still… I mean that previously LEDAS activity had two fundamental aspects poorly compatible with the essence of business. One is what can be called stubborn adherence to developing what you can invent as opposed to developing what you can sell. It is obviously a heritage of our academic history.
Deelip: But you were quite successful in bringing your LGS geometric solvers to the market, weren’t you?
David: It depends on what you call success. Yes, LEDAS implemented and brought to the global market LGS 2D and 3D constraint-based geometric solvers competitive with DCM, the solvers from D-Cubed/UGS/Siemens. Moreover, some customers moved from Siemens to us and our customer base can be characterized as good enough taking into account that we began 13 years later than D-Cubed and obviously didn’t have marketing facilities comparable to UGS or Siemens.
You see, in the 90’s all leading CAD vendors could or had to buy geometric solvers for millions of dollars. But somewhere around 2005 the prices were drastically reduced, by the way, also because LEDAS broke D-Cubed’s monopoly. The result was that we were extremely proud of our capability to build a very cool and market-approved industrial component, which however was requested by very few customers and thus could not bring enough money. Therefore, in the long run, we had to decide to be a cool, recognized but not profitable developer of unique technology or to switch into a business which actually meant producing noticeable income.
Deelip: And what about your direct modeling technology?
David: This maybe is even more telling example because as you know more people talk about direct modeling than geometric solvers. Our own version of constrained-based DM called variational DM was explicitly announced several months before Siemens unveiled its synchronous technology. With its increasingly extended support of design intent, VDM became, by our estimation, at least at the level of a corresponding mechanism of Inventor Fusion. We consider it as more advanced than a DM component of Creo and more prospective than the DM approach of SpaceClaim.
Deelip: Wasn’t that a success?
David: Well, the same as with LGS, we didn’t have our own mass market product, in contrast to all known developers of DM, to embed our innovative component and then sell it as a profitable added-value. A project involving the incorporation of a VDM component into a certain CAD costs about 3 to 5 man years and thus is impressively equivalent to something like couple dozens of annual licenses for LGS. Unfortunately there are still few CAD vendors who want to have direct modeling in their products and most of them prefer to develop this cool component in house. As you know, Bricsys and ASCON agreed to use our technology, and in the long run Bricsys loved it so much that proposed us to acquire it.
Deelip: A while ago you mentioned two aspects that incompatible with the essence of business. One is your academic approach to business. What is the second?
David: It is our long-term contract with Dassault.
Deelip: Do you mean that this contract was not profitable?
David: Oh no, it was very profitable. But actually LEDAS was working like a department of DS with a regular stable financing which some years covered up to 90% of our revenue. In addition, the contract restricted us in relations with the DS competitors. In some sense, it looked similar to being a well-financed and strictly controlled state organization.
Deelip: So were you not happy with this cooperation?
David: No, I was very happy. Let’s say I was very happy with 75% of this cooperation (laughs).
DS gave LEDAS projects of critical importance. This means that they were very difficult and demanding very high competence and skills. So this was a very effective many-year practical training which finally and effectively transformed us from an academic lab to industrial team. This included necessity to constantly look for new methods and technologies or to invent them. This included stringent requirements for performance, scalability, reliability of developed software, severe acceptance tests. This also related to project management in general. By the way, all this finally resulted in a quite funny situation. When we started to deal with our new customers, we looked more committed than them. We began working under strict project management discipline: regular reports and phone or Skype calls, maintenance of a detailed test base, using a well-developed incident management system, and more… In some cases, our customers considered the LEDAS approach as somewhat bureaucratic. But subconsciously or in the long run they understood that it was reliable and effective. As for requirements of performance, just imagine, one day we received a requirement from DS to outperform a component of one of the DS competitors by 100 times.
Deelip: And did Ledas do it?
David: I don’t know how DS did comparison of those two performances but they were very satisfied by our result. The second aspect of cooperation with DS which I very much appreciate is a lot of close contacts with very interesting, sometimes outstanding persons. Note that I mean not only professional aspect, which goes without saying. I also mean general very high cultural and human level.
Deelip: Can you name a few people?
David: Let us begin with the veterans. I have written quite a lot about Francis Bernard, CATIA inventor, the DS founder and its first CEO. We were also lucky to contact with Andre Clemant, who can be considered as a key methodologist for many of current geometrical solutions implemented at DS. Andre was also a university professor of many today’s DS top technological managers. Andre was the first who understood the real potential of LEDAS and convinced the DS management to outsource to us a key industrial project. I believe that Francis and Andre are excellent examples of the profound French culture considered in a broad sense. At the same time, a Russian word intelligentsia can be perfectly applied to them. Philippe Herbert, one of the founders and first employees of DS, while keeping a position of the company’s director for business development, was a real practical founder of our cooperation – a person who is very impressive in many domains.
I am also pleased to mention Pascal Lecland (who is currently at Delmia), Arnaud Rebadeau-Dumas and Xavier Gourdon. They were very important persons in practical management of R&D at DS. Absolutely brilliant guys. Both in outstanding professional competence and effectiveness, general intelligence and in all kinds of communication.
Last but not least, it is worth mentioning that due to our agreement, all these years there were regular 3 month exchange of working visits of the developers and managers from LEDAS and DS. As I could see, people from DS liked visiting Siberia, and of course I know definitely that a lot of visits to Paris and France were very useful and pleasant for people from LEDAS.
Deelip: Now what about 25% which you were not happy with?
David: Before I answer, please remember what I told about absolutely positive 75% and let me clearly state that DS is a great company with outstanding products and solutions, with a lot of brilliant developers and managers. It was a great school and honor for LEDAS to implement critical industrial projects for such a customer. In particular, before partnership with DS, the LEDAS team already had a very high competence in math and software engineering. But those years of intense work improved our competence.
Now about the 25%. First, it’s not surprising that some formal paragraphs of our framework agreement with DS restricted LEDAS from any serious work for any company which can be considered as a DS competitor. Not less important was the informal aspect. Many potential customers who expressed their interest to cooperate with LEDAS finally did not agree because they associated us with DS and were afraid of potential consequences.
Second, my geographical and historical positioning gave, and unfortunately still gives me, a huge knowledge and experience about authoritarian style of management. I have had an experience of contacts with 4 or 5 billionaire corporations, and not only CAD and PLM. I am hardly original in seeing a kind of analogy between some aspects of management in, say, USSR and a Big Corporation.
Deelip: You mean that DS reminded you of the USSR?
David: In all Big Corporations that I worked with, I could find some features of this. Sometimes it was just funny. But sometimes very difficult. Roughly speaking, all of them are quite closed, somewhat suspicious, more or less bureaucratic, inclined to always dominate, be this effective or not, and very much concerned about their image, which not always coincide with the real content. Look, I am not naive. Corporate discipline, image and all this stuff is obviously inevitable in serious big business. But it is really difficult or even harassing when an answer to a simple question you put to your counter partner requires a good month of coordination in some mysterious spheres of a Corporation. It is especially difficult when such Corporation is your main, decisive customer. Besides, as always, any general phenomenon has its concrete embodiments.
Deelip: Could you give me some examples?
David: Again, I ask you to remember my absolutely sincere statements. DS is a great company and LEDAS was lucky to come through 12 years of this very useful cooperation. This estimation can’t depend on whatever particular negative impressions but it can affect the cooperation itself.
One day we asked DS to standardize processing of LEDAS invoices. Not to minimize the time but to transform the processing from unpredictable to just regular in order to facilitate our financial planning. As an answer we received a 15 slide detailed and confidential PowerPoint presentation which described a 7 stage general scheme of invoice processing at DS and thus explained a terrible workload of DS financial department, likely to appeal to our understanding and sympathy. Then, for the sake of determinism, LEDAS agreed for the longest variant of our invoices processing. But it did not very much help. The new procedure was several times broken and caused some unpleasant problems with our Russian bank which is very strict about non-compliance with the schedules specified by a contract.
Now imagine how one Sunday, at three o’clock in the morning I was awakened by a phone call from Alexey, a LEDAS guy who had an early morning flight to Moscow and then to Paris to perform at DS a routine working visit to deliver recent results, pass through acceptance tests, and specify the next 3 month statement of work. Another guy was already waiting for Alexey in Moscow. Both of them were accompanied by their wives because after completing the visit to DS they were going to have a carefully planned vacation in Europe. Alexey told me that he has just received a call from a DS project manager who said that late in the evening the boss suddenly cancelled the visit. The partner did not have any explanations and was very much upset because he was responsible for continuous implementation of this very important project. Alexey was already waiting his taxi to go to the airport and wanted to hear my decision.
Deelip: And was your reaction?
David: I told Alexey to take his flight to Moscow, join his colleague, fly with their wives to Paris and on Monday morning to call on DS and ask for confirmation whether the visit was really cancelled. If yes, they could immediately begin their family vacations. Their expenditures for the flights plus something else would be compensated by LEDAS because they definitely deserved the bonus for their excellent work. On Monday, the guys made the call to DS and by the afternoon the cancellation was cancelled. They spent a fruitful week with a dozen working seminars that enabled to specify the next efficient work package. By the way, a similar occurrence symmetrically happened with a delegation from DS to LEDAS. Their visit was cancelled at the last moment without any reasonable explanation neither to LEDAS nor to those upset DS people who have already packed their luggage.
Another interesting episode occurred during the last year of our partnership. In the process of organizing of one of the regular visits, DS asked LEDAS to enhance the usual delegation by an additional manager because, as they said, it was planned to discuss a very important new development they wanted to outsource to LEDAS. Well, the guy joined our delegation. But in Paris he was invited to a restaurant and proposed to leave LEDAS and become an employee of DS. Our guy found several reasons to decline this honorable offer, and not the least of the reasons was his unwillingness to work with those who can behave so insidiously.
Before we go further, note again my refrain. DS and its main competitors are definitely great companies. And I do not wish them to lose their leadership because of potentially possible enlargement of bureaucratic component and regeneration of their bosses into unchangeable infallible chiefs with voluntaristic behavior.
Deelip: Did DS ever want to acquire LEDAS?
David: Yes, they asked me whether I would agree to discuss this topic. I answered why not. A couple years after this, DS made a due diligence but never send me a single word about their opinion and conclusion. I will never recognize that is normal to work with someone for several days side-by-side on many details and then keep absolute silence instead of saying a simple open no.
Deelip: Did you receive any other proposals on acquisition?
David: I have no right to disclose names but there were very positive assessments of technology and competence made by very big companies. However they did not result in a deal. Maybe the potential buyers felt uncomfortable with a distance to Siberia or they still associated us with DS.
Deelip: I want to come back to the beginning. Were all these happenings in 2011 somehow planned or they occurred spontaneously?
David: For quite a long time we have discussed the LEDAS business development scenarios with several experts well known in the world market of engineering software. In brief, they recommended us to split our business into three parts – working for DS, all other software service providing, and selling our own components. It’s interesting that in magic coordination with our steps into this direction, the environment was making its counter steps to contribute into our plans.
As for DS-LEDAS, both sides felt like spouses who spent together some fruitful nice time but tired of each other, ready to peacefully divorce and start new fresh life contracts. We at LEDAS were just waiting for a formal date when, according to the contract, we had right to notify DS about our decision to stop the agreement. We’ve received a symmetric notification from DS just 3 days before we could send our prepared letter? Very nice mutual sensitivity, isn’t it?
Just at the moment when we decided to emphasize and intensify our service providing direction LEDAS received a very good contract proposal from JETCAM. During one year this contract demonstrated excellent results and has been several times extended. This and some other contracts enabled us to exceed the income level which we had within the DS period.
At last, as soon as we have clearly split our service business from the non-profitable component business, Bricsys made its proposal to acquire the latter one. So decide yourself, whether it was planned or spontaneous.
Deelip: Does all this mean that you have already finished your business transformation?
David: Not completely. We have one more step in mind related to another rather serious potential accumulated by LEDAS during quite a long period but not yet implemented. This is for 2012.
Deelip: What is your strategy now that LEDAS has become purely a provider of software development services?
David: Look, such positioning makes our strategy and plans very simple. In 2012, we are going to at least double the LEDAS average annual income and achieve this both by extending some current contracts and establishing relations with new customers.
Deelip: Could you outline some key advantages of LEDAS as a service provider?
David: I believe that LEDAS is globally the optimal provider of advanced, math-intensive software development. I mean optimal with respect to quality, reliability and price.
Deelip: Are you saying that you offer software development services at low rates?
David: No, we have competitive rates. But you can hardly find a team with considerably lower prices that can produce a comparable quality of software. For sure one can always find somebody who will, in a reasonable time, develop a running program for quite a low price. But a serious customers need something which differs from just a plausibly running program.
Is the program well debugged against a sufficiently test base? Is this test base equipped by a convenient measurement and reporting facilities? Is the test base extendible?
Can a customer see and control a process of development, debugging and testing via a transparent and efficiently maintained incident system?
How is a process of convergence organized? Are the developers competent, skillful and smart enough to find or invent appropriate algorithms that increase a program performance to the level required by application environment?
What is the interface of the developed program with its environment? Is it designed and implemented as a thought-out API which can later be adjusted to changing environment?
Is a developed program scalable? Is it flexible enough to meet changing requirements of its forthcoming life-cycle?
Is a program ready for a non-trivial maintenance? Is the code written in a competent style, with adequate structuring, reasonable mnemonics, non-formal comments? Can it be maintained only by the team of its developers or it is sufficiently and qualitatively documented to be maintained by the others?
All this and more is perfectly mastered by LEDAS along with all relevant instrumental and technical skills. Being a software developer yourself you know all this stuff as much as me but not all customers recognize the problem.
Deelip: Sure. If you pay peanuts, you get monkeys.
David: I can give you a recent example. A company addressed LEDAS with an urgent request. Their very big, although not very cool, program was outsourced to a foreign team of 9 developers. This development came up to a dramatic, or better to say, catastrophic point. A very important customer was waiting for delivery of the initial version but debugging was hopelessly not convergent. To make a long and not easy story short, although the code was terrible and needed considerable competent re-writing, we quite rapidly rescued the company, satisfied the customer, took further development to LEDAS and efficiently proceed with 3 developers for a double rate compared with the foreign team.
Deelip: You mentioned that currently LEDAS has revenue higher than a year or two before. So I highly doubt that you have access programmers on your payroll. In addition, several of your programmers are being transferred to Bricsys. And now you mentioned plans to double your contracts. Are you going to hire a bunch of new developers? Won’t freshers affect the quality of the LEDAS services?
David: The ability to absorb newcomers is one of the decisive advantages of LEDAS. First, note that LEDAS office is located in a 15 minutes’ walk from one of the best Russian universities whose relevant faculties – math, IT, physics – always have good connection to LEDAS. Practically all our employees are graduates from this university. Also, since we are operating in a world-famous research and technological center, Novosibirsk Academgorodok, in the neighborhood of 1 km from us there are dozens of IT companies there is always a certain competition for human resources. LEDAS does not pay the highest salaries but the company is known by its highest irreproachable reputation, friendly human and infrastructural atmosphere and other aspects that attract talented people. For some of them, it is also important that LEDAS always has some projects that require intensive application of mathematics and development of new algorithms, software architecture, and so on. So those who are able and wish can make a thesis and become a PhD. This is definitely not an objective but nevertheless an important indicator. Typically there are about ten PhDs in the LEDAS team of experts and developers. Yes, LEDAS is attractive place of interesting work and personal progress but we always organize very serious testing and hire only the most capable and knowledgeable candidates.
Deelip: LEDAS is well-known by its competence and achievements in development of geometric or broader engineering software. Do you restrict its domain by this type of software?
David: The team which has built industrial geometric solvers and leading direct modeling technology, has obviously proved its ability to develop practically any engineering software. And we would prefer to deal with the domains where we have unique competence and skills that require computational and discrete mathematics, advanced software architecture, advance technology of software engineering, and so on. On the other hand, as we talked earlier, if you in business, you should not reject projects that are feasible and profitable. Today, we have about 40% of our resources allocated in the projects related to data base management systems, data extraction and translation, and not in math or computational area.
Deelip: Can you outline what types of customers you invite to work with LEDAS?
David: I do not see any serious restrictions other than ability of a customer to intelligibly formulate requirements, distinguish between good and bad software, and realistically estimate a cost of quality. I also do not see restrictions for the types of customers. Maybe it’s useful to mention three obvious groups: software vendors, industries and service providers.
We are ready to work with any software vendor. From ASCON to Siemens and from Autodesk to Microsoft.
Deelip: What about DS?
David: Why not? DS would be one of the most welcome and honorable customers.
We have had a very good experience with providing high level software development service to the JSC Reshetnev Information Satellite Systems which is a Russian leading space enterprise specializing in the design, development and manufacture of high performance spacecraft and satellite systems. Industries are obviously inexhaustible source of high-tech tasks. Be they adjustments and integrations of some standard software or implementation of internal particular sophisticated projects that can be of crucial importance but can hardly be solved by standard software providers – especially for a reasonable price.
Deelip: You also mentioned also service providers. Don’t you consider them as your competitors?
David: I don’t know small software providers which can compete with us. As for big or huge service companies, they rarely encounter requests for the projects that need either specific competence or very advanced efficiency or both. We had already received some requests for sub-contracting from such Russian company. Our negotiations failed because the big company wanted too large margin. It’s strange because in their place I’d be much more concerned not about a margin in one of their thousand projects but rather about expansion of their customer base as well as ecosystem of competent partners. Interestingly, one of our current customers came to LEDAS after it failed to outsource its project to that big service provider. So all big companies including service providers have their specific problems. But I strongly believe that we can cooperate with those who know that smart labor division is a sign of maturity and will be glad to include LEDAS into their eco-system. By the way, it would be nice to cooperate in this style with Geometric from India which looks like a very well developed and reasonable company. Maybe you can connect LEDAS with Geometric.
As for the contacts with industries and big service providers, we are currently looking for a competent and experienced person who could represent LEDAS, partially on a high level piece-price base, at relevant active markets, first of all, in USA.
We definitely have no visible plans to extend LEDAS up to a big service company with hundreds of employees. However we will soon have considerable reserves of perfectly equipped infrastructure since we are preparing to move our office to the Technopark of Academgorodok
Deelip: OK, so maybe you can hire a bunch of talented freshers. But to efficiently work in math intensive projects they need to accumulate special knowledge and skills. Besides, more projects need more managers. And I’m sure you know as much as I do that not every talented developer is cut out to be a manager.
David: You know, we have a 12+ years of very successful experience of absorption of novices. The key points are how a project is balanced between novices and people with sufficient labor experience, what initial and subsequent tasks are given to newcomers, how attentively you take into account personal inclinations of a novice and many other factors. In such a process, a role of managers is very important for sure. You may believe or not, but we have the same successful experience of upbringing new managers.
Look at the recent renewal of the LEDAS top management. It brought to the upper positions the guys who are at average 30. However they have an impressive work experience of 8 to 10 years in the most difficult and responsible projects. These guys are have profound expertise in a very broad range of projects and effectively coordinate and supervise their implementation.
I can say that the LEDAS team has the same features as a skillfully and efficiently developed software system. It has good interfaces, it is extendible, flexible and scalable. It works with a high performance, and it is adjustable to new requirements of environment.
LEDAS will never have a shortage in efficient human resources and that’s why I am eager to welcome new customers as we take the company along this new direction.
Ledas’ new office at the Technopark in Academgorodok