LISA: You first came to Boulder in 1979, after working for IBM and TRW. What prompted you to start a translation service business at that time?

Bernard Gateau: I was working in localization—except that we called it highly specialized technical translation back then—without really knowing it: I was responsible for adapting products for European markets, which involved the translation of software. This made me realize that I could be a damn good translator, and that led me to set up the business. Actually, I ended up starting two businesses because I didn't know which one would succeed. The other one was called Polyglot, and produced linguistic tools such as spell checkers. In fact, it was difficult deciding which company to focus on, because early on both had merits.

LISA: What were client priorities back then, as opposed to today?

Bernard Gateau: At the beginning I had so few clients that it's hard to give you a general picture! I started off by myself with one client, a company called Innovative Software in Kansas City which later became part of Informix. I did my first 1200-page translation all on my own, plus everything else such as typesetting, because there was no DTP back then. That made me realize that I needed some employees, and that, in turn, brought more clients.

Client priorities then were "let's get something out to the market based on one specific large order or the need in one specific country." There was no global vision for products. This sounds bad, but it made it easier to get decisions done quickly: projects were not based on market evaluations or what a product could bring in the long term, but on one single criterion.

LISA: What do your clients expect of you today?

Bernard Gateau: I think they expect more innovation, thought I don't think they'd put it like that. As an industry, we have failed to provide enough technological momentum, innovation and sense of direction. Sometimes I feel that we (as service partners) have permitted a few large clients to influence us too heavily. While this has its good side, we need to organize ourselves as a separate industry, independently of the clients we're serving. And we need to be self-sufficient in terms of technical innovation.

LISA: So innovation for you means technical innovation?

Bernard Gateau: Not only. Our industry is essentially based on human beings, which means that the places in any process where you can actually inject good technology are limited. Where you see that this is possible, you can build a new process around the technology while trying not to disturb the human element too much. I firmly believe that the technology comes first; the result is that technology forces you to find out what works and what doesn't. It is a delicate balance to maintain, but I think that you need to look at what technology can do for you and then see how you can use it. If you reverse the order and squeeze the technology into a process that you believe is correct, the result will be too limited. The downside of this approach is that you have to spend more R&D money to be able to select the tools that are going to be successful out of a large number of candidates. You cannot simply say ahead of time, "this is the tool I need." You need to let people be free to invent, and to imagine what could be one.

LISA: What about organizational innovation and structures?

Bernard Gateau: In a very human-orientated business, structure is mostly a way to implement your ideas. If you build too rigid a structure, you may limit your sense of vision. You need ideas first. However—just as with R&D in the software world—you must be willing to accept that not all of your ideas will materialize into good applications. In the localization business we should constantly be pushing our companies toward new visions for future. We cannot be simply a "support group" to our clients. You have to have a vision. And then, on the basis of your vision, you either buy tools if they exist, develop your own if you have to, or encourage other people to develop what you want. This hasn't been done very well yet. We need to give developers more feedback about what sort of tools we need.

LISA: ILE has traditionally been extremely strong in proprietary tool development. You use your tools internally and you implement your technology with your clients. What advantage does this have versus reliance on commercially available tools?

Bernard Gateau: Well, let me start with one disadvantage—you have to invest more money! This means that you have to have faith in what you are doing, and there is more risk involved. The reason why we originally started developing tools was simply that there were very few available commercially. The situation is different now. Today, we have a choice between translation memory technologies from outside and in-house technology. We use whatever fits our needs best. I don't have the "Not Invented Here" syndrome. Of course, the advantage of being able to control your key tools is that you can update and maintain them as needed. You are not at the mercy of a third-party developer. So it is a trade-off. In my opinion, a good solution is a mix of in-house and external tools.

LISA: Where is ILE's development effort now focused?

Bernard Gateau: You need flexible processes that are client-dependent. We cannot avoid commercial tools, but need to make them fit into a modular process; right now, the amount of flexibility we currently have with commercial tools is still too limited. Companies like us, who also develop their own tools, have to provide tools manufacturers with feedback about what we need. One problem is that, since many of the tools developers do not provide localization services, they have a less intimate knowledge of what needs exist and what problems have to be solved. If we can close that gap, we will optimize the development process and this will benefit all users. This is our focus.

LISA: Do you see ILE and the localization industry moving towards high-end machine translation? For example, do you think that MT will replace TM, or, will there be a good symbiotic relationship between the two?

Bernard Gateau: We have never been involved in a partnership with an MT vendor, although it's exactly what I think should happen. Considering the increased volume of translation that is needed, MT simply has to work. The challenge is to convince the translation community at large that MT is a good solution. This is a shift of focus for them, and their inertia has slowed down the number of MT implementations. Whether TM is just a temporary measure to fill a gap between purely manual translation and "super MT", I don't know. But in the long run, the increased frequency of maintenance releases means that the percentage of new text to be translated decreases, and then an MT/document management solution might become more appealing. There again, though, I don't expect any tool to last forever—that would be a childish assumption. TM provides a solution to a particular problem; other problems may necessitate other solutions.

LISA: ILE has been extremely helpful with regard to LISA's OSCAR TMX initiative. How does TMX fit in with Tag?

Bernard Gateau: TMX and OpenTag are, in fact, complementary. Our main assumption when we looked at this two years ago was the reality that large clients would use multiple vendors more and more, and that they would never accept being locked into a single vendor relationship because of a TM or MT format. We needed an exchange mechanism, not just because one client might change his mind and shift production from one vendor to another, but to ensure that the assets represented by translations done over the years could be shared among multiple providers. At a very high level, we needed a way to do justice to the fact that text has become an asset of the corporation and not of the services supplier. This may not be exactly what some people in this industry want to hear, but it's what this industry needs. The growth that will be generated by standard information exchange far exceeds the drawbacks to individual vendors. That was why we developed OpenTag and then got involved in OSCAR.

LISA: What value does OpenTag provide?

Bernard Gateau: If you look at a document, you'll see it's made of two components: the text itself and the tagging. At a sentence level, you may reuse a sentence one hundred percent, but your tagging may be defined at the paragraph level. So what do you leverage? Depending on the challenge you're facing, you may want to optimize your memory at either level. And you cannot impose a specific format on your clients—large corporations with multiple locations have rarely standardized around one single DTP solution. Instead, they maintain a mixture of Word, Interleaf, HTML, XML and many other formats. So our goal was to reuse, as well as possible, both the words and the way that they were formatted, using a common standard format representing both the text and its formatting. This means that you can maintain a TM that is the result of multiple, disparate DTP formats.

LISA: What other areas of standardization do you think could be next?

Bernard Gateau: I wish I could answer your question—I'd have the next great format for us! Of course, there's always the idea of text databases comprising all the help and error messages and documentation in all languages. We are all getting tired of stripping everything out of the product. And we also need a way of handling and reusing English text more efficiently: a high-level document management system. But apart from that, I don't see much that can really be automated in the area of project management and the other issues facing us all the time.

LISA: Let's switch to the business side of things: Where is vendor consolidation heading now?

Bernard Gateau: There has been a lot of talk, but if somebody were to come down to earth today not knowing anything about the industry, would he notice any impact from what has happened these last couple of years? The answer is probably no. Certainly I don't see any obvious manifestations of this consolidation. Yes, there are some larger players now, but is it changing the way we do business? No! I'm waiting for the next stage: after all this money has been spent, we need to see a solid ROI for the investors. And on the client side, the ROI should be faster service, better service, better tools, or all three. However, the jury is still out about whether consolidation is better for the client and for the industry (and it should be both!).

LISA: So where is the money from the venture capitalists, the corporate and private funding, and the investment bankers being applied? How "healthy" is this for the industry—in relation to the attention being focused on it?

Bernard Gateau: The motivation behind investment is vastly different depending on the company concerned. If you take the four or five major mergers/investments done last year, no two of them were really alike in terms of the nature of the money coming in, the reasons for the investment, or the expected results. This is a blessing, because it means we are attracting the attention of a wide circle of investors. However, I'm afraid a lot of these investments were purely done to make a quick killing, and that's not good enough for this industry. Improving your core business and getting it to a size at which it is extremely efficient for your clients and therefore allows you to generate more cash seems to me to be the most logical way to go about investing. I don't want to point the finger at any one here, but I don't know who in the past eighteen months would fit that model. You cannot have investment that is purely focused on investors or shareholders. There has to be an equilibrium, even if it isn't equally shared.

LISA: So, you don't think size is the issue?

Bernard Gateau: I hope not! Of course, depending on your goal, it can be critical—such as if you are a small company needing to reach a critical mass to go public. And you simply cannot go after large contracts if you don't have the capacity to handle them. But size in itself is not sufficient, contrary to popular belief. At ILE, we want to be able to accommodate a greater number of clients, not just to compete with the big guys. We have 200 people in-house, plus a large network of over 1,000 translators who work for us part time. We are expanding rapidly in Asia and will be moving more production to Europe. Anyway, reaching revenues of one hundred million dollars is not that difficult if you do it through acquisitions. But internal growth and acquisitions are two totally different things. ILE is taking a mixed approach. We still want internal growth to be the core component and if we add to this through acquisitions we can easily reach the fifty to sixty million dollar mark quite rapidly.

LISA: So how has ILE been spending its money in the past eighteen months?

Bernard Gateau: We've been opening new offices and attracting senior management professionals from inside and outside the localization industry. One thing that is happening is that this industry has become a legitimate business, and you are seeing more client executives switching to the service side of the business. This is a normal healthy cross-fertilization process. However, these are steps relatively small companies cannot afford to do without outside investment. A lot of the investment proceeds have actually been used where it is invisible to the outside world, focusing on new internal systems and the like. You invest money where you think you are going to save money. You invest in more efficient workflow processes for the corporation, in the backbone infrastructure that will sustain a much larger and global business, because if you expand before the appropriate infrastructure is in place, you just end up with an expensive catastrophe.

LISA: ILE has a very strong reputation as an engineering and localization outfit. Are you going to maintain this or are you moving into vertical businesses?

Bernard Gateau: IT is so vast an area that you cannot look at it as a homogeneous market. There are some distinct vertical niches within the industry. I don't see anything in common between a real-time security system for a nuclear power plant and a spreadsheet, even though they're both relying on some software. If you take a simplified look at the industry we are serving, there are two different models. One is the shrink-wrapped, software package industry and the other model is the large scale vertical application. The time, cost pressures and project complexity of the two are vastly different. While you still need the same basic skills—translation, engineering, testing and the like—you apply them differently at the project management level.

The extreme case would be where localizing a piece of shrink-wrapped software and documentation into 25 languages would cost less than localizing a highly complex software system into two languages. For each project, you're dealing with different client expectations, different lead times and different QA requirements.

LISA: How can we as an industry avoid the peaks and valleys associated with this?

Bernard Gateau: It has been like this for eighteen years! It's the biggest challenge we have, and it's increasing. The size and complexity of products aren't helping, which make it harder to keep a staff of highly specialized people busy. If you look at this problem in the context of vertical applications, it becomes crucial: if you can hardly keep a team of people with multiple and more generic skills busy all the time, how in hell are you going to keep a team of highly specialized people going? Of course, clients want the ultimate in specialization, but they can't afford it. So, in practice, it is a trade-off between ultimate specialization, quality and shorter lead times on the one hand, and cost control on the other.

LISA: Is the client-vendor relationship maturing?

Bernard Gateau: Slowly. Clients have started to realize that the need to allocate resources ahead of time is critical, yet—for perfectly sound corporate reasons—they don't really want to take the risk of securing the time of your employees. So we have to take that risk. And while we now do have better information about what may happen one or two years down the road with some of our clients, there is a big difference between such knowledge and an agreement which says "if you, the client, delay the start of a project by even one day then you're going to have to pay for the idle time that may result." Of course, from the customer's point of view, if you have to take that risk you might as well have in-house staff. Clients contract work out because they don't want that financial responsibility. And as vendors, we have to accept that. To come full circle, this is why I believe that larger companies will have to maintain multiple suppliers, because they cannot assume that a particular supplier will be able to start a given project at a given time.

LISA: Of late, there have been moves within the industry to set up arbitration facilities to deal with complaints. Can you comment on this?

Bernard Gateau: It's sad. Arbitration means that something has gone so wrong that a client and vendor cannot even discuss it. We should be avoiding this, and talking openly about problems earlier during the course of the project. The next question is whether you need special arbitration procedures just for this industry. In my opinion, arbitration can be done very effectively by anyone who is trained and experienced in those techniques, and you don't need to create something separate just for localization matters.

LISA: Do you feel that contracting procedures are sufficient?

Bernard Gateau: They can never be sufficient because of the differing levels of expectations from one client to the next, and because of the difficulty judging that elusive thing called quality. What's more, expectations constantly change between the start and the end of a project. A service business is essentially based on trust. If you don't have that trust, you may as well forget it. Whether you go to an arbitrator or to court, you won't solve the problem.

LISA: Summing up, where do you think this industry will be in three years?

Bernard Gateau: I wish I knew! Increased specialization is certainly one issue: I wish the industry could be large enough for some companies to be known, for instance, as the CAD/CAM specialists. We're not ready for that yet. I also hope we will become a less cost-conscious business. If clients don't allow vendors to reinvest, they jeopardize their own future. Of course, they don't want their vendors to get rich at their expense, but suppliers have to be able to serve their customers now and in the future. Up to now, most localization companies have been driven by entrepreneurs who naturally regarded their company's financial wellbeing as their own as well, with the resulting pressure and responsibilities that this entails. The danger is that the industry will lose that sense of urgency, because of investors with deep pockets. It could result in a more lax attitude toward the best use of funds. As an industry, we have to continue to spend money wisely, regardless of the extent of investment funds we may have.

LISA: Good luck to you and ILE, and thank you very much for your candid evaluation of the industry.