My interview with Wordbee


TF: 00:07 Welcome to the International Buzz podcast by Wordbee. This is your co-host Tanja Falkner.
RR: 00:13 And I’m your co-host Robert Rogge.
TF: 00:16 For today’s episodes, we invited someone who has been in the industry since 1982. He is a translator, localizer, a technical writer, and he’s even been running his own consulting business for 16 years, called sQuid. Not only that, he was also teaching terminology and localization at university in Rome, he’s written several books and the articles and he’s still speaking regularly at conferences and holding webinars and workshops. Today’s guest is Luigi Muzii. Welcome Luigi.
LM: 00:48 Hi. Hello everybody. Thank you for inviting me.
TF: 00:56 Thanks for being here.
RR: 00:54 Yes, no problem. We’re super excited to have you on the show. So, I mean to open and put the first question as I don’t want to take credit for this hard-hitting question. So, that was definitely Tanja who had the idea. OK, so you have a reputation for being particularly critical. Someone says that you’re perhaps a contrarian, although I’m not so sure about that. In your article Reality check there’s this line that is just amazing. I’m pretty sure it was from the Reality check article. It says whether a translator, an LSP or an advisor, a bullshitter is always there, mouth breathing, his or her words like mere vapor, which is awesome. So, we were just talking before we got started and it was almost impossible to match your voice in person to the voice in writing. So, I guess we were wondering how you see yourself and have you ever had any consequences or problems for being so critical or the tone of your works?
LM: 01:57 I’ve always had problems with my opinions because I’m not really a contrarian as someone would like to define me, but I’m simply an outspoken man. I’m not really afraid of mouthing my opinions even when they are against the common thought. My latest book, which is a collection of my blog posts over the years, is titled Upstream because I never liked to be mainstream. And the translation industry is a mainstream industry. If you have the chance to browse the blogs and articles and even books in this industry, no one is taking a stance, a clear stance against something or someone that could harm or compromise his own reputation or job and the only stances are taken against customers or agencies at large, so the abstract figures that no one can identify precisely.
RR: 03:20 It’s always someone else, right?
LM: 03:21 Yes, but this is a huge problem for the industry because in this way, there’s no competition, no real competition. All profiles are flattened out, so everyone more or less looks the same as the others. And in a way this reflects even, for example, in the usage of quality as the USP of the industry, the unique selling proposition of the industry. So, there’s no real differentiation. When I wrote the article you are referring to, I wasn’t actually expecting to receive any real reaction because, what I was thinking at the moment was, “OK, well, I’m going to write this piece and people will say, «Oh yes, the same old Luigi as always, the same old guy»”. So, find a word starting with an A and ending with a hole.
RR: 04:19 Oh, no! We don’t think at Wordbee, definitely.
LM: 04:21 Of course not. I don’t specifically mean you, but I’ve been actually insulted more than once over the years for my opinions, for having been outspoken. I actually thought that that article would be received just like any other article I wrote in the past. And I was quite surprised that it was not.
RR: 04:52 Are there any particular actors in the industry that you think are doing a good job of being open-minded with how they communicate about things and being perhaps more self-critical or more honest or is it pretty much everybody?
LM: 05:07 One I never met, and I would possibly never agree with him, if he were still alive, was a Spanish financial translator that died a few years ago. It was just as outspoken as I am but on opposite side compared with me. So that was a really intellectually honest man. There are also other people in this industry that do a very good job, very good people. I’m in very good relationships with most of them, and I could name over a dozen although they, for one reason or another, even when they tell me that they appreciate my articles, that they share my opinions, when it comes to offering me some job or referring me to people for consulting, they’re not the same willing because I’m scaring. I am, but I don’t know any other way. I was just grown and educated like this, to be honest, not to be a hypocrite. So, I grew up thinking that hypocrisy can put one’s work and one’s career at risk rather than being profitable.
TF: 6:39 Was there ever a time or do you recall a specific case where you think maybe you shouldn’t have said something in particular or maybe something that was not so great for your business? That you would change?
LM: 06:58 I’m going to answer you with a song. One of my favorite song, a most famous song, Non, je ne regrette rien, I don’t regret anything. I have no regrets. So, it’s hard to me to find a day or time or whatever that I could have said something that I would not say again.
RR: 07:22 Is that the original for Frank Sinatra’s My way?
LM: 07:28 Well, first of all, it was not Frank Sinatra’s. I was not even a Canadian singer-songwriter’s whose name I can’t remember right now. It is Edith Piaf’s. It is a great song. It was the Legion’s song. It is the song the legionnaires were singing when leaving Algiers. It’s a wonderful, wonderful, wonderful song. You have to pay attention to the lyrics.
TF: 08:03 Definitely. I liked the idea of just not having any regrets.
LM: 08:08 Once you’ve done anything, it’s done so it’s no use in having any regrets.
RR: 08:18 Totally agreed. I thought you were going to sing a few bars for us for a second, and I thought yes, that’s going to be the first time someone sings on The International Buzz.
LM: 08:26 Oh, dear no! Actually, my French is even worse than my English…
TF: 08:28 OK, next time then.
RR: 08:35 We like the stuff that you write regardless of tone and at least it’s interesting. It’s provocative. It stirs the soup. And you have usually a pretty good idea of what you’re talking about, I think. That’s my personal assessment. So, like moving on, we’re going to transition sort of pivot towards technology and language technology and to kick off that part of the discussion, we have a quote here from your website that says technology should be functional. It must integrate in processes, not shape them. However, processes must be designed according to the technology that is expected to be employed to manage them. And we were wondering if you could unwrap that concept for us.
LM: 09:16 Yes, that’s the curse of translation technology and of translation technology companies, because the problem with translation technology in particular, and with technology in general, is that we expect technology to solve our problems. Most problems do not require technology to be solved. To be solved they just need a different approach, a different state of mind, and to be capable of identifying the issues in a process, try to correct them and move on until you have a perfectly working way of operating, and then find the tools, not just one, not just a platform, but the tools, they could be many, that fit into your way of working. Technology should fit you as a suit, not the other way around. And that’s why we have this problem with translation technology because all those platforms and tools simply replicate the century-old process of every freelancer, every translation company, even every translation bias, and so that’s just one thing that I would like to tell everybody in this industry and it is not to be obsessed with the technology. Technology must be useful, must be functional. If you get trapped, if you get caught with technology, that means that you have been having problems before adopting technology.
TF: 11:09 As you said, we don’t need technology to solve problems. I get that, but don’t you think that technology makes it easier to solve them? Even though we could do without?
LM: 11:23 Definitely, but if you rely solely on technology to solve your problems, there’s something wrong in your approach.
RR: 11:28 Sorry. Do you have any specific examples of how technology should integrate into processes and not shape them? Like something specific about a particular tool that you’ve used or seen that you’ve tried to achieve?
LM: 11:39 Well, a couple of days ago when I was in Athens, one the attendees asked me whether she could exploit my presence there and have some piece of advice for free. And I said, OK, and she asked me which is, in my opinion, the translation tool that offers the best quality-price ratio in the world right now.
RR: 12:08 And you answered Wordbee, right?
LM: 12:09 No, I answered that her question was wrong because there’s no one tool. There’s no one tool and no one platform. There’s just the tool or the platform that best fits you because you are unique with your processes, your education, your knowledge. I could give you another example. I have been fighting for years against quality standards in translation. In 2001 or 2002, I can’t remember right now, CEN/BTTF 138 task force was established to draft the new quality standard that later became the EN 15038. I was in the Italian committee. I asked the Italian committee not to be chosen to represent Italy at general meetings, but I really worked hard on my side, and I remember that we Italians, the Brits and the Swedish were the only three groups pushing for adopting ISO 9000, for drafting a standard that fit into ISO 9000 and forging a certification standard specifically for translation. So, at the time ISO 9003 was the only standard that could fit the needs of the translation industry. So, we Italians, the Swedish, and the Brits said, “OK, just use the existing standards.” The objection was that ISO 9000 was too costly and hear, hear too stiff, too rigid, not enough flexible. Well, the outcome, EN 15038 and even the following standard ISO 17100 were both even stiffer, even more rigid than the original ISO 9000. In fact, ISO 9000 has been made more and more flexible and lean over the years. So, the icing on the cake is that the requirements of ISO 17100 are more or less the same as EN 15038, and if you are willing to pursue certification you will have to certify to ISO 9000 because we still don’t have a certification standard for ISO 17100. And the same goes for ASTM F2575, ISO/TS 11669 or the new ISO standard on post-editing; they all just freeze the same modus operandi that is a snapshot of the way of working that the whole industry has been following for decades, for centuries even. This approach is replicated in technology.
RR: 15:22 I think that kind of relates to your article from 2014 called The red pen syndrome, where you talk about TEP and in that article you suggest trying to be more agile to try to replicate agile processes instead of these classic processes. And you also mentioned checklists. Can you go into some of the sort of lateral thinking that you think would help the translation industry escape from these age-old processes and why they might work?
LM: 15:54 Agile is another buzzword in this industry, because there could be no agile with TEP as it’s a serial process and it allows for no circularity, no parallelism, no parallel processing and so on and so forth. Also, we have no way of measuring anything in translation. KPIs are still very little adopted in this industry. We have no specific translation-related KPIs. Probably some LSPs have their own way of tracking some KPIs but there’s no shared effort to draw some KPI that could be useful for the industry, to measure the efforts done by everyone in this industry, so that’s just one of the many problems we have. We still have the old-fashion way of measuring quality that is measuring errors, when we have no common way to define errors. We have error types, class of errors, importance and levels and so on and so forth. ISO 9000 came out for the manufacturing industry, to answer the need for standardizing processes, not to make every company work in the same way, but just to give companies the same basis to work on. One of the most serious problems during World War One was ammunitions because there were no standards for ammunitions, for projectiles, especially for light weapons, but even for artillery. So, the number of accidents were impressive. So, the need for standards came from the need for reducing errors and this process evolves and now gets to the Six Sigma, for example, or the CMMI where the focus is no longer on catching errors but in preventing errors. We are still at the error-catching stage, but we have no common definition for errors. It’s a typical school teacher approach, where the school teacher says, “This is a serious mistake, so I’m going to use the blue pencil. This is a minor mistake, and I’m going to use the red pencil.” And so on and so forth. This doesn’t help the industry progress. We are stuck in the past.
TF: 19:07 Don’t you think that technology could help here? I mean, when you say there is no common solution?
LM: 19:20 Yes, but if you want to develop a software to measure quality, what would be the base for measurement? The metric. Again, catching errors, so you have to start all over again. It’s a vicious circle. Anyway, this is my idea.
RR: 19:38 Is there any new technology in particular that exists today that you find particularly promising in those regards?
LM: 19:44 Oh yes. Now that machine learning has become a reality there could be many, many potential advances in this area. Well, I have to admit that a good friend of mine took some advice I gave him and is making a good job with his platform. I won’t name this platform here, it wouldn’t be nice. But machine learning could be a path to go and try a new way of measuring. For example, TMS producers might think of a way to allow their users to customize their KPIs, but this again would require standard metadata, and this is something we don’t have yet. This is something that is going to be dealt with inside TAPICC, the pre-standardization effort the GALA association has set up, but, again, as a pre-standardization effort a consensus has to be reached on the kind of metadata to use in this industry that would allow the players to reckon, calculate, and process their own KPIs and have a common way to display them to buyers, so that, when talking about quality, quality really means something.
RR: 21:28 Just out of curiosity, have you ever use Wordbee, or do you know anything about Wordbee?
LM: 21:33 Yes. I must admit that I most often disguise myself when registering for a trial account. I actually explored Wordbee the same I did with other platforms, like Memsource or Smartling or other platform of the same kind. I must say that the feature I like most in Wordbee is the different approach with respect to freelancers and LSPs. What I mean is that even though the platform is the same, freelancers can use it in a totally different way than LSPs may use it. So that’s very important. The user interface is some way different. That’s the thing I like most in Wordbee. The only problem is that you still have to rely on a reliable connection and that’s could be a problem even in some areas here in Italy or in Spain or in France.
RR: 22:55 The kind of the challenge I think at Wordbee right now is because some of the things that you mentioned are true, but then some of them are not true. There’s so many features in Wordbee that it’s hard to sometimes get them out there. For example, you can export a bilingual Word file and you can translate it in Microsoft Word with Wordbee or you can export an XLIFF file and work offline. Same with KPIs. We have a pretty thorough set of analytics in Wordbee, whether it’s like time to delivery or how long it takes a translator to accept a project, which translators turned them around the fastest, the throughput of any given project manager. How efficient are each project manager projects…
LM: 23:46 Anyway, the thing I was describing a moment ago about the different approach for the single freelancer and the LSP is one of the things I would like to see in most TMS or most translation environment tools in general. When you are an LSP, you may want to use a single platform so that when a vendor manager logs in he can see a different user interface, exposing different items, different features, even though there will be the same database shared with project managers. In fact, as a project manager you should be able to work on yet another different user interface, for budgeting, vendor selection, or planning. This would implement real collaboration between business units. So, what is the job of a vendor manager? It is not just recruiting, it is vetting, it is keeping in touch with vendors, keeping track of changes in a vendor’s experience, specialization and so on and so forth. This is the sort of things that a project manager should not be interested in. He should be interested in seeing a list of candidate vendors and pick the one that best fits for the project. How? Based on rating, which should come from the activity of the vendor manager and from objective data. So, the KPIs that are relevant to a specific vendor, specific freelancer or whatever, and that’s the kind of thing that are now quite easy to achieve with a machine learning algorithm. They would have required thousands of lines of code in the past when you can just rely on SQL. So, it’s supposed to be much easier. Now.
TF: 26:09 I think what you just mentioned before, it’s kind of relating to an article you wrote at the beginning of last year where you said that you think that TMSs are generally lacking the feature of fully workflow management systems in terms of that they need, you know, when you mentioned the project management, but also about a management resource allocation and that sort of stuff. How come you don’t see the full workflow management in today’s TMSs?
LM: 26:40 Well, I must admit I spent many years in large companies with plenty of resources and, in some cases, spending a lot of money on software, and I have the chance to learn from the greatest software companies in the world, from 3D database management systems to real time database management systems and so on and so forth. So, the things I wrote in my article are actually a reflection of the things I have seen in the past, when they were just affordable for a very limited number of companies. I’m talking of the years before the year 2000, but now they’re quite cheap even for a small-size technology company, so I’m surprised, not just because I think that these technology companies have not the knowledge and the experience to develop those kinds of features, but just because they are trying to anticipate their prospects who belong to this industry. So, these tech companies develop the typical attitude of the industry. The one that we were talking about a few minutes ago. I don’t blame translation companies in this respect.
TF: 28:13 Do you see this changing towards being more of what you describe as a fully working workflow management system as they used to be?
LM: 28:34 It could be. That could be a starting point. Yes. I’ll give you an example. I was attending and speaking at the first memoQ fest back in 2009, and at that time memoQ was a wonderful, wonderful tool because it was lean, light, agile, fast. It could still be missing something, but it was really a novelty. But the reference at the time was Trados. It was not yet SDL Trados. I wouldn’t say it would be easier, but it could mark a difference. I’m still a memoQ user, for the very occasional translation projects I could be asked to do, but it is now clogged with features that I can do without. I mean my dream with software is the package that makes me say, “OK, this is the whole bunch of features I have in my package. So, let’s start with this one.” That is the core, the kernel, like that in an operating system. So, the kernel is launched, and then the module you have selected will add on the kernel and you have a very, very light environment, very light tool that is lean and easy to use.
RR: 30:07 Like a granular control over what you see and what you don’t want at any given moment.
LM: 30:13 Yeah. That was what I was hoping to see in web-based tools because you can’t clog a web-based tool with too many features that you can’t have the resources to handle.
RR: 30:30 So, you should check out the new Wordbee editor actually because I think it does.
LM: 30:35 I did, and I liked it. I must say I like it.
RR: 30:37 Have you seen the new one though? The one with the colors in the background and all that?
LM: 30:42 No, I must say not the one with the colors.
RR: 30:46 Yes, there’s a new editor, which I’m not sure when it’s available. Tanja do you know if the new editor is available right now? Like if a freelancer signs up?
TF: 31:01 Yes, I think so. I believe you can choose. You can still choose between the old and the new version.
RR: 31:03 It’s worth looking at because I think they’ve made a decent effort at exactly what you’re talking about, about trying to minimize a little bit the interface and not overwhelm the translator with too many tools and bells and whistles and stuff. So, you can do, in Wordbee, anything you can do in memoQ and probably more. But that doesn’t mean that there is a clogged interface, they’ve tried to hide it. In today’s age of single-page applications it should be totally possible to have a feature-rich interface that doesn’t overwhelm the user.
LM: 31:43 Yes. You know, what I love most in smartphones is that they help software developers a lot slim down their applications. Even Windows has profited from smartphones. In fact, Windows 10 is now a very smart and fast operating system. It is not yet the perfect operating system but it is much better than its predecessors. I’m used to say that we’ve got real super computers in our hands today. When I started working in this industry, in 1982, I was using a typewriter and I considered myself lucky because I had a portable Olivetti Lettera 22 typewriter that was really, really light at the time even though it weighed over three kilos, so it wasn’t really profitable. My first PC had 64 KB RAM, and a 20 MB HDD; it was a few years back, but not that many. I bought my first PC in 1988. So, the exponential trend we have been facing since the Apollo program is bypassing us in the translation industry. It seems like we’re still at Gutenberg’s time.
RR: 33:32 Well, we’re working on it at Wordbee. I don’t know what else to say. I don’t think we could pump out features any faster than we are. I think if we had 20 more developers, we still couldn’t pump out features faster.
LM: 33:50 Again, I don’t blame translation tech companies. I think they are trying to give the people what they want, and, in my opinion, people are not asking for the right things. I mean, they should be asking for something different.
RR: 34:07 Well, that’s true. Customers don’t always know what they want. That happens all the time actually. But let’s go ahead and start wrapping up. So, what’s the most controversial opinion or controversial prediction that you can think of at this moment?
LM: 34:23 Predictions.
RR: 34:28 Well, controversial ones?
LM: 34:33 No, just predictions. I think it was at Danish comic writer who once told that predictions are always hard, especially when they have to do with the future. This is a statement I like very much. It is always attributed to Niels Bohr, the physicist, but this is one of the favorite sport in the industry. At the beginning of every year, you can be awake all night to be counting the complex predictions for the coming year. The only prediction I could make is that we will continue to see mergers and acquisitions in this industry and this is because the organic growth is becoming harder and harder every day.
TF: 35:25 That’s an interesting prediction and we’ve actually just talks about merges and acquisitions in one of our last podcast episodes.
LM: 35:30 Yes, I listened to it, with Renato Beninatto. I know him, and we have different views on that.
RR: 35:38 The very last podcast we did, we talked about M&As with Peter Argondizo from Argo Translation and he’s going through the process of looking for acquisitions to make. So, I think with Renato is a little more like theoretical, but with Peter we got into the nitty-gritty details of how you go about trying to make it. I totally recommend listening, it was a pretty sharp podcast. Why do you think there’s so little M&A going on in language technology? For me, I started working with Wordbee first, with my previous company back in 2009 or something, and then I started collaborating with Wordbee in 2011-2012. And, like in other industries, it seems like you make a good tool, because Wordbee is a pretty good tool, they have huge customers, like Air France, they got huge enterprises, they got small companies, they got off a great business and I would have expected that someone would have said, “It’s about time to go buy Wordbee, and one of these big companies would come along and buy Wordbee. And I would have expected the same thing to happen like in other industries it happens to pretty much anybody who makes any tool that works. So, why do you think there’s not so much M&A going on with technology providers in particular?
LM: 37:00 Because of the attitude, the general attitude of the people within the translation industry. As long as translation technology companies are still looking more at translation industry players rather than at translation industry customers, the translation bias will always lead people to underestimate this kind of technology. The greatest and the worst thing at the same time that Google Translate have done to translation technology was to popularize it. But when you talk about translation technology to everyone outside this industry, to any outsider, the only thing this outsider can think of is machine translation in general. This reminds me more or less of the attitude I dealt with at the beginning of my career when I was asked, “What do you do?” My answer was, “I’m a translator and a localizer.” And then the talk went on like, “You what? What do you really do?” “I mostly localize software.” “Please, tell me more about it.” And so, I had to go into details to explain what I did. Then the next question was, “Which kind of studies, which kind of education you have?” Most people were astonished at knowing that I have no formal education in languages, no formal education in technology, no formal education in engineering, but I that I “only” made some specific experience on the job. And when my translator friends said that they were translators, the usual question was, “And how many languages do you speak?” Many still confuse translating and translators with interpreting and interpreters, and this is something that in the translation industry never ends. And yet, the thing I hear the most in this industry is the invitation to educate the customer, as if you could educate someone who is not willing to be educated. But even in this case, it’s not you who is supposed to educate anyone. The industry as a whole, that is trade associations should educate the public at large, but I never saw newspaper ads, a page in a magazine or a TV ad talking about translation. Because it’s universal, there’s no need to explain.
RR: 39:58 That’s true. Even as a language technology provider in a, in a day to day conversation, “OK, we build this platform for businesses to make the translations” and the people will ask you actually the same question “How many languages do you speak?” It’s like, well actually “I’m a developer” or “I’m a marketer” or…
LM: 40:24 Exactly. This is a consequence of the fragmentation of the industry, which also reflects in the fragmentation of trade associations. We have ELIA, GALA, TAUS, FIT, EUATC, and ALC and this and that…
RR: 40:44 Yes, there’s like a hundred and 50 or something.
LM: 40:48 Why can’t we get together? We can engage, be involved in a single effort that have our industry be better known, be better understood by the public?
RR: 41:04 Yeah, that’s interesting. It would be kind of nice this would be a good article for the future maybe to compare the translation industry in terms of associations with another industry. Like, for example, the pork industry in the United States, which pays so many commercials on TV. I think pork industry is a lot bigger than just translation industry. So, I don’t think that the translation industry is going to have a Super Bowl ad, but they could. They could have something.
LM: 41:39 You know what? I got milk.
RR: 41:41 Yes, exactly. That was the one I was thinking of actually what I did want to pick the obvious one.
LM: 41:49 There’s no bad in picking the obvious. You know, a few years ago, Common Sense Advisory compared the translation industry to the bicycle industry. We all know that there’s not just one bicycle manufacturer in the world, only in Italy there might be maybe 50 or 60 companies producing bicycles, but everyone is sooner or later facing with the need for better lanes for bicycles or respect for cyclists and so on. And, you know, I am sort of commuting between Amsterdam and Rome. There are probably more bicycles than people in Amsterdam, but I don’t think of the brand of the bicycle. I just think of the bicycle. In fact, manufacturers are joining their resources and their efforts to promote the industry. That’s why the industry is still pulverized the same as the translation industry without suffering that much. Can you tell me which is the first exporting country of bicycles in the word?
TF: 43:20 I don’t know, I have no idea.
RR: 43:21 Let me guess. That’s a fun one. Well it wouldn’t be Holland even though they have a lot of bikes.
TF: 43:36 I think of Asia. I don’t want to specify a country, but I think an Asian country.
RR: 43:45 Maybe Spain?
LM: 43:47 No, that’s the problem. We are thinking of a country where bicycles are so many or where there could be a large industry to produce them. It’s Portugal. So that’s the proof in the pudding. That there’s no need to be huge to be important.
RR: 44:03 Totally. Well, I think that our customers would agree as well. I mean we’re a pretty big company these days, but I don’t think any of the language technology companies are huge. I mean, they’re not as huge as technology providers are in other industries where they have several hundred employees and stuff like that, but I guess we’re all pretty important, so that’s a pretty good way to wrap up. So, hey, thank you for coming on this show. It was super interesting.
LM: 44:31 Thank you for inviting me.
RR: 44:32 Yeah, no problem, Luigi, we left topics on the table for next time, actually.
TF: 44:38 Definitely. Thank you for joining us.
LM: 44:40 Thank you for the opportunity.
RR: 44:41 Thanks, Luigi, have a great day!
LM: 44:43 Thank you to you two.
TF: 44:46 All right, so let’s wrap up this episode of the International Buzz podcast brought to you by Wordbee with today’s episode Luigi Muzii. Bye everyone.
RR: 44:54 Be sure to check out the Wordbee translation management system. Not that this is a commercial podcast. It’s been pretty beyond commercial so far with Luigi. I think we do, but if you are listening and you need a translation management system, check it out.


Author: Luigi Muzii

Luigi Muzii