|
Hi, As a boutique LSP, selling raw MT to your clients might completely transform your business. (And if you keep reading, I’ll tell you how to do it.) That’s because you can deliver machine translation for less than your clients can do it themselves… And you can do better work… And then you can upsell some of that to MTPE, or even, on occasion, to premium human translation and transcreation. Oh, and did I mention you can increase your margins at the same time? And because you’ll be handling exponentially more volume, you’ll even grow your revenue. Not just this month, but next month and every month after that. And not just for you. But for your translators as well. No way, you say… “Why would my clients pay me for raw machine translation if they can get it free in Google Translate or DeepL?” But it’s not free; shall we list out the costs? 1 – The Cost of Data Insecurity If they use free services, their content goes into the training set for the future, their confidential company data being stored and shared inside the matrix. Maybe they don’t care. Fine, let’s keep going. 2 – The Cost of Non-Structured Multilingual Content They can’t process their work in a CAT tool; they don’t know how. Even if they know it’s possible, they won’t want to learn it or be bothered by it. This means they can’t leverage past content with translation memory, which also means they can’t efficiently update their translations later. They can’t run additional automated quality checks. They can’t build in their branding, or terminology, or internal consistency. They can’t output their translations into original formats, and they can’t maintain formatting. They’re just shoving documents in and getting them out, each time a little different. That’s not structure; it’s the Wild West. 3 – The Cost of Segment-Locked MT But what if they do use standard CAT-tool processes? They still won’t learn what’s really possible by breaking out of the segment-by-segment confines and leveraging context more robustly. (Heck, most translators don’t even know what’s possible outside the established workflows.) As a result, they can’t produce premium MT, which is certainly NOT available with mainstream tools, no matter what the toolmakers say. That means they’ll be left searching for fancier and newer tools to solve the problems caused or not addressed by mindless automation. 4 – The Cost of Data Limits If they use free tools, they’ll get throttled as soon as they sit down to a big job. The tool will limit how much they can do; tell them to come back tomorrow for more. So if they have any volume at all, they’re going to have to pay for something. Fine, so they get the paid version of DeepL. Still throttled: 20 files per month. They upgrade to Enterpirse… Hmm… Have to use the LLM models provided… and not so free anymore, is it? Still cheap, but now they’re starting to get in a bit over their head. Maybe they try Robert, or Algebras… or any of the other DIY AI translation solutions popping up like mushrooms after it rains… Those cost more. Why? Maybe they’re better. But now they’re up to $0.01/word, or more. Remember that number. 5 – The Cost of Time And who’s going to do the work? They think it’ll be easy (looks easy.. niece just studied in France for a semester; she can probably do it!) so they give it to an intern… or a low-level employee? This person is now going to take their technical content (manuals, spec sheets, promotional videos, etc.), stuff they’ve spent tens or hundreds of thousands of dollars creating in the first place, and plug it into a tool. Do they know how to use glossaries? Other fancy functions? They can figure it out. Then they can pump those files through into every language they can think of… It’ll be “good enough”… But if it is, don’t for a minute think it won’t have taken a lot of time to learn and execute. And then… time for updates… Oops, their content isn’t in translation memory. That means they have to go back and redo everything from scratch. More time spent… 6 – The Cost of Non-Reliability And is it even right? They’ve heard they can get cheap translators on ProZ.com or Upwork to check it…. Great, go hire for that sight unseen… Get it certified correct… Not worth the virtual paper it’s written on. We know how that’s going to end… Or maybe not. They might get lucky. 7 – The Cost of Lost Opportunities Six months later, where are they? The content is a mess… or maybe it’s not. But what about everything else that employee (and their boss, cuz let’s face it, it’s going to take some management) could have done instead of this “translation project”? The opportunity costs of having taken so long to get it done? Or of the inconsistent quality and content organization. Somebody was a little distracted from their main job, weren’t they? This could be the biggest cost of all… And we’re not even done yet… Because… 8 – The Cost of Non-Sustainability Maybe that employee actually did a fantastic job. Let’s just assume they did. It’s still six months of salary, but it’s good work. That employee is ready for a promotion. Or they leave. They’re gone. Now it’s back to square one. New person, back on the project. Or not… 9 – The Cost of Workflow Complexity And what about when they actually want premium human translation too? Maybe they come to you for that occasionally. And they process the MT inhouse and outsource the MTPE on Upwork… Or just ask you for sporadic “proofreading” services… They’re now managing a rolodex of vendors just to stitch together their Frankenstein-like DIY approach. 10 – The Cost of Fixed Overhead Building a localization department is what they’re really doing. But honestly, they don’t even have enough work to justify that much. So they’ve got staff split between this task and all the other stuff they need to do. Or maybe they’re just so inefficient that they do have a full time person or two in charge. That’s not cheap. Or maybe nobody’s in charge; everybody just uses ChatGPT on an ad hoc basis. OMG!!! Everything I said above times ten… … Truth is, your clients don’t want DIY translation, free or not. What they really want is someone to just “solve” the language problem for them affordably so they can get on with their business. But if all you offer is one rate for one level of high-priced full-human translation, you can’t deliver the value they want. And then they’re not going to choose you. They might stick with DIY if their needs are light and the costs haven’t snowballed yet. But more likely, they’re going to hire someone that gives them more value than you do. But why can’t that “someone” be you? No, seriously. What if you can give them a three-tier service structure without the overhead? What if your cheapest service really is cheaper and better than “free” MT, and you can now give them honestly verified “good enough” translation, as well as sooper dooper full-human translation? That would be the Holy Grail, wouldn’t it? You’d be their one contact for everything…. just like the good ol’ days. |
|
What if this lets you get back to doing everything you love about running your translation agency? This is the transformation that you’ve been waiting for. And until today, you may have thought it wasn’t attainable anymore. But the fact of the matter is, you CAN deliver raw MT cheaper in any language than your clients can do it themselves, and you can deliver it better…. And then you can deliver better of everything else too. You’ve lived and breathed language for decades. You know best practices, you know your TMS, you’ve built your systems, and you’ve got a roster of excellent linguists who are wilting on the vine while your sales drift downward. In fact, watching these assets depreciate in real time as the AI hypsters have told you you’re toast has probably kept you awake a few nights, hasn’t it? I’m a translator. I know the feeling. It hurts. And so you’ve been trying to come up with a way to become relevant again. But around the time you realize that you can compete against free MT, perhaps another doubt creeps in. How are you going to compete against the gods of localization? After all, the MT providers, CAT tool makers, and mega agencies are building out an infrastructure to bring the costs to zero and they’re going to gobble up the market. Costs to zero? Really? There’s no money in that. Of course they’re not going to bring the cost to zero. In fact, they’re aiming to keep the costs as high as they can get away with, while eliminating linguists and small LSPs from the market if they can. And they’re offering their fancy, bespoke solutions to big-time multinational multilingual content buyers, since they’re the only ones who can afford to cover that overhead. But you never sold to the mega buyers. Your niche was the niches. That means your previous market is still in play! It’s just waiting for more value. Give your old clients 10x, or 30x, more value for the same money, and they’ll come rushing back, for all the reasons I’ve listed above. You just have to figure out how to deliver all that value. What are your options? Well, you can go get RWS’s MT thing. Or memoQ’s plugin. Or something similar. Uh-huh… Now you’re competing against the big guys, using their tools. How’s that going to end up? You know the answer… In your gut at least. So you get creative… Assemble a suite of tools and connect them together. Lots of automated LQA out there, it seems. Or agents. Or special models. Or fully automated end-to-end solutions. If so, where do you add your value to avoid becoming just another undifferentiated provider? Be the one with the most complicated Rube Goldberg workflow? There’s no future in that. You’re not an engineer. You’ve got an agency to run. You need something else. Something that leverages your human-touch strengths, not something that makes you dependent on automation. |
|
Workflow iteration and context curation are your silver bullet. There’s one weakness all of the automation tools suffer from: They are stuck in the segment-to-segment straitjacket. That’s the price of full automation. It means they can’t translate with robust context leverage or adapt flexibly at the document level. That’s why they make such a big deal about this fancy feature, or that new model, or whatever other “extra” automation they throw on the process. But context leverage is what Generative AI does best. It’s its superpower. Leveraging context properly lets you bypass fancy tricks and push the LLM to the frontier of its machine translation capabilities in one go. It can be done using the Generative AI Iterative Translation (GAIT) Workflow. Human driven from the beginning. A manual workflow that depends on the human going first. And the price you pay for this native quality? Manual processing. Takes tons of time. It takes so long that processing MT through GAIT costs a lot. In fact, it costs… $0.01/word What? That’s cheaper than the state-of-the-art automations. Surprise! They don’t tell you that. By the time they build out their massive infrastructure, put their engineers on the job, and take the process completely out of your hands with their esoteric technology, there’s no way the big guys are going to tell you there’s a better way. But now that you know, this is how you can compete against them. All you have to do to prosper is deliver more value to your clients than anyone else; then you have a business again. And doing it with a human workflow is how you succeeded in the past, and it’s how you’re going to succeed now too: – With better service You just need to get it done. That’s where I come in…. I’m your on-call MT department to make you super competitive again. If we’ve worked together before, you know me for my premium Korean translation. What you may not know is that I’ve spent the last few years engineering AI-human workflows specifically for translators at CotranslatorAI.com. Right now, independent agencies are being forced to buy out-of-the-box MT solutions from the very tech giants and mega-agencies that are actively trying to automate us out of the market. But to survive, we must compete differently. As a career-long vendor, my business relies on you prospering. That is why I adapted my commitment to human-level quality into this outsourced, secure MT solution in any language pair. By operating from Korea, I act as your overnight processing department, preparing files with intention and arming you to compete with the giants on your own terms. You do the marketing; I’ll process the content. At a base rate of $.01/word. – No upfront capital investment Curated premium raw MT… that feeds into better MTPE and full-human premium… in any language pair. That’s a winning service combination. And you can start selling it tomorrow! —- Right now, I am personally working my way through my client list to introduce this on-call MT service. (And just to be 100% clear: I am not stepping away from the translation desk! My core Korean full-human translation and post-editing services remain exactly the same. This MT pipeline is simply a powerful new engine running in parallel.) I plan to continue reaching out to agencies one by one until my dedicated MT production schedule is completely full. If you want to jump to the front of the line, just tell me. I’ll send you my service sheet. When I send it over, I highly recommend flipping straight to Page 4. It outlines the ready-made Three-Tier Service Model (premium raw MT, verified MTPE, and full-human translation) that you can plug into your existing setup and start offering to your clients immediately. Just hit reply if you’d like to find out more. Hope to hear from you. Steven PS – To be completely transparent, I am a single professional operating a highly customized pipeline, which means my daily processing bandwidth is strictly capped. Once I onboard a handful of partner agencies, I will have to close the doors to new MT clients. Reply right now to get the service sheet and secure your spot before my schedule is locked. |
