The Translation Tech Trifecta

Many localization teams struggle to help others understand the work they do. Yet, quickly grasping how to work with localization is very important in high-growth digital companies. Communication is quick and the pace is fast. No one has the time to learn about a niche discipline like localization, but yet, they depend on it directly to achieve their goals. So, it’s important to use terminology that makes it easier for people to easily understand the work a localization team does. How can we demystify localization and make it more accessible? That’s where the Translation Tech Trifecta (T3) comes in.

Categories That Are Easy to Understand

To make it easier for people working in SaaS and other digital companies to understand what we do, here is a simple framework:

This “touchless” terminology is widely known and accepted terms at tech companies when applied to many other areas, especially sales, purchasing, and monetization strategy. It isn’t so common to apply these terms to localization, but maybe it should be. The holy grail in tech would be what is known as “true touchless” revenue, in which automation happens to the maximum degree in order to minimize the need for human involvement. “Touchless” is aspirational. It’s a goal that many digital companies strive for but will never fully achieve. Others do achieve it, but rarely does it make up 100% of all revenue.

“Touchless” translation is something localization teams also strive for, even if aspirational in this corner of the world too. Localization teams aim to minimize human involvement not just on the linguistic side, but on the human file shuffling side as well. What is useful about these categories is that most people working in tech also intuitively understand that there is a spectrum of what “high-touch” means. Optimizing processes to such a level that they can be touchless is not always possible, or even desirable, if you want to achieve the best outcomes for your customers.

Here’s a bit more about each of the categories in the Translation Tech Trifecta, or T3. (By the way, T3 is just a convenient shorthand for localization folks since we love abbreviations for the sake of economy of language. I don’t recommend you throw this or any other numeronyms around with stakeholders unless you want to bore and confuse them! Stick to the three categories, which are easier to understand without explanation, instead!)

Touchless = Fully Automated

A touchless translation workflow is fully automated. No humans have to touch it, not even to kick off a workflow initially. You kick off the workflow with some other automated trigger, such as hitting a certain tier of website traffic, view counts on a page, click-through rate, or even just publishing a new article or an update that meets certain criteria specified in advance. A human might have to hit a “publish” button somewhere, of course. And a human might have to configure things up front, like frequency of translation batches, and so on. But bottom line, touchless translation is the type that happens instantaneously and ideally, continuously, to handle not only net new content, but also content updates in an automated and touchless way.

Touchless translation reminds me of a modern version of the old adage, “measure twice, cut once.” But instead, you measure 2000 times up front, so that no human has to ever manually cut anything. You’re going to have to spend a lot of time getting the process and tech infrastructure just right to enable any truly touchless translation workflow. Planning is critical to enable touchless translation to be sustainable, and ideally continuous.

What touchless means for localization practitioners is that you’ve set up content to flow from a content management system (CMS) into a translation management system (TMS) to leverage your translation memory (TM), followed by a machine translation (MT) workflow, or perhaps you’ve configured it to go straight through MT, preferably neural machine translation (NMT) if we’re talking about most types of web-based content that can change on the regular. To do this, you’ll need a CMS that actually has multi-language functionality (I’m smitten by the HubSpot CMS for that purpose, of course). You’ll also need access to your own developers in order to do this, because very few content tools come with out-of-the-box MT functionality.

Low-Touch = Partially Automated

What low-touch translation means is that humans don’t get overly involved in the process, but are still in the loop. This is the perfect way to describe a large number of workflows in my opinion, since human involvement is truly minimal in many cases, but still necessary. Low-touch also still sounds attractive to stakeholders, because it sounds more tech-enabled to them. It also reassures them that you know something about automation, and that you realize human involvement isn’t always desired.

For localization folks, what this means is that any machine translation post-editing (MTPE) workflow you have running, you can bucket under low-touch translation. Some MTPE workflows are actually so involved that you could argue whether or not they truly deserve the moniker, but hopefully in those cases a team would flip things over to a human translation (HT) workflow anyway, if it’s not yielding the desired results.

High-Touch = Tech-Enabled

With high-touch translation, automation possibilities are limited and humans need to get involved. However, the process can still leverage technology to a high degree. Even in highly creative projects, technology is usually utilized by any localization team. But, by acknowledging that it’s high-touch, you’re speaking on the same level as your stakeholders, communicating in terms they can easily understand. You’re flagging that humans are involved here, even though that might not be optimal in terms of speed or other factors.

When you talk about high-touch translation, the focus with stakeholders should be on showing them how, in spite of human involvement, your team is leveraging technology extensively. For example, it’s possible (and quite common) to do continuous localization with a high-touch workflow. At any company that embraces continuous deployment, most product UI strings are localized this way. Ironically, developers might assume everything is done through automated translation. For them, the process might seem touchless. They deploy, and strings “automagically” appear in other languages. They might not even realize there is a high-touch, human-led translation process going on behind the scenes. That’s ultimately the goal here, that even the high-touch workflows are so tech-enabled that they can happen seamlessly and continuously — even with heavy human dependencies.

Localization friends, this category is where the majority of your localization projects will likely fit, especially if you’re working on any marketing localization projects. If you handle more than just marketing, you will likely have sub-groups here, of “high-touch continuous” and “high-touch on-demand” workflows, and the many variations of these that exist within each sub-group. Those details won’t matter much to your stakeholders, but the fact that it’s “high-touch” will help them understand why these types of projects take longer, or are more expensive, opening the door for you to explain why certain tasks still need to be performed by humans, just like in any other “high-touch” process.

Primarily, high-touch translation would almost always rely on computer-assisted translation (CAT) tools, terminology databases, and usually a two-step translate-edit (TE) workflow for the linguistic production process or in some cases a three-step translate-edit-proof (TEP) workflow.

Speaking the Language of Growth and the Dialect of Digital

The other thing I like about using these simple T3 categories to describe the varying levels of translation automation is that they link localization to something positive that everyone always feels good about: growth! Because the term “touchless” goes hand in hand with revenue and growth, it evokes something pleasant, something way more aspirational than localization, which sounds boring to most people. Revenue is the oxygen that feeds the rest of the muscles in your company. There is strong alignment around the fact that things can be done faster and more efficiently when you minimize the need for human involvement. “Touchless” conveys this easily.

The goal of any localization team in using such a model might not necessarily be to shrink the volume or percentage of “high-touch” projects at your company. Perhaps you want to leave that volume the same or even grow it. But acknowledging and measuring what projects and volumes fall into each of these categories is, I think, a helpful way for others to understand that you are on the same team, working toward the same mission, and speaking the same language as they are: the language of growth, and the dialect of digital.