If your job has to do with translation—and especially with handling large volumes of content—then it’s probable that you’ve heard about machine translation (MT). Some industry professionals will tell you that MT can translate anything—fast and for cheap. 

But don’t fall for the hype.

Yes, MT has improved in the last few years and become easily accessible to the masses through platforms like Google Translate and DeepL. But it’s by no means 100% perfect. Though many people assume MT only involves machines, the engine itself is just one piece of the translation puzzle.

For example, it’s often too risky to use without human involvement, so you may have to pay for wordsmithing and QA to achieve the level of quality you need for your market.

With so many questions and misconceptions surrounding MT, we’re happy to share with you our six-step process below, designed to help you determine whether MT is right for your translation program before rushing to jump on the MT bandwagon.

Step one: Analyze content

The first step is to analyze your content's purpose. You should be able to place each piece of content into one of four categories based on what we call the “scale of emotional weight”:

1. Inform: This type of content has no emotional weight. Its purpose is to provide readers with quick information. Examples include submission orders, contractual positions and codes of conduct.
2. Instruct: The content’s purpose is to provide explanation or instruction in plain, emotion-free language, such as policy updates and orders to initiate.
3. Interactive: The content contains user interface elements that lead users through an experience, like adding a product to an e-commerce shopping cart.
4. Inspire/influence: High-emotion marketing and sales content. (This type of content rarely shows up in the legal context.)

The purpose of your content often maps to the way it was written: simple language versus highly branded, creative content. The more creative the content (category 4 above), the riskier it is to use MT. Most legal materials fall in categories 1 and 2. 

For example, “inform” and “instruct” content is prime material for MT because readers only need a general understanding of its meaning, though you might need human editing depending on the translation quality you require. At “interactive,” human validation becomes critical to ensure users can navigate experiences correctly.

For content in the fourth bucket, we don’t recommend using MT at all—unless fast turnaround is more important than quality.

Step two: Define business goals 

After sorting your content, you must decide where and how to apply MT. It’s not as simple as setting up MT for all the content types in the first two categories above (inform and instruct). Your language service provider (LSP) will need you to describe specific business goals that can be met with MT.

For instance, you might want to translate only your high-volume content (like general legal documents that do not require certification) or only language pairs for which you have scarce resources. Perhaps you only need to translate the gist so a reader can get the general idea. But for certain types of content—like ones requiring certification, which need to be clearly understood by all parties—you might have zero tolerance for errors or permit only non-critical errors like grammatical inconsistencies. 

It’s important to define your specific business objectives for each type of content to apply the right MT technology, trained for a specific use and language pair, to the most appropriate situation. Even then, you don’t know if MT will work well until the pilot stage.

Step three: Conduct a pilot 

Before running an MT pilot—testing how well a chosen MT engine handles translation for specific languages—you need to decide what you expect to get out of it. What does success look like in terms of translation quality, for example? Is success about speed? An LSP can help you figure out your criteria.

Then comes executing the pilot. Your LSP will choose which engines they recommend you try, run them on a test set of files (a sample of your content), and compare the results against what you consider a “good” human translation.

For example, you might provide NDAs in English and already translated into German by humans that meet your quality requirements. Your LSP will evaluate the outputs of various MT engines—collecting data along the way—to see which comes closest to the quality of the German translation.

Step four: Analyze the results 

With the data collected in step three, your LSP can analyze which engines are most effective for your business goals. A mix of human and machine-automated quality analysis can determine which engines, if any, meet your success criteria.

The one that comes closest to the quality of your human-translated reference material will be the winner, but there is no one-size-fits-all solution—different engine outputs might “win” for different language pairs and content types; your eventual MT program might use several. 

If none of the outputs meet your success criteria, perhaps you aren’t ready for MT yet—or as we prefer to think, MT isn’t ready for you yet, because it’s not up to the level of quality you need. (And that’s OK—you have other options.)

And if your tests are a success…

Step five: Deploy your engines 

…great! You’re ready to move on to deploying MT. That means choosing your engine(s), setting them up for the languages you want, building a post-editing program (if you need it) and training project managers on processes.

Importantly, the deployment model varies for each company. You have some unique decisions to make: how to integrate MT with your existing workflows and translation management systems, whether to deploy on-premise or in the cloud and how many review passes translations need in each market, to name a few. Talk to an LSP about how to deploy MT—they can help develop the right roadmap.

Step six: Optimize, rinse, repeat 

Even after choosing an MT system that meets your present needs, it may or may not evolve with your business growth or hold up long-term. It’s like buying a pair of shoes that seem to fit at first, but over time, you might notice discomfort or wear and tear and realize they’re not so perfect after all.

So, while we at RWS Alpha believe that every growing translation program should use technology to advance their goals of handling more volumes, reducing costs and improving quality, successful MT warrants continuous testing, analysis and optimization. Use these steps to assess the technology’s effectiveness for your business and to plan initial goals, and your LSP can take care of the rest.

Come to us with questions about any stage of machine translation planning or deployment.