Automatic website translation is getting a lot of buzz these days. Unfortunately, we’ve seen many people not understand what it does or how it’s supposed to work. It can be an extremely valuable tool, provided you know its limitations and plan in advance for them.
Automatic website translation can significantly improve manual translation methods that take time and are hard to maintain. However, introducing automation will also introduce some risks, like the potential for minor errors to grow into massive, multilanguage catastrophes. The best way to leverage it is through a hybrid process that brings in machine and manual translation aspects.
The Pros and Cons of Manual vs. Automatic Website Translation
Before moving on to the pros and cons of manual versus automatic website translation, it’s vital to define both. The manual method is a step-by-step, human-run process–you download the files and send them off to a translator. The linguist sends them back; you check them and then upload them yourself.
Automated translation is a bit more complex as it is often incorrectly used interchangeably with machine translation. A key difference is that automated translation is workflow-driven. A connector and a code work together to grab content in need of updates. Jobs are then triggered based on changes, assigned to linguists, and translations are automatically uploaded to a website once complete.
Machine translation uses a computer program like Google or Microsoft Translator to complete the update. It’s not automated, as you will have to copy/paste text and upload it yourself. The simplest way to remember the difference is automatic translation covers the workflow–human linguists still complete content updates. Machine translation covers the text, but humans will still have to manage the workflow manually.
When comparing manual, machine, and automatic website translation, there are pros and cons to consider that will impact your results, workflow, and processes.
- Total control: A manual process gives the project manager complete control at every step.
- Strict oversight: As this is entirely manual, every step will involve supervision, limiting the amount of mistakes.
- Lower upfront costs: There are no significant upfront investments like in localization management software. The initial upfront costs of a manual process are much lower than an automated one.
- Challenging to manage: With every step, the project manager has extensive follow-up and processing tasks like checking quality, uploading and downloading documents, and monitoring content for needed updates.
- Time-consuming: Managing all of these tasks is often a full-time job that will require additional staffing and resources.
- Increased long-term costs: With increased staff comes increased cost. Any in-house manual project will be an expensive long-term endeavor.
- Low cost: Many programs are available for free and offer reasonable levels of accuracy. Google boasts accuracy rates of 81% for Chinese translation and scores as high as 92% for Spanish.
- Near-instant: Even extensive translations can be completed in under a minute using a basic machine translation program.
- Suitable for low-priority work: Machine translation can be beneficial for low-priority, simple to translate projects like support pages.
- Low quality: While machine translation is always improving, it’s not great for complex tasks like translating similes and metaphors. The results will be disjointed, nonsensical, and in the worst case, offensive or misleading.
- Limited applications: Machine translation is only appropriate for low-priority content, testing, and other simple tasks.
- Unsuitable for high-priority work: Things like marketing material, or items with a lot of jargon, like medical or legal documents, are not appropriate for machine translation. It can make serious or even dangerous mistakes.
- Consistent updates: Automatic processes ensure that content is always up to date as jobs are triggered by the discovery of changes.
- Lower long-term costs: Many tasks are automated, so you won’t have to check quality or follow up and upload documents manually.
- Improved results: Overall, the automatic translation results are seamless and provide high-quality content compared to other methods.
- Limited control: Automation takes many options out of your hands, as content is parsed and managed by machines.
- High-impact mistakes: A small mistake could filter into other languages, becoming exponentially worse during each transition.
- Higher upfront costs: Establishing an automatic translation system can be expensive upfront as you have to map out so many processes and strategies accurately.
When considering the pros and cons of all these options, it’s important to remember that you can mitigate many issues by combining strategies into a single, end-to-end localization program.
Mitigating Problems With a Combined Translation Process
Hybrid translation automation is a process that combines both automatic strategies and manual methods. You may, for example, trigger a job with a crawler using automation. Then, you review and approve that job via a manual strategy. On top of that, you can bring in machine translation to act as a tool to aid linguists and improve results. Combining all three methods allows you to mitigate the problems and take advantage of the benefits.
Automatic website translation can be a very valuable tool, but you must be aware of its limitations. By adding both manual and machine translation aspects, you can streamline your workflow and eliminate mistakes. This combined approach ensures excellent translation on a timetable that works for your organization.
April 29, 2021