Website Translation

How MotionPoint Efficiently—and Affordably—Localizes Dynamic Content

Discover how our superior solution finds, and translates, hard-to-detect dynamic content within complex application code.

Todd Michalik's avatar
Todd Michalik

October 09, 2019

5 MIN READ

MotionPoint’s localization solution understands and handles the technical complexities of modern websites—particularly how sites present dynamic content to visitors.

This is a big deal, because it’s something other translation approaches can’t do well … or at all. They usually wind up using inefficient workarounds that increase translation costs in unexpected, unwelcome—and unbudgeted—ways.

MotionPoint dramatically reduces the cost of translating dynamic content with:

  • Superior content-detection capabilities
  • Content-parsing technologies with pattern recognition
  • A unique translation approach

Let’s take a closer look.

MotionPoint’s Superior Content Parser

First, here’s a primer on how MotionPoint’s solution gathers and translates digital content:

  • Our web crawler detects a website’s translatable content
  • Our solution parses any relevant translatable text into phrase- or sentence-long units called segments
  • Segments are translated by software or human linguists and stored in a special database called translation memory

With MotionPoint, segments are translated only once. They can be re-published countless times on a localized website via translation memory at no additional cost.

MotionPoint provides additional savings—which can save customers between 10% and 40%—thanks to how we’ve optimized our technologies to handle dynamic content by increasing segment repetition.

Segment Repetition

People regularly encounter segment repetition on websites, in templatized product descriptions. This content often appears as sentences or headlines with predictable structures across multiple pages.

These descriptions are usually dynamically generated from databases, powered by complex application code. Here are a few examples; the repeating segments are underlined for clarity:

  • Long-sleeve green dress shirt by ManStyle, XL
  • Short-sleeve blue dress shirt by ManStyle, small
  • Short-sleeve gray polo shirt by ManStyle, medium

How Other Approaches Handle Dynamic Content

While the sentence template in the examples above is clearly identifiable to a human mind, it is not obvious to most website translation solutions. They also don’t understand that the template’s discrete content variables (like sizes and colors) are dynamically assembled in real-time.

This means that their solutions treat every single instance these variations are combined and presented to end-users as unique phrases that require translation. In essence, every time a website visitor’s on-site actions dynamically generate a sentence, a request is sent to the vendor to reactively translate it.

This creates all kinds of inefficiencies and troublesome UX issues. The most significant problem: Depending on the quantity of templates and variables, the number of resulting unique phrases can become astronomically high.

And so does the cost for linguists to translate each one of them.

MotionPoint’s Superior Approach

In contrast, MotionPoint’s smart parser recognizes these template structures, and notifies our team when it detects them. We identify all of its discrete units of content—those individual variables such as sizes and colors—and mark them as reserved terms. We then proactively translate them.

We only need to translate these special terms once. Our solution understands that no further translation is ever required for a particular template, since all content variables have already been translated and stored in translation memory.

The translations are republished at no additional cost, no matter how many times they appear on-site. This completely eliminates the inefficiencies and astronomical translation costs generated by other vendors.

Reserved Terms: A Closer Look

As previously mentioned, segments that contain translatable keywords are identified with reserved term placeholders. When this content is presented to customers on the localized site, text within the placeholders appears as translated.

For instance:

  • People who liked Baby Estella Glitter Dress also viewed these products in Girl

Here, the phrases in green are reserved terms that are considered translatable. Reserved term placeholders have been placed around the product name (Baby Estella Glitter Dress) and the product category (Girl).

Similarly, our technology ignores non-translatable text within segments, which means these dynamically generated phrases don’t enter our system for translation. For instance:

  • We’ve sent your password-reset email to V****th@cable.comcast.com.

Here, the content in red has been identified with a placeholder as a non-translatable reserved term. While the rest of the phrase will appear as translated on the localized site, the red text is ignored by our system.

Here’s an example of dynamic content we might see for an American railroad company:

  • Train route: PBI to ATL on west track
  • Train route: FLL to NYC on north track
  • Train route: MIA to ATL on southeast track

We then define which instances of dynamic content contain translatable words or phrases (seen below in green) versus non-translatable words (seen in red):

  • Train route: PBI to ATL on west track
  • Train route: FLL to NYC on north track
  • Train route: MIA to ATL on southeast track

We then translate the relevant segments (everything that isn’t red):

  • Ruta del tren: PBI a ATL en la via oeste
  • Ruta del tren: FLL a NYC en la via norte
  • Ruta del tren: MIA a ATL en la via sureste

This process is ongoing, meaning MotionPoint’s technology continually looks for the above patterns, and we implement the appropriate reserved term placeholders as needed.

Conclusion

As you examine your options for website and digital CX localization, take a good look at how potential vendors handle dynamically generated content. Ask about:

  • The thoroughness of their content-detection capabilities
  • Their solutions’ ability to detect translatable dynamic content within application code
  • How (or if) they save money for their customers along the way

MotionPoint’s approach uses superior content-detection and -parsing capabilities, coupled with an efficient translation approach, to dramatically reduce the cost of translating dynamic content.

Last updated on October 09, 2019
Todd Michalik's avatar

About Todd Michalik

Todd Michalik manages the pre-sales technical process and evangelizes MotionPoint’s technology to existing and prospective customers. He spearheaded the creation of our data-driven culture, mentors & motivates engineers and data analysts, oversees the operation of our collaborative sales tools, and more.

Keep Learning

Here is some related information that you might enjoy