We spend a lot of time thinking about and advising customers on how to achieve the highest quality translation at the lowest total cost. Quality is important. After all, if your team puts a ton of energy into bringing a product to the market and the last mile - translation - is fumbled, all of that hard work would have been for naught.
The challenge of quantifying, measuring and addressing translation quality is a frequently discussed topic within Smartling and with our customers. How can a business remain in tone, brand and efficacy in multiple languages? It turns out there are a handful of ways to understand or measure translation quality.
Different types of content warrants different methods of evaluation. That critical email your CEO is sending in response to a natural event has a bit more weight than a new web page you're launching in a few months, right?
So how do we work in a standardized method to measure quality across different languages and content?
Of course, we weren't the first to ask this question, and certainly won't be the last. While there have been different scoring models, none of them were completely satisfactory.
TAUS created their Dynamic Quality Framework in 2012 to address some of the "moving targets" of translation quality, like content type and machine translation leverage.
Since translation quality remains a top concern, especially as brands continue to translate vital content during this time of disruption, we wanted to take the opportunity to sit down with Smartling's Aisling Nolan to learn more about why brands leverage DQF in Smartling.
As Smartling's Director of Customer Success, Aisling has a ton of exposure to different use cases and priorities of our customers. Let’s see what she has to say about DQF.
So, Aisling, for people who are new to translation, what exactly is the DQF? How does it work in Smartling?
DQF is an industry standard process that global businesses can depend on to test for translation quality.
Customers can use DQF for a human review process, completed by credentialed reviewers based on the TAUS DQF framework without leaving the Smartling platform - it’s built right in.
If DQF adds another layer of quality review, is this meant to replace an internal review process?
It completely depends on your objective. Internal review is typically used for a larger volume of content to ensure brand continuity whereas DQF is a process typically associated with a sample of content to understand, on a thematic level, the quality of translations.
When should localization managers consider using DQF?
While internal review might make sense for ensuring consistency and quality as a final step before content is published, DQF can be used as a larger-scale auditing and quality assurance tool.
For example, brands that are refreshing and re-writing large volumes of content can leverage DQF to audit these bulk updates. We see this typically when brands are redesigning or launching an entirely new website, they'll use the DQF framework within Smartling for a sample of content instead of sending each individual page for internal review.
Brands can also rely on DQF for overall quality compliance to ensure continuously evaluate translators and content standards set by localization teams.
Interesting, do you know of an example?
Of course! A lot of our clients will use internal review for emails. It's content you can't get back once you send it, and it is also typically being created and deployed at a fast pace.
DQF is a better fit for auditing the quality of your translations across your brand's entire website.
Your website is high priority, but it can also be edited and changed whenever necessary unlike that email that you can't recall after it's sent.
Seems like DQF can be a really powerful tool for localization teams. How nuanced is the DQF configuration in Smartling?
I get a lot of questions about how DQF works in Smartling. We've built it as a simple error-counting framework for reviewers to access directly in the CAT tool.
As a reviewer reads through the translation, they mark down any errors in a dedicated quality evaluation panel.
Reviewers work through content one project at a time on a string level, and when they submit their work, Smartling creates a report for your team to analyze.
When reviewing a translation, how do translators determine the weight, or score, of each error? How do they keep track, or know when a translation has passed or failed?
Our implementation allows reviewers to evaluate and mark down any error typology, and we leave it up to the customers to determine how error typology will affect scoring so they can determine the most relevant framework.
We’re also happy to assist with tuning the weights and severities to determine what is acceptable.
Does Smartling implement any guidelines or scoring from TAUS?
Yes! We use TAUS' default values in our template for severity weights, as well as assigning severities to different error categories.
But we also make it possible for customers to configure their methodology to best fit their needs.
How do localization managers access all this info?
There is a dedicated tab within that CAT tool where the DQF scores are added at a string level, and a detailed DQF report is housed alongside your other dashboard Reports that analyses the results.
So with all that, would you say the DQF is better than internal review?
They’re different, and what you choose depends on your objective. Internal review is typically used for a larger volume of content to ensure brand continuity whereas DQF is a process typically associated with a sample of content to understand, on a thematic level, the quality of translations.
Aisling’s insights really shine a light on what DQF as a service within Smartling, andhow localization teams can take advantage of this framework being built into the tool.
If you're curious to learn how Aisling and her team can help your brand, reach out and chat with us!