"A Slippery Slope: Measuring the Quality of Extra Virgin Olive Oil and Translation" - thats's how Don DePalma put it in a nutshell on the Common Sense Advisory Website. Here's our point of view to this important issue:
 |
Klaus Fleischmann
on quality management |
In the last
decade, translation tools have not only made impressive progress in terms of
user friendliness, productivity and automation, but have also incorporated a
wide array of quality control tools, making it easier to detect and correct
linguistic errors in the translated text. This has been extended into reviewer
tools and portals, such as our own globalReview product, which allows in-country
(or other) reviewers to check and approve translations online.
However,
for a “holistic” company-wide approach, we need to look beyond simple quality checks
to a strategic level and implement a real quality assurance process in terms of
error prevention, quality evaluation, and ultimately quality improvement. While
quality control is a post-process endeavor, quality assurance starts well
before the actual translation with aspects like clean source texts, solid
terminology, project preparation and intelligent query management during the
translation process. And it continues well after the translation to provide
quality tracking, process improvement and an objective quality assessment. All of
these processes involve stakeholders outside the core translation process:
authors, clients, in-country reviewers and others.
Incorporating
these stakeholders and “fringe” processes into the highly industrialized
translation production process, however, is easier said than done. For
instance, Common Sense Advisory recently found that in-country reviews “…are
notorious for causing delays and frustrations for all parties involved. “
I believe
that we can do two things to improve this review process, apart of course from
technological assistance by simple online tools or in-layout reviewing:
1) Shift
the focus of reviewers from “reviewing” to “assessing” the quality of the
translations. Rather than painstakingly reviewing every segment, they can
review predefined samples and assess them. The error scores will be calculated
and, most importantly, continuously tracked. This gives us business
intelligence to find out which languages, vendors, processes, or projects we
need to work on to improve quality. It also makes the whole process much more
objective and less emotional.
2) From a
process point of view, we need to make it clear to reviewers that they are not
necessarily responsible for getting the text ready and fit to publish. They
assess quality, and if corrective measures have to be taken, they will be
taken: by the translators who are ultimately responsible for the final text.
The reviewers help us to assure the text meets the market conditions and general
expectations and to develop an ever-improving product over time.
Holistic, collaborative,
quality-driven
Sounds like
a rather complex project? Not necessarily, if you, as a precondition, integrate
all stakeholders and smoothen processes. Thus, you avoid frustrations, delays,
or, even worse, total blockage. Furthermore, we recommend implementing a collaborative
language quality process by means of a web-based, centralized platform. The
back-end of this platform needs to be integrated into CAT tools, while the
front-end has to be much more “collaborative” and communication-oriented. The
latter enables and motivates non-linguists to participate in the process, making
it much faster and much more “strategic”. So, say good-bye to the days of painstaking
reviews and hello to terminology processes for the pre-translation, query
management during the translation and quality assessment for the
post-translation phase.
Integrating
these three working areas in one collaborative place creates a healthful boost for
the overall language quality and following significant benefits:
- It
is very easy and highly efficient to manage these very different workflows with
one unique task-based platform.
- Collaboration
and community features make it a lot easier to keep track of tasks and open
issues. The communication-driven process also connects all stakeholders in one
team and thus avoids individual “black boxes” criticizing each other.
- The
quality assessment platform enables the stakeholders both to give input and to
actively track performance. Due to this tracking, the stakeholders are bound to
witness improvements on the overall process. Objective quality indicators replace
occasional, mostly emotional input like “This translation is bad!” This will lead
to a beneficial boost in motivation to continue working on it.
Further benefits
go as far as terminology, queries, and in-country review. For example,
terminology can be defined, “translated”, voted, and approved collaboratively.
Queries can be channeled to the right people, answered and then be shared with
all translators and reviewers working on the project - or even with the
authors, who might want to learn from translator queries to improve their
source text.
How to Implement Quality
Assessment
After
establishing this company-wide collaboration, translation quality assessment can
elegantly be tied into the process. We suggest incorporating approaches from
TAUS and QT21, such as content profiling or error typology:
·
Content profiling takes into account that not the company’s
entire content needs to be treated equally in terms of quality: While
emotional, brand-relevant public content might need to be reviewed completely,
for internal content with a short shelf life a 10% sample should suffice. Clients
and translation providers agree on the kind of content profiles that need to be
reviewed and also define to what extent.
·
Error typology builds on the fact that not all
errors are equally sensitive. A meaning error in a prominent spot is more
severe than a stylistic error in a support article.
Combining
these error type assessments with content profiling enables efficient evaluation
of any text in terms of quality. Based on this evaluation, subsequent measures
can be decided, ranging from instant publication to re-translation.
At the same
time, each evaluation produces a quality score that will be tracked over time,
thus providing business intelligence insights into the translation process.
This makes it possible to recognize what kind of quality process certain
vendors have implemented, which languages or resources are problematic, and
whether certain quality improvement measures have been successful or not. And
more importantly, it increases transparency and objectivity in a field that is
notoriously ambiguous and highly emotional. Reviewers - who usually complain
about having to “correct” all translations although having no time for it - should
actively get involved in the long-term language quality assurance process,
rather than “reviewing” translations or “translating” terminology.
Conclusion
By
regarding quality as a team effort and an ongoing improvement process rather
than an occasional corrective measure, we can make all stakeholders allies
rather than emotional opponents – and finally move away from occasional quality
checks to consistent and collaborative quality assurance.
Klaus Fleischmann
More Information?
Darüber hinaus bietet der SCHEMA ST4 DocuManager in der Version 2016 viele weitere neue Funktionen wie ein überarbeitete moderne Oberfläche, Erweiterungen des CallOut-Designers, Text- und Ressourcengruppen, Unterstützung des COTI Level 2 oder zusätzliche Kommentarfunktionen.
Posts
