Redefining (And even Renaming Quality Assurance) With the Analytical LQE Model

“LQA.” It’s one of those ubiquitous terms in the localization world: Used by everyone, tossed around in nearly every meeting, heads nod with each mention. But often times, the definition is hard to pin down – and, of course, varies from each person using the term.

Which can leave your business open for inconsistent translation and missed opportunity.

What’s even worse, LQA is sometimes confused with QA check, a set of automated checks that produce a report listing potential errors in translation. However, it’s important to note that automated QA checks are not the only step in ensuring quality translations. To achieve better results, LQA should be performed both by a linguist and by automated quality assurance checks.

So instead of “LQA” and its nebulous definition, we prefer the term LQE. And to make sure we’re all on the same page, we want to define it so you can understand exactly what you need to be getting when it comes to ensuring consistent translation quality no matter what vendors you are using.

LQE Process: A True Statistical Measure of Quality (pass/fail ratio is also a part of an analytic LQE model)

LQE provides an analytical method for measuring the quality of a translation by first letting you set an objective measuring system and then breaking translation text into smaller segments to track errors in specific words and phrases so that an absolute and objective quality score can be assigned.

Our industry has made progress, and we’re moving faster than ever. With the advent of technology, global sourcing, and distribution, plus the emergence of markets spanning the globe, the entire localization industry is growing at an incredible pace.

But with Growth Comes Struggle – and Opportunity

Companies are anxious to expand markets and find opportunities – which means they’re increasing the translation rate. And, with increased work comes an even more pressing need to evaluate the quality of all these translations and create objective measures that will improve translation efforts over time – because, as we all know, a poor translation can have an all-too-sinking effect on sales, customer satisfaction and even lead to lawsuits.

In short, we need a new Language Quality process that is faster to apply, more objective in its grading, and has adaptable standards to suit the different types of translation work performed in each market, especially translations containing risk-prone content referring to legal, geopolitical, and/or data protection items.

This is where Language Quality Evaluation (LQE) takes over from LQA (Language Quality Assurance): Expanding the definition of quality from mere pass/fail to a complete system of checks and balances that improve what really matters: your customer’s experience with your brand.

Transitioning From a Pass/Fail and Subjective to Statistical Objectivity

LQE provides an analytical method for measuring the quality of a translation by first letting you set an objective measuring system and then breaking translation text into smaller segments to track errors in specific words and phrases so that an absolute and objective quality score can be assigned.

LQE framework allows you to select and adjust evaluation criteria for specific content type and specific needs but, whatever you chose, and however you tailor it, the results are still comparable (if you have different LQE models for different content types, they still fit together, giving you a detailed view of localization quality throughout different departments)

To think of it another way, TQE is applying the same statistics-based quality control system that revolutionized the manufacturing world to translation.

3-Steps to Success: An Overview of the LQE process (as backed by MQM.org):

While it might sound intimidating, building an effective LQE model for any project can be broken down into a fairly simple process roadmap that first identifies the quality level to achieve for each content type and the criteria that will be used to determine if that quality level was met.

Once the level and criteria are set, the project can be scored, and follow-ups can be delivered. Let’s take a look at the basic steps:

1. Preliminary Stage: Defines your content types and quality goals per type. It helps you select the dimensions of quality that are most relevant to the project, whether it’s fluency and accuracy, internationalization, or other considerations.

Within the preliminary stage, you will also set the thresholds for 3 key measurements that make LQE an objective measuring tool:

• Threshold Value: this is the 0-100 score of the necessary quality. Setting this score is based on the end use of the translation. Many companies set a lower TV on informal communications, while legal documents and contracts require a much higher rating.
• Evaluation Word Count: This is the number of words you evaluated in the project. In the TQE model, this number is based on the word count in the source text.
• Absolute Penalty Total: This is the total number of errors found in the project after being weighted for severity level. Industry-standard severity penalty multipliers are: Neutral = 0, Minor = 1, Major = 5, Critical = 25.

After your thresholds are identified, the text is broken up into Translation Units (TUs) or phrases that can be checked for errors between the source document and the translation.

2. Error Annotation Stage: In this stage, each of the translation TUs is checked for errors against the source document, and the errors are noted. A “scorecard” that classifies the errors by type (examples: Mistranslation, Omission, Punctuation, Unidiomatic Style, Organizational Style, and Awkward Style), and penalty points are assigned based on the severity of each of these errors. Using this scorecard, the Absolute Penalty Total can be calculated.
   

   https://themqm.org/introduction-to-tqe/concrete-example-with-formulas/

3. Calculation and Follow-up Stage: This is the true “magic”’ of LQE. Because we’ve identified metrics and captured data in steps 1 and 2, we can now objectively evaluate translation quality through a series of formulas that lead us to an Overall Quality Score (OQS).

The OQS can be compared to the starting Threshold Value to determine if the project is objectively a pass or a fail, but that’s only the beginning of TQE’s follow-up utility.

Because the OQS is an objective measure, your team can compare quality results between different projects to determine the success rate of translation providers across multiple projects and languages, note trends over time or find areas for targeted improvement.

What’s The Benefit of LQE Over legacy LQA? True Insights vs. Subjective Opinion

Defined goals create defined processes with measurable metrics to judge quality from. With a true LQE philosophy guiding your process, subjective definitions of quality are replaced with objective data, so all team members can see the same end goal – and have the proper tools to measure their results. And by digging into the stats, each team member can see what steps they can take to improve their performance and deliver greater value.

https://www.istockphoto.com/vector/3d-isometric-flat-vector-conceptual-illustration-of-evaluation-gm1418874295-465421627?phrase=performance%20evaluation

Need Help Making the LQE leap? We’re Here for You

With over a decade of experience in the localization quality world, we’ve become, if you’ll pardon the bragging, masters of implementing quality evaluation and control processes that break down the walls between translators and other stakeholders and get the results that today’s fast-moving business space demands. We’d love to explain more about how adopting an LQE philosophy can help – and we can guide you through every step along the way.