Language Quality Assurance || Logrus IT

Logrus IT’s suite of services includes linguistic content audit / quality evaluation (LQE), which entails evaluating the quality of translations performed by in-house employees or freelancers.

Why is LQE necessary?

LQE can be an appropriate solution if you:

Frequently translate into multiple languages, but have no employees who speak those languages and can evaluate the quality of your translations
Are uncertain about the quality of a translation provider’s services and would like to get an objective expert opinion.

LQE can help you determine:

Exactly how accurately a translation conveys the meaning of the source text
Whether or not there are errors in a translation and how critical they are
Whether or not a translation corresponds to the cultural, linguistic, and legal characteristics of the target country.

As a rule, in order to determine the quality of a translation, the linguists from the LQE team only need a few fragments of the translated text. This approach is especially convenient if you need to evaluate a large amount of translated material in a short period of time or on a limited budget.

What you get after LQE

A numerical score for the translation’s accuracy and readability

A list of errors, each with its own category and degree of severity

Various corrections for any errors

The linguist’s comments and conclusion.

OUR EXPERIENCE

Perception assessment of payment platform interface

Goal: Understand how the product’s user interface (UI) looks and feels to local users in local languages:

is the purpose of the screen clear to local users and consistent with local cultural and national standards;
are the texts in the UI displayed correctly, are there any grammar, spelling or other errors, including brand consistency;

Task: Assessment, recommendations for UI improvement, error correction. 

This entailed developing a holistic metric with the following parameters: Ease of Navigation, Clarity of Purpose (General), Clarity of Purpose (Controls) to assess UI elements such as buttons; Intelligibility, Complexity, Locale Convention, Tone/Voice, Technical Quality, Language Appropriateness, Spelling and Grammar, Capitalization and Punctuation.

Workflow: 

In-depth discussion with the client about the purpose of the assessment
Development of an assessment metric in line with the goal
Creation of a user walkthrough scenario to navigate the UI and instructions on using the metric
Platform assessment by local users

Content Audit / Quality Evaluation of LinkedIn posts

Goal: Understand how LinkedIn posts look and feel to users from a sales perspective. Posts should interest the TA enough to follow the client’s account.

The TA is businesspeople of interest to the client, because the latter’s services can help them run their business.

Task: Assess text quality, carry out analytics on the basis of the assessment results with recommendations for improving the texts of posts.

Workflow: 

In-depth discussion with the client about the assessment purpose, creation of TA portrait;
Development of assessment metrics in accordance with the terms of reference. For this assessment, a holistic metric was developed with the Intelligibility, Complexity, Content Appeal, Locale Convention and Language parameters (only the texts of posts were provided, so we did not include such parameters as Ease of Navigation and Technical Issues);
Assessment of texts by local users.

Content Audit / Quality Evaluation

Goal: Determine game translation quality on the basis of 2000-word text fragments, and coordinate errors with the translation team.

Workflow: 

Creation of a three-dimensional metric with atomic parameters for line-level assessment of Adequacy, Intelligibility, Language Issues, Style, Locale Convention, Technical Issues, Terminology, Tone/Voice and two holistic parameters for overall assessment of fragments for Informativeness/Relevance and Consistency;
Drafting of instructions for reviewers describing both the assessment purpose, parameters and scales, and the work steps on the Logrus IT portal;
Text assessment by reviewers;
Arbitration (counterarguments from translators);
Responses to arbitration.

Get a Quote

How it differs from the traditional approach

Traditional metrics permit assessment of text only, exclusively at the sentence level, with errors typologized into standard categories and assigned a severity level.

Assessment of any content and visual 
We assess not only text strings, but also website, platform and application interfaces, marketing and advertising materials (posts on social networks, announcements, ads), articles, plus audio and video materials.
Multidimensional assessment of content, from different perspectives
We go beyond mere grammar and compliance. Our multidimensional approach measures such factors as relevance, engagement and intelligibility, among others, providing a complete picture of how effective your content is.
Custom metrics
We can create metrics to suit your goals and requirements, and adjust the complexity of the assessment scale, saving you time and expense.
Any type of assessment
You receive not just an assessment, but recommendations on how to improve your product.
Achievable goals: 
- Assessment for quick decision making on product quality;
- In-depth assessment for content correction and recommendations for improving the product;
- Assessment for subsequent AI training;
- Assessment for statistics collection (for example, to find out which users like your product and which don’t), etc.

LANGUAGE QUALITY ASSURANCE

Multidimensional metrics

Each object can be assessed from multiple perspectives:

Atomic approach (at the segment/sentence level).
The metric contains:
- Customizable error catalog;
- Customizable error severity scale;
- Assessment thresholds depending on expectations and content type;
- Assessment rating formula with customizable normalization factor (allowable number of errors per certain number of words).
Holistic approach (at the level of the object as a whole).
The metric contains:
- Customizable individual parameters for product assessment;
- Customizable scales;
- Assessment thresholds depending on expectations and content type.

LQA metrics

Examples of Parameters

Intelligibility: how readily the target audience (TA) perceives the content (sentence, line, unit, screen, form, etc.). This parameter measures the extent to which the content is clear and readable, is free of double meanings, and sounds natural.
Complexity: how readily the TA understands the content, given the expected level of technical knowledge or area of expertise.
Content Appeal: how appealing the content is to the TA. Above-average appeal can be achieved through the use of scenarios.
Locale Convention: adherence to date formats, addresses, currencies, calendar, time, names, numerals, phone numbers, units of measure, punctuation rules, keyboard shortcuts, sorting and text markup.
Tone of Voice: how the overall style matches the characteristic tone/voice of the product or service in the given context. The TA must not experience any negative connotations or feelings of inadequacy.

Why Logrus IT?

Our LQE team includes specialists with many years of experience in the linguistic testing field. We complete about 300 LQE projects a year and evaluate translations from over 80 languages. Microsoft, McAfee, Nexters, and other companies use our services on a regular basis.

Logrus IT’s experts use a proprietary method called the Triangle of Quality to evaluate translation quality. We use this method to determine the overall quality of a translation, as well as the severity of each individual grammatical, stylistic, and terminological error. Moreover, this method can be flexibly adapted to meet the needs of various kinds of materials, as well as limited time frames and budgets.

In addition to LQE, we can also:

Develop an individualized LQE strategy for you based on our own method
Help you embed quality assurance within your company using tools and utilities developed by Logrus IT, as well as third-party solutions

If our LQE evaluation determines that a translation’s quality is insufficient, Logrus IT’s specialists can edit your text or retranslate it from scratch. You can find a full list of our services here.

FAQ — Language Quality Evaluation (LQE)

What is Language Quality Evaluation (LQE)?
LQE is a structured process that evaluates the linguistic accuracy, consistency, and compliance of translated or localized content.
What does LQE include?
Linguistic testing, terminology and style verification, functional testing of localized interfaces, and structured QA reporting.
How does linguistic testing differ from functional testing?
Linguistic testing checks language accuracy and context. Functional testing verifies that localized content displays and operates correctly within the product.
When is LQE required?
For software, websites, mobile applications, marketing materials, and large-scale multilingual projects with defined quality requirements.
How are LQE results documented?
Issues are categorized, quantified, and documented in QA reports to support corrective actions and quality monitoring.