Translate | ISO 9241-220:2019 Human-Centred Design Processes, Usability Evidence and UX Handover

Translating ISO 9241-220:2019 human-centred design documents requires preserving what users actually did, what designers intended and what an organisation has demonstrated. People looking for human-centred design process translation, ISO 9241-220 UX audit translation, usability research report translation or multilingual design handover translation often face documents full of similar words—need, requirement, outcome, activity, observation and approval—that have very different roles. A polished UX summary can become false if ‘users requested’ turns into ‘users successfully completed’.

ISO 9241-220:2019 describes processes and outcomes for enabling, performing and assessing human-centred design within organisations over the life cycle of interactive systems. The official ISO public scope addresses process purposes, benefits, outcomes, activities and work products, and describes use by small and large organisations, including for internal systems, products, services and procurement. Its focus is the human-centred aspects of design, not a blanket statement about software security, engineering reliability or general business quality.

This practical guide explains how to translate multilingual user research, research consent, usability findings, design rationale, accessibility observations, process assessments, procurement requests and UX handover notes without inventing results. Fictional examples show how an international reader can distinguish an intended outcome from verified evidence and a participant’s statement from a researcher’s inference. No fictional example claims to reproduce the purchased ISO text or establish that a product meets a certification requirement.

Quick answer: what must a reliable translation preserve

Keep user groups, tasks, environments and accessibility needs in context.
Separate stakeholder preferences from observed user needs.
Distinguish process activities, work products, intended outcomes and verified results.
Preserve participant quotations, researcher interpretations and uncertainty levels.
Retain test sample sizes, task definitions and outcome metrics.
Label prototype findings separately from released-product evidence.
Respect privacy, research consent and accessibility accommodations.
Have UX owners and target-language users test the final document or interface.

The official publication, its role and its boundaries

The official ISO catalogue entry identifies the cited edition and describes the public scope. An ISO title does not, by itself, demonstrate that a particular supplier, individual, machine, process, test or product conforms. Keep the cited standard separate from local regulations, contracts, manufacturer instructions and the competent authority’s actual decision. Where the exact wording of a proprietary requirement or table matters, consult a licensed copy and the responsible reviewer instead of treating this explanatory article as the standard.

The intended readers include UX researchers, product managers, localisation teams, procurement officers, interaction designers and multilingual service teams. They may need different levels of explanation, but they should reach the same conclusion about what happened and what remains uncertain. A translator can add a clearly identified explanation of a specialised word; that is different from adding a technical obligation or declaring success that the underlying evidence does not establish. The goal is not to make the report sound more impressive. It is to make the original facts usable across languages without shifting responsibility.

A question-led method for readers who need an answer immediately

Before working on an human-centred design processes for interactive systems document, identify what the reader actually wants to decide. Is the file an operating instruction, procurement criterion, safety warning, result, field observation, proposed change or authorised sign-off? List the sentence or table that permits each decision. Translate the narrow answer first, then provide the background needed to avoid misunderstanding. If the source does not permit a decision, say so plainly; a translation cannot create missing authority.

There is a difference between a source fact, an editorial explanation and an unresolved question. Label each separately. A sentence such as ‘the sample was inspected’ needs its sample identity; ‘the procedure was followed’ needs its procedure edition; and ‘the result was acceptable’ needs the criterion and approved decision. The reader should not have to guess whether those details came from the original record or from a translator’s helpful but unsupported inference.

The connected technical reasoning

User needs are claims about real contexts

An organisation may think its customers want faster onboarding, while actual observations show difficulty understanding a confirmation step. Translate each claim with its evidence source. A stakeholder hypothesis is not a verified need, and a need is not automatically a particular interface solution.

Create a context-of-use sheet with user group, task, equipment, environment, capability and relevant constraints. Translation becomes easier when ‘user’ refers to a defined population rather than an imaginary average person. Preserve exclusions too: a study of experienced office staff does not demonstrate suitability for first-time users with low literacy.

Process outcome is different from activity

A workshop, interview or usability test is an activity. A research note, design specification or finding report is a work product. A demonstrated capability or achieved result is an outcome. ISO 9241-220’s public description explicitly distinguishes processes, outcomes and typical work products.

A translation that says a workshop ‘delivered an accessible service’ can be seriously misleading when it only produced a backlog of hypotheses. Keep activity verbs such as conducted, drafted, reviewed and tested separate from completed outcomes such as verified, implemented or accepted under named criteria.

Do not mistake usability for popularity

Participants may like a screen without successfully performing a critical task. A task completion rate, time on task, observed error, satisfaction score and qualitative quote measure different aspects of experience. Translators should preserve method, participant group and task definition rather than converting every positive result into ‘easy to use’.

When the data is incomplete, say so. A pilot with six volunteers can uncover meaningful problems but cannot establish universal performance. Preserve confidence limits and caveats in the original report; an internationally polished summary is not permission to erase the study’s weaknesses.

The user journey has boundaries

Journey maps are useful but can conceal who owns each step and which parts were directly observed. Translate touchpoints, channels, tasks and backstage dependencies with careful actor attribution. A service failure caused by an identity-check requirement cannot be reduced to ‘confusing copy’ without evidence.

Keep cultural context in view. In some regions a postal address, surname or mobile number is not structured in the way the design assumes. Translation should preserve the actual user problem and flag data model assumptions to the product team, not disguise them with locally familiar examples that mask the original limitations.

Research ethics travel with the text

Interview transcripts, recruitment records, consent forms and usability videos may contain personal or sensitive information. Translation increases the potential audience for those records. Determine lawful access, permitted purpose, storage controls, de-identification and retention rules before exposing the material to vendors or automated services.

Do not turn ‘pseudonymised’ into ‘anonymous’, or ‘participants agreed to this study’ into unrestricted permission to republish quotations. Translate withdrawal rights and research conditions accurately and refer questions to the responsible ethics or privacy owner.

Accessibility is not a sticker

An accessibility review must identify the user capability, task, assistive technology, observed barrier and test condition. A high contrast button does not by itself prove a complete service accessible. Translate evidence, not an unqualified label. Distinguish keyboard access, screen-reader semantics, language clarity and cognitive load.

A localisation can also create new barriers. Text expansion may clip button labels, reading order may change for a right-to-left language, and screen-reader labels can become inconsistent with visual controls. Test the actual language build with affected users or competent accessibility reviewers.

Process capability needs evidence

A team may describe its practice as mature, consistent or human-centred, but the assessment needs evidence that activities are planned, performed, evaluated and improved. The official ISO description links the processes to the ability to assess an organisation’s human-centred practices.

Keep version history and roles. A team that ran user interviews once should not be portrayed as running continuous validated research throughout an entire product portfolio. Distinguish current organisational practice from an aspirational future process.

Handover is a design experiment’s memory

Product requirements, interaction decisions, study findings, dissenting comments and open questions must remain understandable when a system is procured, outsourced or maintained in another language. Preserve why a decision was made, the alternatives considered and the evidence still missing.

The final translator should conduct a reconciliation meeting with the UX and product owners. Verify that every critical claim is linked to a research artifact, design revision or decision log. Keep the translation’s own change history so the next team can trace how key terms were approved.

Twenty-two worked translation decisions

The following are constructed learning cases, not quotations from an ISO publication or evidence about real products. They are designed to expose a recurrent challenge: the same term can be a label, a variable, a reported observation, an authorised instruction or an unverified claim. Each case gives a source situation, the likely mistranslation and an evidence-led repair. In a live job, use the real source document rather than adopting any fictional values here.

1. Human-centred design

Meaning and boundary. An organisational approach that considers human needs throughout an interactive system’s life cycle. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Fictional document situation. A vendor advertises a single focus group as full human-centred design adoption. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Where a fluent translation fails. A translation upgrades a one-off workshop into enterprise capability. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

A defensible rendering. State the actual activities and documented organisational processes. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

One final reader check. Ask a reader to identify the named object, the condition under which the statement holds and the action it authorises. If the answer differs from the source, revise the exact line rather than adding a vague disclaimer elsewhere. This approach tests meaning transfer, not just grammatical fluency. It also leaves a repeatable correction rule for future versions, languages and staff rather than depending on one translator’s memory.

2. Context of use

Understand the source term. The users, tasks, equipment and physical or social environment relevant to interaction. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Consider this handover. A kiosk test occurs in a quiet office while production units will stand in noisy stations. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Notice the risky implication. The report claims performance was established in real station conditions. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

What to retain. Keep laboratory and real-world contexts distinct. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

Verify in context. Compare the target-language table or label against the controlled source with both a domain reviewer and a second linguist. Record the revision number and whether the issue affected a value, a qualification or an action. This approach tests meaning transfer, not just grammatical fluency. It also leaves a repeatable correction rule for future versions, languages and staff rather than depending on one translator’s memory.

3. User need

Why the distinction exists. A human requirement grounded in a situation, not an already chosen screen component. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

A realistic example. A researcher observes that people need to recover from mistaken form entries. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

The dangerous shortcut. A translation states users specifically requested one button design. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

The preferred treatment. Render the underlying need and keep the proposed solution separate. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

Reviewer’s question. Read the sentence without its surrounding paragraph, as a busy buyer or operator might. If that isolated sentence promises a stronger result or wider scope, restore the missing qualifier in the same visual location. This approach tests meaning transfer, not just grammatical fluency. It also leaves a repeatable correction rule for future versions, languages and staff rather than depending on one translator’s memory.

4. Stakeholder requirement

Define the controlled meaning. A demand or constraint from a sponsoring organisation or stakeholder. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Imagine the record. A legal department wants an acknowledgement screen before submission. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

What goes wrong. The target summary attributes that choice to users. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

How to fix it. Name the stakeholder and note whether user validation occurred. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

Test the conclusion. Give the page to a competent target-language reader and ask for a short paraphrase of the technical decision. Compare their interpretation with the authorised reviewer, not merely with a dictionary definition. This approach tests meaning transfer, not just grammatical fluency. It also leaves a repeatable correction rule for future versions, languages and staff rather than depending on one translator’s memory.

5. Process outcome

Meaning and boundary. A result expected or achieved from a defined organisational process. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Fictional document situation. An audit lists outcomes and evidence gaps. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Where a fluent translation fails. The translation treats an intended outcome as already achieved. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

A defensible rendering. Preserve modal verbs and assessment status. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

6. Design activity

Understand the source term. Work performed such as interviewing, prototyping or observing tasks. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Consider this handover. A project plan proposes five interviews next month. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Notice the risky implication. The translation says interviews were completed. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

What to retain. Keep planned, in-progress and completed tense distinctions. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

7. Work product

Why the distinction exists. A produced artifact such as a report, journey map or requirements set. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

A realistic example. A team creates a draft usability report. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

The dangerous shortcut. A translation calls the draft a validated design. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

The preferred treatment. Preserve document type, status and approval. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

8. Usability goal

Define the controlled meaning. A defined user-performance intention rather than a vague hope. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Imagine the record. A product targets completion of a critical payment task without assistance. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

What goes wrong. The target version says the goal has been met. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

How to fix it. Separate goal from measured outcome and named test. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

9. Task completion

Meaning and boundary. Whether participants achieved a defined task under stated conditions. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Fictional document situation. Seven of ten fictional participants complete an application flow. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Where a fluent translation fails. The press copy says 70 percent of all customers succeed. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

A defensible rendering. Retain sample, task, method and uncertainty. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

10. Time on task

Understand the source term. Duration measured for an identified activity, often with start and end conventions. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Consider this handover. A usability test times navigation from landing page to confirmation. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Notice the risky implication. A translation uses the clock time of the session instead of elapsed seconds. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

What to retain. Specify the measured interval and unit. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

11. Error rate

Why the distinction exists. Frequency of a defined error within a tested task or participant set. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

A realistic example. Researchers log wrong-field entries and recovery attempts separately. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

The dangerous shortcut. A translation combines mistakes and abandoned sessions into one rate. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

The preferred treatment. Maintain denominator, error definition and observation method. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

12. Participant quote

Define the controlled meaning. A person’s reported words that may not represent the research team’s conclusion. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Imagine the record. A volunteer says a screen ‘felt faster’. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

What goes wrong. The translated report states objective performance improved. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

How to fix it. Keep quotation attribution and distinguish feeling from measurement. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

13. Researcher interpretation

Meaning and boundary. An analytical explanation inferred from observations. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Fictional document situation. A UX researcher suspects terminology caused task failures. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Where a fluent translation fails. The target report calls the cause proven. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

A defensible rendering. Preserve ‘may’, ‘suggests’ and the alternatives not ruled out. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

14. Prototype

Understand the source term. A design artifact tested before product release, with limits of fidelity. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Consider this handover. A clickable mock-up has nonfunctional payment behaviour. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Notice the risky implication. The translation calls it a launched production system. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

What to retain. Keep the prototype type and unimplemented functions visible. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

15. Design iteration

Why the distinction exists. A traceable change made in response to feedback or evidence. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

A realistic example. A button label is changed after usability sessions. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

The dangerous shortcut. The summary says the change solved all errors without retest. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

The preferred treatment. Describe implemented change and state whether it was evaluated. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

16. Accessibility issue

Define the controlled meaning. A barrier linked to a user capability, tool or task. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Imagine the record. Screen-reader users cannot hear an error message in a fictional test. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

What goes wrong. The translation calls this a minor visual problem. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

How to fix it. Preserve assistive-technology context and severity evidence. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

17. Reading order

Meaning and boundary. The sequence content is encountered visually or by assistive technology. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Fictional document situation. A multilingual screen moves a field visually but not in keyboard focus order. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Where a fluent translation fails. The target sign-off says the flow is fully localised. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

A defensible rendering. Test both reading direction and operational focus sequence. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

18. Cognitive load

Understand the source term. The demands a task places on attention or memory under specific conditions. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Consider this handover. A form requires remembering an earlier reference number. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Notice the risky implication. The translation claims users are cognitively impaired. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

What to retain. Describe the task demand without pathologising participants. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

19. Pseudonymisation

Why the distinction exists. Replacing direct identifiers while retaining a possible means of re-linking data. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

A realistic example. Interview files use participant codes with a separate protected key. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

The dangerous shortcut. The translation says the files are completely anonymous. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

The preferred treatment. Preserve the actual data-protection state and access limitations. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

20. Research consent

Define the controlled meaning. Permission for defined research purposes, not automatically for public distribution. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Imagine the record. Participants agree to an internal usability review. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

What goes wrong. A translator sends raw recordings to an external public service. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

How to fix it. Maintain permitted processing scope and seek privacy approval first. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

21. Process capability

Meaning and boundary. The organisation’s demonstrated ability to carry out recurring human-centred work. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Fictional document situation. A supplier claims mature UX practice from one old project. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Where a fluent translation fails. The translation certifies organisational maturity without current evidence. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

A defensible rendering. Report assessment basis and avoid unsupported certificates. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

22. Procurement handover

Understand the source term. Transfer of requirements, evidence and responsibility to a new supplier or owner. A translator should first decide whether the phrase names a classification, measured fact, proposed operation, observation or final disposition. Those categories are not interchangeable simply because their English terms appear close together. Where a term is an identifier or controlled category, retain that identity and explain it rather than replace it with a fashionable approximate synonym.

Consider this handover. A public agency hands off a multilingual citizen-service design. In this example the document would normally also have an author, date, revision and evidence trail. Match each of those to the claim being translated. Treat nearby numbers and table headings as part of the sentence, not as decoration, because an international reader may rely on the translated version without seeing the original sheet.

Notice the risky implication. A translated package removes open research risks to appear complete. That is more than a stylistic issue. It alters the apparent object, timing, population, physical action or level of proof. An attractive target-language sentence is not successful if it changes the decision a professional reader would make. Mark the disputed source span and identify who can authoritatively resolve it.

What to retain. Keep unresolved findings, owners and next verification steps. When the source is ambiguous, do not hide the ambiguity with fluent prose. Keep an issue record with the exact source wording, the proposed interpretation, the technical owner and the approved resolution. If a more explanatory expression is needed for a nonspecialist, make it clear that the explanation is about the documented fact and not a new operational command.

Eight realistic editorial disputes and the safer response

1. A stakeholder asks to convert ‘suggested’ into ‘proved’

Keep the evidence status and request a new authorised conclusion if the research genuinely supports it. The important distinction is between solving the reader’s communication problem and silently rewriting the underlying evidence. The bilingual document should remain a reliable record of its source, even when that source is incomplete, inconvenient or written in an awkward style. Preserve the responsible role, any stated condition and the exact version being corrected.

A useful escalation note contains the source phrase or field, a clear description of the risk, two plausible interpretations where relevant, and the decision requested from the technical owner. Avoid an email that merely says ‘please clarify’. Once a decision is returned, incorporate it into the controlled glossary and regression checks so the same error does not recur in another language or release.

2. A study sample is too small for a marketing claim

Report the actual group and task; do not claim universal usability. The important distinction is between solving the reader’s communication problem and silently rewriting the underlying evidence. The bilingual document should remain a reliable record of its source, even when that source is incomplete, inconvenient or written in an awkward style. Preserve the responsible role, any stated condition and the exact version being corrected.

3. A translated button no longer fits on a phone

Change the design with product approval, not the essential meaning of the instruction. The important distinction is between solving the reader’s communication problem and silently rewriting the underlying evidence. The bilingual document should remain a reliable record of its source, even when that source is incomplete, inconvenient or written in an awkward style. Preserve the responsible role, any stated condition and the exact version being corrected.

4. A user quote could identify a vulnerable participant

Apply approved de-identification and confidentiality safeguards before translation or publication. The important distinction is between solving the reader’s communication problem and silently rewriting the underlying evidence. The bilingual document should remain a reliable record of its source, even when that source is incomplete, inconvenient or written in an awkward style. Preserve the responsible role, any stated condition and the exact version being corrected.

5. Research terms have no exact local equivalent

Use a controlled descriptive term and note how it corresponds to the source concept. The important distinction is between solving the reader’s communication problem and silently rewriting the underlying evidence. The bilingual document should remain a reliable record of its source, even when that source is incomplete, inconvenient or written in an awkward style. Preserve the responsible role, any stated condition and the exact version being corrected.

6. The team translated an accessibility claim but did no target-language test

Label the claim’s evidence boundary and commission appropriate local testing. The important distinction is between solving the reader’s communication problem and silently rewriting the underlying evidence. The bilingual document should remain a reliable record of its source, even when that source is incomplete, inconvenient or written in an awkward style. Preserve the responsible role, any stated condition and the exact version being corrected.

7. A prototype success is described as production success

Correct the lifecycle stage and reference the tested build. The important distinction is between solving the reader’s communication problem and silently rewriting the underlying evidence. The bilingual document should remain a reliable record of its source, even when that source is incomplete, inconvenient or written in an awkward style. Preserve the responsible role, any stated condition and the exact version being corrected.

8. An external vendor receives the UX archive

Verify permitted access, contracts, version controls and privacy before sharing. The important distinction is between solving the reader’s communication problem and silently rewriting the underlying evidence. The bilingual document should remain a reliable record of its source, even when that source is incomplete, inconvenient or written in an awkward style. Preserve the responsible role, any stated condition and the exact version being corrected.

A complete multilingual production and review workflow

Register the source. Capture title, edition, relevant amendment, file hash or revision, document owner, intended market, language direction and permissible use. Identify separate source records when there is more than one authoritative version.
Map meaning to decisions. For each critical sentence, ask what it permits a reader to do or conclude. Tag terminology as classification, quantity, process step, evidence, claim, limitation or disposition before rewriting it.
Control symbols and numbers. Freeze identifiers, dimensions, dates, formulae, references, units and decimal conventions in a QA register. Record approved conversions rather than quietly altering them during stylistic editing.
Draft for the target reader. Use clear conversational technical prose and introduce the necessary concepts in the order a new reader needs them. Short sentences are useful, but they should not delete relationships such as exceptions, dependencies or conditions.
Resolve uncertainties. Raise precise source queries about missing criteria, conflicting values or unidentifiable actors. Keep unresolved issues visibly open rather than guessing from another project’s documentation.
Review through two lenses. Have a competent domain owner check technical meaning and a proficient language editor check readability, grammar, terminology and cultural pragmatics. Neither role substitutes for the other.
Test the final format. Examine rendered PDFs, tables, warnings, captions, screen strings and right-to-left text where relevant. Check line wrapping, embedded figures, page references and any inaccessible wording.
Approve and preserve the record. Store approval date, signatory role, accepted exceptions and final file identity. If the source changes, reopen its dependent translated sections and issue a traceable revision.

For high-consequence documents, a second independent interpretation test is particularly useful. Hide the source text from a competent target-language reader and ask them to state which entity or product is affected, what was measured or instructed, which exceptions apply, and what action the document allows. Then compare their answer with the source owner. This catches false certainty, missing negation and displaced conditions that routine spelling checks often miss.

Frequently asked questions

Is ISO 9241-220 only for software?

No. Its public description concerns interactive systems and applicable human-centred processes, including some non-computer-based contexts. For a real operational, contractual or assurance decision, connect the answer to the actual controlled source and obtain approval from the qualified owner. A smooth translation can help people understand evidence, but it cannot repair a missing measurement, grant a permission or remove a legal requirement.

Are completed interviews proof of usability?

No. Interviews are research activities; observed task performance and findings require separate evidence. For a real operational, contractual or assurance decision, connect the answer to the actual controlled source and obtain approval from the qualified owner. A smooth translation can help people understand evidence, but it cannot repair a missing measurement, grant a permission or remove a legal requirement.

Can a translation call a prototype accessible?

Only to the extent supported by a clearly defined assessment; accessibility depends on specific barriers and contexts. For a real operational, contractual or assurance decision, connect the answer to the actual controlled source and obtain approval from the qualified owner. A smooth translation can help people understand evidence, but it cannot repair a missing measurement, grant a permission or remove a legal requirement.

What is a work product?

An artifact produced by a process, such as a report or requirements document, not automatically a verified outcome. For a real operational, contractual or assurance decision, connect the answer to the actual controlled source and obtain approval from the qualified owner. A smooth translation can help people understand evidence, but it cannot repair a missing measurement, grant a permission or remove a legal requirement.

Should quotes be translated literally?

Preserve meaning, attribution, register and uncertainties, with appropriate privacy and research review. For a real operational, contractual or assurance decision, connect the answer to the actual controlled source and obtain approval from the qualified owner. A smooth translation can help people understand evidence, but it cannot repair a missing measurement, grant a permission or remove a legal requirement.

Can machine translation process participant recordings?

Only under authorised data-handling conditions and with qualified review. For a real operational, contractual or assurance decision, connect the answer to the actual controlled source and obtain approval from the qualified owner. A smooth translation can help people understand evidence, but it cannot repair a missing measurement, grant a permission or remove a legal requirement.

How does the article relate to the master translation guide?

It addresses the narrow problem of human-centred design documentation, while the master retains general translation architecture. For a real operational, contractual or assurance decision, connect the answer to the actual controlled source and obtain approval from the qualified owner. A smooth translation can help people understand evidence, but it cannot repair a missing measurement, grant a permission or remove a legal requirement.

How can readers verify a UX claim?

Follow its task definition, participant group, method, evidence artifact, version and authorised decision. For a real operational, contractual or assurance decision, connect the answer to the actual controlled source and obtain approval from the qualified owner. A smooth translation can help people understand evidence, but it cannot repair a missing measurement, grant a permission or remove a legal requirement.

Final release checklist

Keep user groups, tasks, environments and accessibility needs in context.
Separate stakeholder preferences from observed user needs.
Distinguish process activities, work products, intended outcomes and verified results.
Preserve participant quotations, researcher interpretations and uncertainty levels.
Retain test sample sizes, task definitions and outcome metrics.
Label prototype findings separately from released-product evidence.
Respect privacy, research consent and accessibility accommodations.
Have UX owners and target-language users test the final document or interface.
All controlled identifiers, numerical units, figures and captions match the approved original.
The target-language layout carries every safety, scope and uncertainty qualifier.
The technical owner and independent language reviewer have signed the final version.
Any unanswered technical questions remain recorded and have not become implied approvals.
Readers can distinguish the official ISO summary, fictional learning examples and real source evidence.

Sources, limitations and next useful reading

Official publication: ISO 9241-220:2019: ISO catalogue entry. This article is an independent educational explanation of human-centred design processes for interactive systems document translation based on public scope information and constructed scenarios; it does not reproduce copyrighted tables, replace the standard itself, or certify compliance in any jurisdiction.

For general language-to-language transfer and document architecture, follow Master Art of Translation — EDKSG-TRANS-MASTER-WORLD-000. For precise technical vocabulary, concepts and the relationship of words to real-world experience, use the protected Vocabulary Learning Hub. For English register, qualification, agency and sentence-level interpretation, see How English Works. These established resources remain the canonical owners of general translation, vocabulary and English-learning methods. This page solves only the specialist task named in its title.

The strongest translation lets the next reader ask the right question, locate the relevant evidence and avoid a false decision. It does not promise certainty where the record offers only a partial observation. Keep terminology stable, give examples that illuminate rather than exaggerate, and make every consequential qualification as visible as the headline conclusion.