How do you translate ISO 11620:2023 library performance indicators when an international library report contains physical visits, digital sessions, circulation transactions, population denominators and cost-per-use calculations? The important question is not whether the target-language sentence sounds professional. It is whether an overseas librarian, researcher or public decision-maker would count the same event, use the same denominator and reach the same cautious interpretation. Library statistics translation is therefore a problem of preserving definitions, sources and reporting boundaries as much as translating words.
This practical guide explains library performance indicator translation, library statistics terminology, collection loans, online usage metrics, user-satisfaction surveys and multilingual annual-report localisation. It is designed for translators, library managers, municipal reporting teams, academic libraries, quality reviewers and readers comparing public services between countries. The article concentrates on the specific task of translating and checking an indicator report; it is not another general manual on translation or a replacement for the official ISO standard.
Quick answer: Preserve the official indicator name together with its locally approved definition; record the numerator, denominator, unit, collection method, institution and time period; keep visits distinct from unique people, loans distinct from renewals and digital sessions distinct from downloads; retain uncertainty, exclusions and changes of methodology; and verify each translated number against its original source table. If an unfamiliar target-language label has no exact equivalent, define it clearly rather than inventing an apparently familiar statistic.
- First identify what was observed, counted or surveyed.
- Translate the denominator and reporting period alongside the number.
- Separate activity, output, outcome and satisfaction.
- Keep methodological footnotes in translated charts and dashboards.
- Obtain bilingual and statistical review before a translated report is published.
What ISO 11620:2023 covers—and what this guide does not claim
The International Organization for Standardization publishes ISO 11620:2023, Information and documentation — Library performance indicators, fourth edition, issued in June 2023. Its public scope describes requirements for library performance indicators, a set of indicators and guidance for their implementation; it also makes clear that not every indicator applies to every library. The official ISO 11620:2023 catalogue entry establishes the edition and scope. The full standard, local measurement manual and any current professional guidance must be consulted when an organisation needs the precise prescribed method. The examples below are original teaching cases and must not be mistaken for reproduced ISO formulas.
This distinction matters. A public library serves different aims from a national research archive or a small special library. The same visible word, such as use, may be measured with door sensors, transactions, website logs, manual reference counts or surveys. If a translator copies a number but changes the measured unit, the report ceases to describe the original reality. A translation can be fluent and still be numerically wrong. Good translation keeps the reader’s understanding attached to the actual observation.
The first decision: preserve the measurement, not the most familiar word
Imagine receiving an annual report with a column titled visits. The instinct is to find an elegant equivalent for visits. The better first question is: what produces one unit in this column? A library entrance counter may register a second visit when someone leaves for lunch and returns. A website report may register a new session after inactivity. A university library may count occupancy snapshots. None of these automatically identifies unique people. If a target translation uses ‘readers’, it may introduce an identity claim the source never made.
A useful working note for every reported measure has seven fields: source name, definition, unit, numerator, denominator, period and limitations. Not every measure has a denominator, but the field should be deliberately marked not applicable rather than left forgotten. Add the collection owner and version where a dashboard evolves over time. Reviewers can then decide whether a target label reflects exactly the source definition or whether a one-sentence explanation is needed. This makes the translation auditable and reusable next year.
This method helps readers as well as publishers. A reader planning a library visit needs to know service hours, not an impressive total formed by adding unrelated online and physical opening times. A policymaker comparing branches needs to know how population served was defined. A manager assessing licences needs to understand whether a vendor counted downloads or sessions. A student researching a national system should be able to distinguish the institution’s achievements from the arithmetic used to describe them. Translation performs well when all these readers can follow the evidence without being forced to guess.
A workable terminology map for multilingual library reports
Library performance indicator
Meaning: An operationally defined measure used to evaluate an aspect of library service rather than a vague expression of success.
Where translations go wrong: The word performance can suggest financial profit or examination achievement in some languages. Translate it as the performance of a library service, retaining any formal label used by the institution.
Practical reading test: A library with more visits does not automatically have better outcomes; the report must say what was counted and why. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Indicator definition
Meaning: The rule specifying the phenomenon, calculation, exclusions, population, reporting interval and source for a reported value.
Where translations go wrong: A polished translation that omits the denominator or a footnote may transform a defined measure into a slogan. Keep the local definition and its version with the displayed indicator.
Practical reading test: Reviewers should reconstruct a reported value from the same input records after translation. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Reporting period
Meaning: The exact dates over which events and costs were collected, sometimes a fiscal year rather than a calendar year.
Where translations go wrong: Do not turn fiscal year 2025/26 into calendar year 2025. Translate whether an interval is inclusive, whether partial-year data were annualised and whether reopening months were excluded.
Practical reading test: A service that closed for renovation needs that context when monthly counts are compared. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Physical visits
Meaning: Entries or visits to a physical library location under the counting rule used by that institution.
Where translations go wrong: A door counter may count repeat entries, not distinct people. Do not convert entries into visitors, patrons or registered members without an identity-based counting method.
Practical reading test: The original equipment, branches, correction factors and closure days belong with the number. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Virtual visits
Meaning: Access to an online library site or service measured by a declared analytics method.
Where translations go wrong: Web sessions, page views, authenticated logins and unique devices have different units and privacy implications. Resist translating all four as users.
Practical reading test: Digital measurement definitions change when analytics platforms or cookie settings change. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Loans
Meaning: Borrowing transactions or circulation events counted under a disclosed lending policy.
Where translations go wrong: Separate new loans, renewals, short-loan services, e-book checkouts and equipment lending where the report does. One word for all of them hides service design.
Practical reading test: Check whether a renewal creates a new transaction in the original data. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Turnover
Meaning: A ratio describing lending or use against the relevant stock or another named base, if that indicator is used.
Where translations go wrong: It is not automatically business turnover, a currency amount or a measure of staff replacement. A translator needs the stated numerator and denominator before choosing a target phrase.
Practical reading test: A numerical ratio of 2.5 should not silently become 2.5 percent. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Population served
Meaning: The defined community for which service is planned, not necessarily every inhabitant of a city.
Where translations go wrong: Municipal population, registered population, university enrolment and eligible residents are different bases. Avoid substituting the most familiar demographic denominator.
Practical reading test: Describe service catchment and geographic boundary in words accessible to international readers. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Active borrower
Meaning: A borrower meeting the institution’s recorded activity criterion over a stated period.
Where translations go wrong: A person with a valid membership card may not be active; a person with several transactions may still count once. Membership and active use are not synonyms.
Practical reading test: Ask whether a borrower identifier is deduplicated across branches. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Registered user
Meaning: A person with a current or historical account according to the library’s administrative definition.
Where translations go wrong: Accounts can expire, duplicate or belong to institutions. Do not claim every registered account represents one currently engaged reader.
Practical reading test: Preserve whether membership is a stock count on a date or an accumulation throughout a year. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Opening hours
Meaning: The counted hours during which a library or service is available under a stated rule.
Where translations go wrong: Staff work hours, building opening hours, live-chat support and overnight self-service access may be separate indicators. A translation should not collapse them.
Practical reading test: For multisite comparisons identify exceptional holidays and service restrictions. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Collection size
Meaning: A specified quantity of collection resources, such as physical volumes or licensed digital materials, with counting rules.
Where translations go wrong: Titles, copies, items, subscriptions and accessible databases can yield different totals. One digital platform is not one individual ebook by default.
Practical reading test: Keep ownership and licensed-access language separate. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Electronic resource use
Meaning: Observed access or retrieval events for licensed or local digital resources under a given statistical method.
Where translations go wrong: A download, search, session, request and view are not interchangeable. Vendor reports can include automated activity or platform-specific units.
Practical reading test: Cite the reporting standard or vendor definition actually applied rather than assuming one universal metric. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Cost per use
Meaning: A defined cost numerator divided by a defined use denominator, frequently presented as a value in a currency per event.
Where translations go wrong: Do not rename it return on investment or translate a lower figure as proof of better educational value. Accounting boundaries change results.
Practical reading test: The currency, period, included personnel costs and definition of use must accompany the ratio. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
User satisfaction
Meaning: Feedback captured by a specified survey or interview method, not a direct measurement of learning.
Where translations go wrong: A mean score of 4.2/5 is not 84 percent satisfied unless the reporting method explicitly establishes that relationship. Preserve scale and response-rate information.
Practical reading test: If respondents were voluntary, describe the sample without claiming it represents all visitors. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Accessibility measure
Meaning: A metric or observation about whether service arrangements enable access for people with different needs.
Where translations go wrong: An accessible entrance, captioned video, assistive technology or accessible website measure cannot be reduced to a blanket claim that the whole library is accessible.
Practical reading test: Translate the particular feature, population and test evidence. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Benchmarking
Meaning: Comparing measured results with a defined reference group while controlling important differences.
Where translations go wrong: Two institutions may share an indicator name but use incompatible branch scopes or periods. Translate methodological cautions instead of producing a league table.
Practical reading test: A good comparison states what the numbers cannot establish. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Data quality note
Meaning: An explanation of missing records, revised figures, estimation procedures or sensor anomalies that affect confidence.
Where translations go wrong: Notes marked preliminary, estimated or revised must survive publication. Never smooth them into unqualified exact values for stylistic neatness.
Practical reading test: Treat caveats as part of the data, not editorial clutter. In a bilingual review, the translated label and its explanatory note should lead the analyst back to exactly the same source field, rather than to a merely plausible dictionary synonym.
Fourteen original worked translation cases
The following figures, institutions and situations are illustrative. They are designed to show how translation changes a decision when a numerical definition, boundary or claim is lost. They are not ISO-certified indicator examples, actual library results or quotations from institutional records. Each case starts with what the source actually measures and then follows the correction a careful reviewer would make.
1. Door counter after an entrance refurbishment
Source situation. A branch recorded 82,400 automatic entrance events in 2025, but the system was recalibrated on 1 July. The source calls the total measured entries with a comparability note.
The translation trap. The draft target version says 82,400 people visited and adds that demand increased. Neither unique identities nor a comparable prior series was measured. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Retain measured entrance events, dates, device change and the comparison warning; ask the analyst whether a revised series has been validated. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. A parent, funder or local official can now understand why the number is useful without mistaking traffic for unique readership. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
2. University library examination season
Source situation. The university reports 19,300 study-area gate movements during a fortnight and 3,400 booking confirmations for seats. The report measures different behaviours.
The translation trap. One translator merges them as 22,700 visitors and another calls seat bookings attendance. That falsely assumes each booking led to a visit. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Keep two named series and the relevant denominators; where a summary is necessary, describe use patterns rather than summing incompatible counts. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. A figure becomes understandable when a reader can see what instrument produced it and whether events overlap. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
3. Library network and catchment population
Source situation. Five branches are funded by a council serving 118,000 residents. A partner mobile library makes visits outside the municipality.
The translation trap. The draft translates service population as customers and includes another district’s 14,000 residents in the denominator without an approved rule. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Preserve official catchment definition, separately describe the mobile service and ask the administrator for the denominator specification. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. International comparison is fairer when boundaries are transparent, particularly in cities with cross-border or metropolitan systems. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
4. E-book platform checkouts and renewals
Source situation. One vendor reports 46,000 loans and 11,000 automated renewals; another reports 13,500 content accesses. Each uses different event logic.
The translation trap. The target-language brochure calls every event a newly borrowed book and combines all three figures as 70,500 readers. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Identify the vendor reporting units, do not count persons from events, and separate renewals and content-access interactions in tables. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. The result may be less spectacular, but it represents the actual interaction instead of creating fictitious readers. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
5. Reference desk questions versus directional help
Source situation. The public library counts 6,200 reference enquiries under a defined staff classification and records 4,100 basic directional queries separately.
The translation trap. A broad local translation uses enquiries for both categories and implies the full 10,300 total required research assistance. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Translate categories as different service activities with their local definitions; document any classification change between years. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. Management can then plan training and staffing using meaningful demand instead of a mixed transaction count. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
6. Online access following analytics migration
Source situation. The report states that online sessions rose from 64,000 to 71,000 while a new consent mechanism began in April.
The translation trap. An editor removes the analytics-change caveat and headlines that the digital audience grew by eleven percent. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Preserve both numbers and the methodological change; if analysts cannot establish comparability, describe recorded counts without treating their ratio as proven growth. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. This approach respects reader trust, particularly where privacy rules make digital metrics uncertain. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
7. Children’s reading programme participation
Source situation. A branch delivered 24 workshops with 510 attendance instances and 290 unique registrations for children.
The translation trap. A translation says 510 children completed the programme, though some children attended several sessions and no completion criterion was recorded. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Distinguish attendance instances, distinct registered participants and any verified completions; retain age ranges only if supported. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. Programme organisers can measure outreach and repeat engagement without inventing educational outcomes. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
8. Collection turnover for multilingual titles
Source situation. A report says a subset of 6,000 physical items generated 15,000 annual circulation events, giving an illustrative circulation-to-stock ratio of 2.5.
The translation trap. The translation makes it 2.5 percent utilisation and treats the ratio as a financial return. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. State the original ratio, its calculation and time interval; keep the distinction between titles, copies and circulation events. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. Readers can reproduce 15,000 divided by 6,000 and understand why turnover is a library-use concept. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
9. Cost per digital download
Source situation. A licensed resource had an allocated annual cost of $18,000 and 3,000 qualifying downloads, producing an illustrative $6 per download.
The translation trap. A promotional translation says the platform saves $6 per download, reversing cost into savings. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Retain $6 as a calculated unit cost, name the currency and denominator, and identify whether staffing, overhead and taxes were excluded. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. The report can inform budgeting without turning a financial measure into a value judgment. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
10. Reading-space occupancy study
Source situation. Hourly snapshots recorded seats occupied at specified times; an average of 72 percent was reported for weekdays only.
The translation trap. The translated infographic suggests the library is always 72 percent full, including weekends and holiday closures. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. State sampling dates, method, snapshot interval and the weekday scope; do not convert average occupancy into occupancy at every moment. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. Space planners can then assess peak demand alongside mean use rather than promising a misleading fixed experience. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
11. Customer survey and nonresponse
Source situation. A voluntary survey had 210 responses from a mailing list of 3,000 recipients, with a satisfaction mean of 4.4 on a five-point scale.
The translation trap. The target narrative calls 88 percent of all library users satisfied, even though response bias and a mean-to-percent conversion are not justified. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Report the scale, response count, invitation base and voluntary design; a percentage must reflect the original method, not a simple multiplication. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. The translation can convey positive feedback while remaining honest about who responded. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
12. Digitisation output and online discoverability
Source situation. A heritage collection reports 850 digitised objects, 1,900 image files and 63,000 item page views.
The translation trap. A summary compresses all three into 65,750 digitised records, adding incomparable production and traffic counts. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Separate physical or intellectual objects, derivative files and user interactions; clarify whether an object can generate several scans. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. Archivists and visitors need distinct evidence of preservation output and discovery, not a larger but meaningless composite. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
13. Staffed desk service compared with self-service hours
Source situation. A branch advertises 46 building-open hours, 28 staffed assistance hours and 70 online self-service hours each week.
The translation trap. The translation presents 144 hours of service without noting the overlapping intervals and channels. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Preserve channel-specific service windows and explain where hours overlap; describe 24-hour digital access only if the actual system is available continuously. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. Readers can plan a visit or request help without assuming every service is available whenever the doors open. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
14. Year-to-year series with missing month
Source situation. The annual dashboard omits February after a network outage and labels the reported total provisional, pending reconciliation.
The translation trap. In the target language the footnote disappears and the provisional sum is presented as a fully audited annual result. The problem is consequential because readers are entitled to assume the translated narrative and the reported table describe the same observation, not two subtly different constructs.
Safe correction. Restore the missing-data label and any allowed estimation method; flag the revision for analytics owners before public release. Keep the original units and reporting notes visible until the data owner has confirmed the revised wording. It may be necessary to change a headline or graphic rather than merely swap one word in the body copy.
Why this matters. A credible translated dashboard keeps uncertainty visible instead of making a table look cleaner at the cost of truth. In a finished multilingual report, the reviewer should be able to reconstruct the source claim, name its limitations and explain why the chosen expression does not increase certainty or promise an outcome that the original data did not establish.
A practical twelve-stage translation and verification workflow
1. Inventory the source package
Collect the source narrative, tables, charts, indicator dictionary, reporting calendar and amendments. Do not translate an isolated slide if its denominator is defined in another file. Ask whether the numbers are approved for publication and whether the report covers one branch or a whole network.
2. Assign data ownership
Identify which team can answer questions about circulation, buildings, digital analytics, finance and surveys. A translator should never be expected to invent a methodological answer that only the person collecting the data can provide. Escalation channels save more time than improvised terminology.
3. Create an indicator ledger
For every measure record the original label, technical definition, source field, period, count unit, exclusions, confidence note and approved translation. Preserve identifiers that allow the figure to be traced to the data warehouse or spreadsheet.
4. Mark incompatible series
Check where measurement methods changed: entry sensors, vendor platforms, catalogue migrations, membership policy, data retention, branch opening times or survey methods. A visual trend should not imply uninterrupted comparability when the source expressly warns against it.
5. Protect numbers and arithmetic
Independently recalculate illustrative ratios and totals from the source values. Check decimal separators, percentage points, thousands separators, currency, rounding and whether the report labels a ratio as a fraction or percentage. Do not allow localisation tools to reinterpret a formula.
6. Choose clear target-language labels
Use the local library profession’s accepted term when its definition matches. When it does not match, write a transparent qualified translation rather than a false friend. A short explicit explanation usually costs less space than a reader misunderstanding the indicator.
7. Separate measured activity from outcomes
Visits, downloads and loans are activities or use events; reading achievement, improved access and community trust require other evidence. Never introduce a causal claim simply because an English adjective such as successful, effective or improved appears nearby.
8. Retain exclusions and footnotes
Display missing-data flags, platform exclusions, self-service limitations, mixed counting methods and estimated values. A footnote can contain the most important condition of a comparison. Verify its placement after typesetting, especially in dashboards.
9. Review charts and accessibility
Check that labels, legends, axis units, alternative text and data tables describe the same series. Long target-language words can wrap into neighbouring labels or disappear in mobile views; an accessible report must carry meaning even without relying on colour.
10. Conduct analyst and language review
Ask the analyst to validate the measurement and a bilingual reviewer to validate the sentence. They are complementary checks. The analyst need not be a language specialist, and the translator should not be allowed to certify an undocumented statistic.
11. Test a reader’s interpretation
Give a short section to a reader unfamiliar with the source. Ask what exactly was counted, which period it covers and what conclusion follows. If the reader claims more than the data permit, revise the translation or add an explanatory note.
12. Archive decisions for the next annual cycle
Store approved terms, source versions, resolved questions and reasons for changes. Next year’s report may use a different vendor or boundary. A record of why an indicator was translated a particular way prevents a helpful correction from becoming a new source of inconsistency.
Questions readers and translators frequently ask
Can I translate physical visits as visitors?
Only when the source measure actually counts people. A sensor may register multiple entrances by one person, accompanying people, or movement related to staff. The safer translation names the event unit that was measured, explains the equipment or counting method and reserves unique visitors for deduplicated counts. If the original report itself uses an ambiguous label, ask its data owner rather than guessing.
Is an electronic download the same as a loan?
Not necessarily. Some library systems register a digital loan, some count download events, and others report content requests or platform sessions. A single person can generate several events. Translate the unit defined by the source system, and keep the vendor or reporting-method note visible. This prevents a digital growth chart from being mistaken for a count of individual borrowers.
How should the word turnover be translated?
Begin by writing down the numerator and denominator before choosing a word. In a collection-use context it can refer to circulation relative to stock, but in general business contexts it often means revenue or staff replacement. A target term that is clear inside a financial report may therefore be wrong inside a library performance dashboard. Add a short definition when the translated audience is unfamiliar with the technical meaning.
Can cost per use prove that a service is inefficient?
A unit cost is one input, not a full verdict. Public libraries have inclusion, education, preservation and access objectives that may not appear in a denominator. Extremely low use of a specialist resource can coexist with high value for a small eligible population. Report scope, fixed costs, method, currency and period before discussing interpretation; do not translate a calculated ratio into a policy recommendation.
How do I translate an indicator when the institution changes its calculation?
Treat the revision as a measurement event. Keep the before-and-after formulas, effective dates, affected branches and any restatement decisions. If the source claims comparison remains possible, identify the evidence supporting that claim. Do not silently connect the two series with a straight trend line, even when the new number looks encouraging. Readers need to know when the measuring ruler changed.
Can translated tables omit footnotes to save space?
Footnotes often contain the reason a figure is not directly comparable. They may explain exclusions, small samples, missing data, a shifted reporting period or changed platform definitions. Removing them can reverse the claim even though every numeric cell remains accurate. Redesign the table if needed, using visible notes or expandable explanations, but retain the qualifying information and review it after layout.
Should the translated report use a local currency?
Translate currency labels and formatting, but never replace the original amount with a conversion without a disclosed exchange rate and date. International users may need both the source currency and an indicative converted value. The original audited number should remain traceable, including whether amounts are before or after tax and which financial year they cover.
Are league tables an acceptable use of these indicators?
Only after the comparison method has been examined. Libraries differ in missions, community sizes, digital licences, opening hours, classification systems and costs. Rankings from incompatible source data create an illusion of precision. A careful translator preserves benchmark group, caveats and denominators rather than promoting a raw figure as universally superior.
Who should approve the final translation?
Ideally a language specialist who understands the reporting conventions, the library’s data owner and someone responsible for publication. Their jobs differ. The translator checks meaning and natural language; the analyst confirms definitions and arithmetic; the publisher checks charts, links, footnotes and accessibility. Record decisions on ambiguous indicators so a later annual report does not recreate an old error.
Does this guide reproduce ISO 11620 requirements?
No. ISO 11620:2023 has a published catalogue description and a controlled standard text. This article explains practical translation risks through original educational examples; it does not certify conformity or recreate protected indicator definitions. Use the licensed standard, current institutional indicator manual and applicable reporting guidance for formal implementation or audits.
Where this specialist guide fits in the wider learning system
The specific skill here is translating library indicator evidence, not learning translation from the beginning. For the broad principles of carrying meaning across languages, use Master Art of Translation. To build the precise vocabulary needed to distinguish visits, users, sessions and denominators, continue with the protected Vocabulary Learning Hub. For the grammar, clause scope, comparisons and qualifiers that control the force of a report, consult How English Works. These are connected learning resources, not competing replacement hubs.
Ultimately the test of a library-statistics translation is simple: when a reader asks what happened, who was counted, over what period, using what evidence and with what limitations, can the translated report answer without changing the original? If the answer is yes, the translation has performed an educational service as important as the report itself. Its figures remain useful because their meanings remain intact.