Lokalise Analytics provides reporting dashboards for monitoring localization usage, task activity, and translation quality across your projects.
With Lokalise Analytics, you can:
monitor processed word usage, OTA activity, and translation memory leverage;
analyze task activity, completion time, workload, and TM coverage at task creation; and
analyze translation quality, review effort, editing activity, and AI Scoring data.
Accessing Lokalise Analytics
To get started with Lokalise Analytics, click the icon in the left-hand menu:
Keep in mind that you'll only see data for the team you've currently selected. To switch teams, click on your avatar in the bottom left corner and choose a different team from the menu.
Choosing product type
Analytics include data from both Lokalise Expert and Lokalise Vantage. Use the Product dropdown to choose which product's data to view:
By default, Expert is selected.
Usage dashboard
The Usage dashboard helps you understand how localization resources are used across your projects.
It includes analytics for:
Processed words — analyze processed word usage and the translation methods used across your projects.
OTA — monitor Over-the-Air traffic, downloads, bundles, and platform distribution.
TM leverage — track how translation memory is used across projects and language pairs and how much translated content comes from TM matches.
For a detailed description of the available metrics, charts, tables, and filters, see Lokalise Analytics: Usage dashboard.
Tasks dashboard
The Tasks dashboard helps you analyze task activity, completion time, workload, and productivity across your localization projects.
It includes analytics for:
Task insights — monitor task activity and overdue tasks, analyze completion time, and compare performance across languages, contributors, projects, and task sizes.
TM Coverage at task creation — analyze how much of your task content could potentially be covered by translation memory before work begins and estimate the expected translation effort.
For a detailed description of the available metrics, charts, tables, and filters, see Lokalise Analytics: Tasks dashboard.
Translation quality dashboard
The Translation quality dashboard helps you analyze review effort, editing activity, and translation quality across your localization projects.
It includes analytics for:
Review and editing activity — analyze post-edit rates, edit distance, and reviewer effort across projects, languages, workflows, translation methods, translators, and reviewers.
Translation quality trends — identify recurring patterns in reviewed content and areas that may require further investigation.
AI Scoring — compare AI-generated quality scores and detected issues with actual reviewer behavior and editing effort.
For a detailed description of the available metrics, reports, and filters, see Lokalise Analytics: Translation quality dashboard.
Special notes
Translation methods
A translation method indicates how a key was first translated. It is determined by the source of the first meaningful translation added to the key.
For example:
View image
View image
In the example above, the first entry for the German language (at the bottom) is an empty translation created when the key was initially added. Empty translations are ignored.
The next change adds actual content to the translation. This is considered the translation method for that key. In this example, the method would be translation memory.
Any later edits do not change the translation method. Since the key was first translated using translation memory, that method remains assigned to the key. Later changes are treated as regular edits.
Lokalise tracks the following translation methods:
Translation memory — translations extracted from memory during imports, bulk actions, or automation. Suggestions from the right-side panel are not included.
Human translation — translations entered manually in the editor. This also includes using suggestions from the right-side panel or ordering professional translations through Lokalise or Gengo.
Standard AI/MT — machine translations applied via bulk actions, automation, or by pressing the Google-translate button for empty values in the editor. Right-side panel suggestions are not included.
Pro AI — translations generated automatically by Pro AI, excluding suggestions from the right-side panel.
API — translations added through Lokalise APIv2.
Offline — translations made offline and uploaded via an XLIFF file.
Other — all other activities, including copying keys between projects, pseudolocalization, find and replace, and restoring translations from history.
Meaningful edits
Some Analytics reports distinguish between edits and meaningful edits. An edit is considered meaningful if it meets all of the following conditions:
It results from a translation method event (excluding Other and Import).
It occurs within 90 days, to maintain data consistency.
It includes at least a minor change that was actually applied to the translation.
The translation remains non-empty after the edit.
The following actions are not considered meaningful edits:
Technical changes (such as find and replace)
Opening a translation without making any changes
Clearing a translation, either manually or in bulk
These actions do not modify the translation content in a way that is considered meaningful for reporting purposes.
Metric differences across Tasks, Analytics, and Statistics
Word-count metrics across Tasks, Analytics, and Statistics may differ because each view represents a different part of the workflow. Some metrics reflect task scope, some reflect processed content, and others reflect translation or review activity.
In addition, these views may use different counting methods. For example, Statistics may attribute translation activity using the source-language word count, while Analytics may reflect words that were actually created or modified during processing. Counts may also be based on source-language words or target-language words, which do not always produce the same totals.
Because of this, these metrics should be interpreted in context rather than compared as exact equivalents.
Reviewed words vs. edit rate
A high number of reviewed words does not imply a high edit rate. Reviewed words show how much content went through review, while edit rate shows how much of that content was actually changed.
This means a user may review a large amount of content, confirm most of it as-is, and edit only a small portion. In that case, reviewed-word totals will be high, while edit rate will remain low.
Testing translations with accurate Analytics data
For accurate analytics, avoid copying both source and target translations when testing. Instead, create a new project, upload only the source texts, and apply translations there.
When testing translations in Lokalise — whether using AI, machine translation (MT), or human translation — it’s important to prepare the project in a way that allows Analytics to track translation activity correctly. This includes metrics such as translation methods, post-editing effort, and overall workflow insights.
Not recommended
Copying both source and target content from one project to another is not recommended for testing or proof-of-concept scenarios. If both values are copied without changes, Analytics may retain the original translation method associated with those translations.
As a result, any new translations applied later may not be tracked as expected. This can lead to incomplete or misleading data when analyzing translation methods, post-editing effort, or workflow performance.
Recommended
To ensure clean and reliable analytics data, the recommended approach is to create a new project, upload only the source texts, and generate translations within that project. This allows Analytics to recognize all translation activity as new and track it accurately.
If both source and target content have already been copied, modifying the source text may help trigger proper tracking. However, for the most reliable results (especially when testing AI workflows or comparing translation approaches) it's best to start with source-only content in a new project.
Known limitations
Data availability
Data updates: Once per day.
Data availability: the current month and the previous 36 months (37 months total).
Usage dashboard: New customers will see the Usage dashboard the day after they create or import at least 10 keys.
Tasks dashboard: New customers will see the Tasks dashboard the day after they create their first task.
Usage calculation rules
Base language requirement:
Usage data includes only translations that originate from the base language.
For example, if a project uses English as the base language, translating German into Italian will not appear in the Usage dashboard.
If the base language does not reflect the imported content, we recommend using Tasks and setting the correct language (for example, German) as the reference language. This ensures the translation workload is calculated correctly.
Branches exclusion:
Branches are excluded from both the Usage and Tasks dashboards.
Translations and tasks created in branches are not counted as usage and do not appear in the Tasks dashboard.
When a branch is merged into the main project, the translation usage from that branch is attributed to the Other translation method.
Filters with short date ranges
Lokalise Analytics allows filtering by date ranges shorter than a full month. However, the feature is optimized for monthly filtering. When selecting shorter periods (for example, a week), the displayed data may be less accurate.