Smartling is our localization platform that is used to translate pages on our website, emails and landing pages in Marketo, and various marketing collateral. The global content team is primarily responsible for our localization initiatives and translations. Follow our progress for fully rolling out Smartling in this epic.
Project Managerneeds to be given access to specific projects to be able to see and modify anything in that project.
Account Ownerscan add
Project Managerscan add additional
Project Managersto the projects they have access to.
RequesterPortal feature must be enabled for your account.
See more details on Smartling's help article regarding user permissions.
We currently support French, German, and Japanese languages in Smartling. We are working on adding support for Russian, Italian, and Spanish at an additional cost per language pair.
Any assets being translated into the supported language pairs will be supported by the localization budget. Any assets outside of this support will need to be allocated from the requester team's budget and include a campaign finance tag in the
Reference Number as well as links to any associated epics or issues in the
Description. Finance requires us to tally the cost and professional services hours at the end of each month for the month so they can approriately allocate each cost to the respective budget. If there is no campaign finance tag associated with your job submission, please use the follow naming convention:
Language Abbreviation - Campaign Name - Date
For information on how Field Marketing uses translations, please head to that page.
Enable email notifications to help stay on top of content changes during the translation workflow process. Notification settings can be customized on a per project basis.
You can reply to issues via email.
Important: The original source asset must be available in the same location from where the translation was requested for the connector to successfully create the translated version in your Marketo instance.
Projects are organized by the integration type:
Smartling categorizes translation jobs by the integration type (Marketo, documents, GDN). This is because each of these integration types parse the strings differently based on the file format (example: HTML vs. PDF). Before you request a translation, you must first start at the integration type and know what type of document format you are translating.
Each job should only be translating to one language pair. For instance, if you have a whitepaper that you would like to translate to French and German, you would create two separate jobs: one for the German version and one for the French version.
Each jobs is parsed differently in Smartling depending on the file type. So if a similar string exists both in a web page (HTML) and a document (PDF), Smartling may not match those strings in the same way.
You can request a translation job either from the main
Jobs tab in the top navigation or from the project level the job will be conducted in based on the file type (Marketo, GDN, Documents).
From top-level navigation
Jobs. This takes you to all translation jobs regardless of what project they reside in.
Request Translationbutton in the top-right.
[Title of asset] - [Campaign tag name] - [Team/individual DRI]
Description, paste the link of the epic or issue related to the translation job.
Save Job and Continue.
All saved jobs must be authorized before they are submitted for translation.
You can upload additional files, screenshots, etc. to provide context to the translators about the job. The more context you provide around a job the better.
In the description of your job, ensure that you are including a link to the respective issue or epic related to your job. For the
Reference Number please include the finance tag for your job so the finance team can allocate the spend accordingly.
Total word count (word count of the source file/language, not the translated file/language) multiplied by the per word rate (rate differs based on language) minus discounted rates from matches on previous translations (repetitions).
Automatically generated based on the word count. Anything under 5,000 words is typically ready in 2 business days.
To download a translated job from the
GitLab Documents project:
Jobstab from the
Filestab from within the job.
Subfolders for languages- file name is not ammended, remains the same as the source file
Languages in file names- adds the locale code at the end of the file name
Subfolders for languages and languages in file names
If any strings are rejected from the translation job, those would appear as issues. If the translators have any difficulties or questions about the source file, these would appear as issues as well. If an issue is raised on a job, all account owners and project managers for that project will be notified. Translators will wait for a day for a response but will continue the job if no response is provided. Once the issue is resolved, please mark the issue as
History tab on a job provides detailed information about the activity that has occured with the job. You also have the option to
Download Translation Activity Report. The same history is available on the individual string level.
A translation memory is a saved record of previously translated content. Each time a translation is saved in Smartling, it is written to a translation memory.
A string that perfectly matches a string in the translation memory, including any tags, placeholders, etc. SmartMatch compares new strings against existing translations in your leverage configuration to automatically apply translations to strings you've translated before.
The process of adapting content from one language to another without losing consistency in tone, intent, and style. Unlike translation, where words or phrases are converted from one language into another, transcreation is based on the conversion of the essence of a message from one language to another, rather than verbatim.
An approach to estimating the cost and effort required to translate a job that accounts for
Translation Memory Leverage and
Repetitions. Calculated by multiplying each word with the corresponding fuzzy match rate.
For example, let’s assume your job has ten source words, and that for words with an 85-94.9% fuzzy match, you will pay 60% of the per-word-rate. If all ten words fall into this fuzzy tier, there are six weighted words in the job. The reason is that 10 x 0.60 = 6.
When Smartling uses an existing
Translation Memory (TM) to match source content with existing translations in the TM, it will often find word matches that are less than 100% identical. These are called "fuzzy matches" and are represented by the percentage to which they match (typical fuzzy match percentages are 50-99%). Translators can see these fuzzy matches in the CAT tool and choose to edit them instead of translating from scratch.
Computer assisted translation tool. This is where all of the translating, editing, and reviewing takes place. See CAT Tool for more information.