Content architecture
Project: Translation of a suite of transactional correspondence templates for a SaaS ecosystem of products
My role: Correspondence Process Owner
Opportunity: Reduce complexity and time spent by rearchitecting the templates and treating content as data.
Problems:
-
Inefficient file structure - our templates were made up of multiple files, one for each section or zone of the message. These files contained both the template code and the translatable text. There was a set of template files for every language we translated into, meaning we were managing around 550 files for the whole suite.
-
Inconsistent terminology - because the content of each template was managed separately, terminology was inconsistent across our messages.
-
Manual process - when a message was translated, the text was copied and pasted out of each file into a single file that was then sent to translation. When the translated text was returned, the text was copied and pasted back into its corresponding file. This was time consuming and error prone.
Here's what this structure looked like for Canada for just the Invoice template:
Solution:
-
Content as data - we pulled the translatable text out of the templates into separate files called data extensions. The correct content string was called from the data extension by the template code, depending on the region and language of the customer. This reduced the number of files to ~50 across all regions.
-
Shared terminology - we had a data extension for content that was common across all templates, which enabled us to ensure we were using the same terminology for the same thing across the suite.
-
More automated process - we partnered with our Globalization team to create a utility that transformed our data extensions to and from a format that their team could work with. This reduced our process by hours.
Here's what the structure looked like for all regions after we finished, again with Invoice highlighted:
