“Set foot on the moon--- but let’s do the baby step first.”
Autotranslation
The Problem
The current content creation process for customer facing content (hotel descriptions, activity descriptions, etc.) is time consuming, labor heavy, expensive, and not optimized for efficient growth.
By Providing a better way for content translation, upload, and download functionality to replace Contentful in Content Team’s current workflow in our System, they can quickly scale customer facing content as Tourlane restart destinations and add new components to offerings.
Forseeable Impact
-
Hugely reduction in content creation cost,
- Allow Content Team to scale more efficiently in a localized manner by generate quality content in volume, build more pages aligned with high travel intent,
- Eventually boosts a goal of 20% opportunity growth.
The Real Challenge, however,
for this project is however not the functionality, as decided to implement by DeepL, instead it’s to find balance between the need from our user, Content Team (they want to have full feature in place) ︎︎︎ the capacity in tech team (but we don’t have enough resource to fulfill every wish).
Design Process
Research
Key Findings
- Still very manual:
At least 2 Excel files are used for collaborations among three departments:
TP <> Content Team <> External Resources/Freelancer
- Request from one party to another is sent or done in bulk; currently the workflow from Content Team perspective is:
- Receiving Requests from Travel Prodct with multiple items from same destination and supplier.
- Creating contents from Original Doc (shared doc., or website)
- Upload .csv file containing multiple items in Bulk (using Contentful)
- Hit “Translate” also in Bulk
- Check/Inspect translations randomly (not all) to make sure the quality is good.
- Edit if needed on wording
- Publish to make the translated content official.
Flow in the happier world:
“Looking into the future, what would be the ideal scenario for the content team?”--- 👩🏻🎨 Designer
By thinking a few steps further, as part of product vision, this flow could help the Content Team get rid of external tool they’re heavily relying on at the moment, as a result, the company save money; for us? we will have better integrated system in one.
Design
First Design Proposal:
“We cannot do everything once, understand.
But at least let’s make a bit more effort in translation side- provide bulk feature.” --- 👩🏻🎨 Designer
This approach serves two possible user journeies:
1. (Create from Scratch) Search --> Multiple Selection & Bulk Translation
2. (Update Content) Editing on each item page --> Single Translation
Iteration 1:
“We only have limited time x 1 developer, let’s cut it down to single item translation, for now.” --- 👩🏻💼 PM
Narrowed scope with Single Journey:
Item Page --> No Selection & Single Translation (Remain flexibility for future when extending to Bulk Translation)
Iteration 2: Even Simpler
“...Well, even simpler, let’s use existing pages without creating more to achieve the the same goal with simply A Button.” --- 🧑🏼💻 Devs, in Team Weekly Discussion
Final Team Decision with single flow:
(Item by Item update) Item Page --> No Selection & Single Translation
- Using existing pages only
-
Remain only two states: Edit and Saved/View
-
Mark translatable fields on the existing page
-
As original planned, still provide timestamp and the last editor/user name on both global and local(description) level.
User Feedback on Final Decision:
👩🏽🦱 :
“This (single translation) can only help much when
1. the request for Content Editing from TP is low (as it is now)
2. We have already quality content in one language to translate directly from.”
🧑🏻🦰 :
“Thank you for iterating on all these and trying to push for the best fitting solution for us; however the simplest solution is still very manual work, because we have to go into each item one by one.”