Oops!
My portfolio isn’t available on mobile just yet
For the best experience, please visit on a desktop.
As a little thank-you for your patience, here’s a cute dog gif
Mistakes in the design process and how we created the right AI translation solution
My Role
UX Design Lead
Team
CSM Operator
Backend Developer
Main Skills
UX Research, User-Centred Design, Product Discovery, Design Iteration, AI-Driven UX Solutions, Content Strategy, B2B2C Design, Phygital Experience, Multilingual Product Design, Design Validation, Business
The Basics
Getting to know the company
Hotel MSSNGR is an SAS company with a B2B2C focus that produces a white label app within integrated back office for more than 100 five-star hotels and resorts across Europe, with more than 2,200 weekly accesses to each app.
The hybrid technology of the native apps uses Material Design as its design system, while the back office uses Bootstrap as its design system, allowing designers to work on both versions of the product, generating white label solutions that use these components and navigation features.
Some real data from this case study has been omitted in order to respect the data privacy of the company in question.
Inputs
A conversation that defined everything
As the product team already knew, one of the biggest difficulties hotels had when using our product was the creation, updating, and quality control of the content entered into their back office and app.
From a business perspective, content creation was never a direct goal of our product—our goal was to increase the viability of that content.
One day, I was called in for a meeting with the company's CEO. At this meeting, he presented some insights he had gained during his last visit to some customers, and with them, the main one: let's create an AI translation tool to help our customers produce content in multiple languages for their guests.
The idea seemed great, we had space in the backlog and a developer available, so why not start investigating and implementing this solution?
Research
Desk Research and our current scenario
To verify that the idea presented to me was indeed a problem, I consulted our Analytics to understand and visualise the number of languages offered in our app, as well as the amount of content offered per language.
Research
Benchmarking
With the validation, through the numbers, that content production in another language was indeed a problem for our hotels, I thought it would also be interesting to take some time to observe who was offering this solution to hotels, either directly or indirectly
Research
Talking with our users
One of the practices I maintain as a UX designer is to keep a contact database of clients, which is updated periodically so that I can validate ideas and solutions. For this project, I decided to talk to some clients in this database who had problems producing content in other languages.
The main questions I asked during this process were:
• Who produces the hotel's content and how often is it produced?
• How is content produced in other languages and who is responsible for this?
• What is the process for producing new content and how often is it done?
• What is the process for translating or writing content in another language and how often is it done?
Research
Key insights of the research
After talking with the leaders of the selected hotels, I was able to gather some feedback that would be crucial to the development of this project.
Hotels need to create and update content weekly
As hotel activities vary, hotels need to update and maintain this information consistently throughout the week, reflecting what is happening on the property.
The people responsible for the content can only produce it in one language
The people responsible for updating and producing the content are those in charge of the areas where the activities will take place, and in almost all hotels, they only produce content in their native language.
Hotels use free translators to generate content in other languages
As content is only officially translated into one language, the employees responsible for producing content need to translate all content produced using an online translator in order to create content in the languages required.
Hotels want to promote the app in another language, but are unable to do
Much of the feedback received by the hotel concerns the content and language available on the app. The hotels are aware of the need to make the app available in other languages, but they are not yet sure how to do so.
Initial Solution
AI Translation
Our initial idea was to transform the current content creation and editing process into one that could be carried out entirely within our back office, without the need for third-party tools, streamlining the work of hotel employees who also created content.
First Numbers
The turnaround of our solution
As this was a major feature, involving an add-on business model in the current contracts of hotels interested in the solution, we decided to launch it in trial format for a few selected hotels to test and give us feedback before making the feature available to our entire customer base.
We only receive positive feedback from customers
All of our testers who used our solution gave us positive feedback on the quality and importance of the solution in their work process.
“So far I am very happy with your translation tool. It's very straightforward and the quality of the translation is impeccable.”
+ 78k translations in total and +14 new languages translated
As we launched the beta solution with a tracking feature, it was possible to know the number of translations and new languages initiated by the hotels that tested the beta version in the first weeks of launch.
All of this seemed very positive. We felt confident to finally launch the solution in an add-on version of our current contracts. Until the biggest problem arose: no customer wanted to sign up for the newly launched solution.
Research
Why it was going wrong
It seemed unreal that none of our customers were interested in our solution, despite conversations, research, and feedback telling us the complete opposite.
That's when I decided to lead an investigation to understand what had happened and find a way to help us change the current reality. To do this, I decided to create a simple CSD (Certainties, Suppositions and Doubts) matrix with the team about our theories, so that I could validate them through interviews with our users.
Insights
Main findings: our first mistake
With the hypotheses defined, I went to talk to some hotels that had interacted with the marketing campaign for the new solution but had decided not to purchase the new feature. The idea was to understand why they did not want to purchase it.
Customers liked it, but not enough to pay for it
The first insight is that customers liked our solution, but not enough to pay for it, because there were already free tools that did the same thing — a mistake in the way we designed the business solution.
They wanted to make the app available in other languages, but did not want to translate all existing content individually
Hotels usually already have fewer employees than they should, which leaves everyone overworked. They would even like to have the app available in other languages, but they didn't want to have to translate everything and do so individually.
It was only then that I realised how biased we were, and that what customers really wanted was not just to be able to translate content, but to have an easy way to make their applications available in other languages.
New Solution
Complete Solutions
After further user research and some internal alignment, it became clearer to us what our customers really wanted: a way to make their app available in a new language quickly—and without too much work.
In the end, we weren't entirely wrong — users would indeed need to translate new or updated content — but that wasn't what they really needed. They needed something else first: to be able to make their app available in another language.
That's how the official solution finally took shape and managed to change users' initial perception of our translation tool.
Learnings
What I learned by making mistakes and redoing our work
Customers were willing to pay for a translation tool that met their real need to ‘translate content into another language for their guests,’ but that meant more than we had initially realised.
This whole process of research, trial and error, led me to some key learnings, which were also experienced by the Squad team during the production and launch of this feature. Among them are:
The longer you take to make a mistake, the more you invest in a solution that will not pay off
No matter how prepared you are, the design process may lead you down the wrong path, but there are more ways to predict this than the one we use, which could avoid some frustration.
There are no direct culprits; there are opportunities for new paths to change the current reality
When I realised we had made a mistake, we talked about it openly as a team and managed to chart a course that finally led us to greater success.
Although everyone knows this: listen to what the user has to say
Perhaps I was too biased and failed to pay attention to the main detail of the entire design process: the user will not tell you what they need. In fact, they did not say anything, I just failed to see the whole picture.
No matter how senior you are, you are still susceptible to error and bias in a project
I would love to say that my more than 10 years of experience have made me never make mistakes in my projects, but unfortunately that is not true. What changes everything is knowing how to adjust and deal with mistakes. It is never about not making mistakes.
Next Project
Aligning Design and Business to build a new phygital sales operation for financial products
How I conducted a Big Data analysis, with support from the Business team, to create a new operation for our product with a yield of more than R$800,000 per year.
Let's work together?
Let’s build, learn, and create something meaningful together
Oops!
My portfolio isn’t available on mobile just yet
For the best experience, please visit on a desktop.
As a little thank-you for your patience, here’s a cute dog gif