2020
Mid Product Designer at Deltatre
MTRIBES is a SaaS platform designed for clients with existing sports and entertainment services, providing real-time audience experience control, targeting and insights. Being able to export user data to Amazon S3 provides users with the ability to gain deeper insight in their other data processing tools, or simply store it beyond the 30 day limit.
Users were unable to gain further insight into their users’ interactions due to the lack of data export functionality in MTRIBES.
Two main use cases were defined based on feedback from existing and prospective clients:
Export data for long-term storage, as MTRIBES only stores data for 30 days per GDPR regulations.
Export data regularly for detailed analysis in external analytics platforms.
Build a clear and developer-friendly data export feature that allows users to easily setup and manage exports of MTRIBES event data for use in external analytics platforms.
I was responsible for owning the design research, approach and testing. I collaborated with the copywriting and engineering team to help finalise the approach.
An essential feature for users who want to get the most out of their data, mtribes supports exporting data hourly to provide users' third-party integrations with a constant stream of data. One-off exports allow users to backup or analyse historical data for a variety of purposes.
Integrating services can be unnecessarily complex; the bucket settings page breaks down each step necessary for users to connect mtribes to an Amazon S3 bucket of their choice. Users can verify their bucket connection with a single click.
Validating a bucket connection has six unique states that needed to be covered to give the user clear feedback to minimise what can be a complicated process. Good copywriting and understanding of the technicalities of exporting data were essential in providing the right feedback to the user.
Create a one-off export by simply selecting a date range and S3 bucket. Recurring exports can be managed by pausing, adjusting frequency, or changing the S3 bucket.
To understand the way export data features work and to document the commonalities across features on existing platforms, I conducted a competitor analysis of eight products with varying specialisation in analytics:
Clevertap: Analytics and personalisation focus
Funnel: Marketing analytics
Optimizely: Experimentation focus
Segment: Data management focus
Mixpanel: Analytics focus
Adobe Target: Personalisation focus
Amplitude: Analytics focus
Heap: Analytics focus
Before UI, I focused on understanding the technical details of exporting data and laid out all necessary minor interactions in user flows; in particular the setup flows. Validating iterations of these flows with the engineering team ensured the necessary details were correct, providing the foundation for prototyping.
I went through multiple iterations of the user flow to confidently define the required interactions before UI ideation.
Working in Axure, I prototyped lofi concepts to get a feel for the three setup flows:
Amazon S3 bucket setup and linking to mtribes
One-off export setup and monitoring
Recurring export setup, management and monitoring
Despite being a straightforward user goal, the technical details and functionality in these flows needed to be presented in an equally straightforward way. The challenges I faced were:
Displaying necessary information without overwhelming the user.
Keeping the user in context despite the need for nested settings and external S3 bucket setup.
Grouping the functionality required logically and clearly
I settled on an approach that aimed to keep the essential export information front and center. I minimised the recurring export feature to a summary of the export statuses. Access to the history and settings were shown on a sub-page, as displaying a table of hourly exports on the main page would potentially overwhelm the user. For the one-off export, as it is simpler than the recurring export, it did not require a separate page, but rather could display its history on the main page and be set up via a modal.
Testing these decisions would help to refine the concept further before release.
I tested the prototype with 7 participants with technical experience but varying knowledge of data:
3 Backend Engineers
2 Frontend Engineers
2 Data Engineers
I facilitated five of the sessions, with two observers in each session. I wanted to validate the:
Feature discoverability and clarity.
Usability of connecting an Amazon S3 bucket.
Usability of setting up a “one-off” export.
Usability of setting up a “recurring” export.
Overall it tested relatively well, with 4/7 having no significant problems, 2 partially completing one task, and one failing a single task. We identified several areas to improve before release.
6 out of 7 participants successfully located and connected a bucket within the developer section. From feedback, the clarity of a few aspects could be improved by adjusting and adding copy.
Improvements included adding more details on required bucket access permissions and displaying the connected bucket name directly on the main page. For clearer guidance during setup, I added a stepped process to the top info section and numbered the cards accordingly.
“I’d expect to be able to confirm I’ve used the right path [...] I didn’t notice the button down there” -Backend Engineer
"...have I connected my bucket? It doesn’t look like the recurring export is connected to one" -Backend Engineer
The bucket settings page from the testing prototype, compared to the revised bucket settings and page header following user feedback.
All users successfully set up a one-off export, but improvements were needed on the setup modal. With the intention of giving the user all the info they needed at this point, it ended up increasing the mental workload of what should be a quick task.
We refined the amount of information in the modal; simplifying the file format copy and removing the “Data” section, instead titling the main page “Export Event Data” to clearly set user expectations from the start.
These improvements would also benefit the recurring export functionality.
The one-off export modal from the testing prototype compared to the revised one-off export modal wireframe based on user feedback.
6 out of 7 successfully completed the setup of a recurring export. A few improvements needed regarding the “Manage” CTA and its subsequent "Manage recurring export" page.
Notably, two users overlooked the recurring export log due to the export settings being located above it and consequently affecting its visibility. Secondly, the ordering of information within the settings section affected usability, particularly during setup.
To address these issues and improve consistency, I restructured the settings into a modal similar to the one-off export setup modal. The default CTA state was changed to “Set up”, changing to “Settings” and “View log” when an export is running to clearly separate the functionality. Hierarchy issues with the settings were mitigated by displaying the toggle and status only once set up.
“[the toggle] feels like I’m immediately switching on something I haven’t configured yet” - FE developer
“I didn’t expect to go [to Manage] to see the logs” - BE Developer
The confusing "Manage" CTA and export settings section on the manage recurring export page.
An important revision; new CTA structure and simplified settings modal for the recurring export.
This project demonstrated the importance of good copywriting, showing the right amount of info at the right time as well as understanding the requirements and expectation of the user at precise points in the user flow and reinforcing this by using consistent UI patterns.