Role
Researcher, Product Designer
Project Type
Usability testing to reiterate design for farm management application
What is Landvisor Brazil?
By Corteva Agriscience, Landvisor Brazil is a high-tech solution for pasture ranchers to control and manage the undesirable weeds by capturing satellite imagery, and using AI to identify weeds.
The Challenge
Test the usability, find errors, validate if the solution actually help solve their problems. Then reiterate the design to solve the found insights before the premier launch.
The Users
The product involves touch-points of 2 key users. A rancher who owns the land and cattle, and a promoter who serves the rancher by helping them make decisions and directly provide crop protection services.
Test the usability before launch
Usability Test
Planning
5 Sessions with a mix of promoters and ranchers.
1 - 1.5 Hour Sessions
Conducted over Zoom
As a team we planned and prepared for usability test sessions. We tested with the actual system which allowed us to surface any bugs early on.
All interviews were conducted in Portuguese by our Product Marketing Manager, the only Portuguese-speaking person in the team at that time. Regardless, the design team joined the session to observe reactions and support the facilitator with technical issues.
Then we had a team debriefs shortly after the sessions where we took notes of the test results.
Recording and Analyzing
We found a variety of insights, imperfections, and bugs during the sessions. They were documented in a spreadsheet, coded by error type, with detailed notes.
Using Miro and Google sheets, we created affinity mapping to group the similarities to find themes.
A Key Insight
Searching for their pasture was a huge challenge for the participants….
Why?
Participants cannot locate the CTA easily to create a new pasture.
They did not recognize the search field to search for the location.
Code bugs led one user to land in Vietnam instead of his pasture in Brazil.
In fact, one user took 10 full minutes to search for his pasture. Even when we said we can move on to the next steps, he was very strong willed on finding his pasture. He really wanted to see his pasture land!
Design
Understanding the Whys
Using the Sakichi Toyoda’s method of 5 why’s, I dived into understanding why users are having problems locating their pasture on the map. Found out we have 2 core problems: the techninal issue, which I worked with the dev team to solve, and lack of empathy in design. The tested design did not provide a good order or flow for users to complete their tasks.
Design and Iterate
Multiple iterations were designed to explore possible solutions. I focused on making task of searching pasture was obvious, has some context to help (search by GPS coordinates & location name, doesn’t shatter our design system.
The Prototype
After facilitating rounds of feedback and testing internally with the design team, and the product team, we got a winner !
Why was this the Winner?
Scanning the page, user will immediately see the “Search for pasture” button. It no longer competes with the primary CTA. It triggers the user that this should be the first step to creating a new pasture.
Since released, we have not heard any negative feedback nor issues regarding pasture locating. 😎
The modal provides context on how to search. Users can either name of a familiar location or GPS coordination.