Developing the MVP for a precision farming product
DX engagement to define, design and develop a MVP tailored specifically for a client in the innovative Agri-tech industry. Through this collaboration, we produced a user-friendly and impactful solution that meets the unique needs of the agricultural sector.
Background
For a client in the agriculture industry, I embarked on a transformation of their precision crop production service provider which produces and interprets variable rate maps. The product allows growers and advisors to view and analyse precision farming data, making it easy to add, retrieve and alter information and allows accurate and effective modelling to test scenarios without the need to step outside.
“Precision agriculture, also known as precision farming or site-specific crop management, is a farming approach that uses data, technology, and agronomic principles to precisely manage agricultural inputs for optimised efficiency, productivity, and sustainability of agricultural production.”
What we did…
Defined a MVP based on jobs-to-be-done identified from user research
Conducted workshops to co-create an MVP, using techniques such as “How Might We’s” and “Crazy 8s”
Refined the MVP iteratively based on user feedback obtained from testing sessions
Developed a high-fidelity prototype on Figma using an existing design system
The objective
Transforming the product to better reflect the shifting user needs and the continuously evolving agri-tech market landscape is essential for success in this dynamic industry.
The objective of this engagement was to rapidly define, design and develop a MVP our client. The existing product has fallen behind competitor tools in the market, in terms of its features, usability, and design aesthetics. Therefore, our goal is to ultimately make the new product to become the front-runner of precision farming tools. We kicked-off this journey by developing a MVP that is rooted in our users (i.e., farmers) to ensure that we start designing and developing features that are most important to them.
The goal
Design and develop a MVP within 5 months.
The challenge
Design and develop a MVP within 5 months.
The team
UX Designer (myself)
UX Manager
Business Analyst
GIS Specialist
Product Owner
Head of Digital Solutions
The approach
This was delivered in a 4 month timeframe that was broken down into two iterations with the following steps:
-
Week 1 - Understand
We reviewed and digested the personas and user journeys that I developed from conducting user research for the client prior to this engagement.
Week 2-3 - Define
We planned and conducted a workshop to understand why our users (i.e., Farmers) come to our product, and what it helps them achieve using the jobs-to-be-done framework. This helped us refine what features are needed for our MVP.
Week 4 - Design
Low-fidelity designs were created and to quickly mock up features that were agreed on in the define stage. They were refined based on discussions during daily feedback sessions and were ultimately transformed into a high-fidelity prototype on Figma.
Week 5-6 - User Testing
User testing was conducted remotely over Microsoft Teams with 5 different users.
Week 7 - Distillation and playback
Insights from user testing were distilled using affinity diagrams and was played back to the team.
-
Week 8-9 - Redefine
Based on our distilled findings, we developed “how might we” (HMW) questions to redefine the problems we should tackle in our next iteration. We prioritised these HMW questions as a team which we then created solutions for using the crazy 8 method during a co-creation workshop.
Week 10 - Design
The top voted ideas from the workshop were incorporated into the existing high-fidelity prototype.
Week 11-12 - User Testing
User testing was conducted remotely over Microsoft Teams with 5 different users.
Week 13 - Distillation and playback
Insights from user testing were distilled using affinity diagrams and was played back to the team.
-
Week 14 - Final Designs
Based on the feedback obtained from users, changes were made to further refine the designs.
Week 15-16 - Development
Hand-over to developers for development.
Iteration 1
Defining the MVP and obtaining feedback from users
Understand
Personas
We kick-started the engagement by reviewing and digesting a number of personas I had generated for a prior engagement for our client. By conducting 1-hour long interviews with multiple user groups, I had developed different personas and their high-level user journeys around using the existing platform. I generated personas for the following user groups:
Farmers
Data Experts
Agronomists
Area Managers
Environmental Reporting Manager
Compliance Team
Sustainability Team
Business Partners
Our objective was to empower users to change their own field boundaries. We thus needed to design a seamless experience for farmers to engage in this task. Furthermore, internal IT specialists may need to provide support to farmers if they are unable to make changes to their field boundaries on their own. Therefore, these two user groups are considered throughout our design process.
User Journeys & Flows
We developed 2 versions of user journeys and flows: As-Is and To-Be. Our intention behind creating As-Is and To-Be user journeys was to highlight the needs and pain points our users have when they currently interact with the existing product, and how their journeys could be improved with the changes we are suggesting (i.e., allowing users to change their own field boundaries). Our aim with creating the As-Is flow was to understand the existing IA of the existing product and how users currently engage in editing and adding field information. With the To-Be flow, we demonstrated how we could restructure the IA to enable users to edit and add their own field boundaries and also improve the overall flow in terms of intuitiveness.
2. Define
“Why are farmers coming to Mapping?”
3. Design
To translate our ideas into designs as rapidly as possible, I created low-fidelity wireframes using Miro. During the course of a week, we had an hour long session everyday where I would present the designs. Any changes made and the team would give me feedback. We continued this until all team members were happy with the designs. We narrowed down our MVP even further into two main features:
Creating maps (using existing templates such as seed, nitrogen, and also blank Maps)
Viewing and editing existing maps
Designs were then turned into a high fidelity prototype on Figma. An existing design system was used to create these designs. The prototype was made interactive for user testing.
We mapped the actions, needs and pains, digital touch points (i.e., the existing product and other digital products) and the empathy levels of the users for each journey step.
The key takeaways of this user journey map is:
Farmers find the product UI confusing, making it hard to access farm and field info.
Farmers wish to edit field boundaries due to the lengthy processing time by IT specialists.
To understand how farmers currently find and edit farm/field information on the existing product (e.g., changing crop types grown on fields, adding polygons/lines to measure distances and areas on their fields), we developed a As-Is User Flow. We mapped the user actions in blue circles and the reaction from the product in green diamonds. It is evident in the diagram that users must follow a long sequence of interactions with buttons and filters to complete tasks/obtain key pieces of information.
We developed two versions of the To-Be journey. The user journey to the left is the To-Be journey for farmers when they change their field boundaries themselves. As some farmers will still require some assistance from the IT specialists, we developed the user journey to the right to account for this. For the To-Be user journeys, we highlighted user actions, their end results, digital touch points and the users’ emotions during each journey step. Both To-Be user journeys pinpoint how both user groups feel more positive when farmers can find relevant information more efficiently (i.e., through an improved IA) and are given the freedom to change field boundaries on their own/given support to make boundary changes.
Our To-Be, or proposed, user flow demonstrates the user actions and reactions from the product, as well as back-end validation that must occur to execute these features. We have developed 4 main flows to enable users to:
Add field boundaries then draw boundaries
Import field boundaries (from files saved on their computer or third party software)
Draw boundaries then add field properties
Editing boundaries
Other flows (e.g., drawing and measuring tools to be used for fields)
Flow 1 and 3 are very similar, they are simply in reversed order to give users the flexibility to add a new field in different manners. We also introduced a number of elements to improve the usability of the website including:
A sub-navigation to help users get to lower-level categories in the IA without countless clicks
Collapsable pop-ups to give users the flexibility to view field information in field or map form (based on their preference).
After refreshing our memory on our primary and secondary persona’s key journeys, needs and pain points, we began defining our MVP. To define our MVP, I planned a workshop for my team using the jobs-to-be-done framework. As a team, we considered the questions:
Interviews transcripts were distilled using affinity diagrams to identify themes that emerged from the sessions. The main themes and sub themes are demonstrated in the picture to the left. Once I played back these findings to the team, the I used the insights obtained from user testing to initiate the second iteration of our design process.
“What does it help them achieve?”
We compiled farmers' goals with our product, grouped them by theme, and prioritized through voting. The top themes selected were:
Viewing Boundaries (map styles, field data, yield data):
Enable users to view field, soil, and crop data on a map.
Limited non-production data (editing, adding, viewing):
Allow users to track non-production data related to farm management (e.g., environmental schemes).
4. User Testing
User testing was conducted with 5 users (i.e., farmers), with each session lasting around an hour. We experienced a variety of users, some who were very eager to tell us their thoughts, and others who did not have much feedback / criticism to give and were happy with redesigns.
I created documentation on Miro highlighting specific feedback we received on design components on the UI.
5. Destillation & Playback
Iteration 2
Refining the MVP based on user feedback
6. Redefine
We embarked on the second iteration of designs by, firstly, redefining the problem space. We developed “how might we” (HMW) questions for each of the themes that emerged from our user testing to reframe our insights into opportunity areas and innovate on problems. These HMW questions were prioritised through voting, and the top three HMW questions were ideated on during a co-creation workshop. In this workshop, I led “Crazy 8s” for each prioritised HMW questions, which resulted in an ample amount of ideas. We presented our ideas to each other, gave each other feedback, and voted on our favourites and / or merged ideas to flesh them out further. We formed user flows for our top ideas which were incorporated in my redesigns.
7. Design
The features and user flows developed during the co-creation workshop were designed on Figma. As the new features and flows did not diverge significantly from the original designs, I decided to make the changes directly on the high-fidelity prototype as we did not require as many feedback session as in the previous iteration. This was due to how our MVP was starting to take a more solid form than before.
8. User Testing
User testing was then conducted again with 5 different users. In this round, we introduced our secondary persona, area managers, in our sessions as well. Our aim was to understand whether there would be a difference in how our designs are received by our primary and secondary personas.
9. Distillation & Playback
Similar to the previous iteration, interviews transcripts were distilled using affinity diagrams to form themes that emerged from the sessions. Some of the main themes and sub themes are demonstrate to the right. Feedback that we received from both user groups were quite similar, and they were mainly around cosmetic issues on the UI and the desire to do more with the product (which was out of scope). Therefore, it was decided to make these cosmetic changes on the design and then to hand it over for development.
Final Refinements & Handover
Finalising the designs and handover to development
10. Final Designs
Our final designs for the MVP involves two main user flows:
Landing Page
On the landing page, users are able to create a new blank map (i.e., a. map without any template) and also view their maps that they have previously created. This feature enables user to view their fields and different data layers on a map. By allowing users to create blank maps, we are not restricting users to create set maps and they will thus have the freedom to map what they desire.
As our product is a significant shift from the existing one, there are informational tooltips next to each title to help users understand what they can do. Under “My Maps” users can use filters and the search bar to find specific maps efficiently, and can also rename and delete existing maps.
New Blank Map
When users select “New Blank Map” on the landing page, they are directed to a page to input a map name and description.
Once the information is inputted in the fields, users can select data layers, view a legend for the data layers, and measure distances on a blank map using the FABs on the right. To ensure that all users understand the iconography used on the map, we have introduced tooltips for each of the FABs when users hover over them. We have also designed a dropdown to allow users to change the basemap. In the bottom right hand corner, zoom tools have been introduced alongside a scale for reference. Users are able to change the map name and description by clicking on the drawer tab to the left.
My designs enable users to engage in a number of key functionalities in a user-friendly manner without taking away significant screen real-estate from the map.
Key learnings
Use Low-Fidelity Wireframes to Align on Designs Quickly
Using low-fidelity wireframes in the first iteration of designs proved to be very powerful, being an optimal method to achieve a solution quickly. The low-fidelity wireframes I designed conveyed enough meaning to clients and stakeholders, as they allowed us to align on the product’s functionality and layout. This meant that design effort was not wasted and we were able to collaborate on creating a product quickly and efficiently.
The Power of Crazy 8’s in Problem Solving
Crazy 8’s proved to be an effective tool to develop a vast number of solutions to challenges we faced. Introducing this playful technique in our workshop sparked innovation, facilitated brainstorming and allowed us to overcome creative blocks. Although some stakeholders in the team were nervous about this rapid technique, we all eased into it and it ultimately facilitated creativity and collaboration.
Dealing with Difficult Users
Some user testing participants struggled to think aloud and hesitated to critique our designs. We adapted our approach to gather valuable feedback by introducing a practice activity for thinking aloud and reassuring participants that honest feedback was welcomed. I emphasized my role in obtaining their genuine thoughts and that the product is designed for them, making their first reactions crucial. These techniques were successfully implemented in our second round of testing.
