PlantScan
A Mobile Application
This is a hypothetical project based on real needs discovered when I worked for South Osborne Community Cooperative (SSOCC). SSOCC is a community driven organization that converts barren land into fruitful garden spaces. Aside from job positions such as garden stewards and board members, the gardens are maintained largely with the help of volunteers.
Problem
Volunteers find it difficult to correctly identify weeds from non-weeds in the gardens. SSOCC believes in the principles of permaculture, and therefore never uses herbicides or pesticides. Hand weeding is often the primary way the gardens are maintained. It's a task that most members, volunteers and even visitors partake in. Being able to correctly identify weeds from non-weeds is vital in maximizing volunteer efforts.
Additionally, SSOCC wants to increase their number of volunteers and members, as well as their participation in community events. These events and workshops are a source of revenue and awareness for the group.
Solution
Design a mobile application that will allow users to accurately identify plants as well as promote community engagement through an events page and newsfeed, where users can post plant care questions.
My Role
Research
UX and UI
Wireframing
Goals
1. An easy to use app that quickly and accurately identifies plants.
2. Increased user participation in SSOCC and other community events.
Research
Challenges
1. Creating the simplest and fastest plant identification process.
During my competitive analyses I looked at the most popular plant identification apps in the App Store and Google Play. I read through the customer reviews and found the main pain points were:
1. Inaccurate plant identification
2. Poor performance i.e. slow, app crashes
3. Not user-friendly i.e. steep learning curve, error messages
My research revealed that a lot of identification problems were due to the quality of the pictures users took. My early iterations tried to address this by providing clear instructions on how to submit images.
Early Iterations
However, even if these steps could increase the accuracy of the results, it wasn't very fast or user friendly. It shouldn't be the user's responsibility to get an accurate scan, the burden should fall on the technology itself. This was less of a "Eureka!" moment and more of a "Facepalm" realization.
The solution was to use the most advanced AI technology to eliminate steps. Similar to Google Lens (a particularly loved photo detection app) the scanning process starts almost automatically, needing minimal input from the user.
FINAL SCAN PROCESS
2. Ensuring the scan function generates accurate results.
Nothing frustrates a user more than an app that doesn’t deliver on its promises.
Due to technological limitations, it's impossible to guarantee 100% accuracy for plant identification. In order to be transparent about this to users, I created a tiered system that would significantly limit inaccurate matches.
There are three scan categories:
1. 0-59% correlation of the scan to a plant in the database: this would result in a no match
2. 60-89%: a list of possible matches
3. 90+%: one confirmed match
It's always a good idea to back up your back up. Incase there is a problem with the scan results, the user is able to reach out to customer support within the app itself. The reviews for every single app I researched included frustrations with software glitches, exacerbated by the fact that users felt unheard. The in-app help feature allows customer support to address concerns directly and quickly, before frustration drives the user to write an unflattering review.
3. Create the ultimate user experience even when the scan is unsuccessful.
When I looked at reviews for competitor apps, many users complained that they did not know what to do when the app didn't have a match. I decided to address this problem by utilizing an "ask the community" feed post that would also boost community engagement.
Turning a no-match into community engagement
4. Integrate social features into the app.
As important as it is to meet user needs, it's just as important to meet business objectives. Goal 2 of this project is to "increase user participation in SSOCC and other community events". It's important to note, that my research showed users are also eager for existing apps to expand their social features. It's not surprising that people want to share their pictures, like posts, and follow one another.
The Newsfeed turns PlantScan the app, into PlantScan the community. Users can post questions, information, announcements etc. This allows for virtual interaction between users, SSOCC, and the larger community.
Let's keep the momentum going and turn that virtual interaction into real life revenue generating engagement! Can I get a woohoo!
SSOCC already does a great job putting on amazing, affordable workshops like vermicomposting for beginners, and at home kombucha brewing. The Events section allows them to advertise these and other community events to their target market.
Next Steps
Let's recap. Goal 1: create an easy to use app that quickly and accurately identifies plants. Check! Goal 2: create appropriate channels to increased user participation in SSOCC and other community events. Check! As stated before, this is a hypothetical app, but imagine it's real. In this world PlantScan has developed a healthy following. Users are happy with the apps scanning performance. Most excitingly, they love the social features. Eager to hear about local events, share recipes, and swap seeds. With a solid foundation in place, let's imagine the future possibilities.
1. Private messaging within the app.
2. In-app purchases.
3. Sponsored contests with prizes such as personalized garden manuals with your very own images!
4. A premium version that allows you to connect with industry experts such as botanists, chefs, and environmental scientists.
Not only do these options open up many more sources of revenue for SSOCC, they'll also keep users engaged both inside and outside the app. Thus fulfilling an even bigger goal, igniting curiosity and compassion for the environment.
