Helping a Startup Company Identifying UX Issues for their AI Landscape Inspiration Tool

Supporting Neighborbrite, a startup company, in finding usability issues and desired features
Overview
Neighborbrite Inc. is a startup that offers an innovative AI-powered landscape inspiration tool. In this project, we conducted moderated remote user testing to assess the usability of the user flow for generating inspiration. We identified the 4 major issues and provided recommendations for each of them.
Outcome
Our client adopted and deployed one of the recommendations (recommendation 1, which I was in charge of) in one week after the final presentation.
Keywords
UX research
User Testing
Client Project
AI
Project Type
Client Project for Neighborbrite Inc.
Duration
November 2023 to December 2023 (5 weeks)
Team
Yuri Minami (designer) + 3 designers
My Role
UX research. All members collaborated. I was mainly in charge of recommendation 1 and 2.
RECOMMENDATION 1
Let users know what the product offers more easily.
RECOMMENDATION 2
Show users what's possible in the AI customization feature.
OUR CLIENT & THEIR PRODUCT

What's Neighborbrite?

Neighborbrite is a startup company that provides an innovative AI-powered landscape inspiration tool. The users can get landscaping design inspiration by uploading their yard images, selecting areas to work on, and preferred style. It also offers a premium customize featurethat allows users to further customize their yards.

Neighborbrite User Flow
PROBLEMS & GOAL

Find usability issues and user needs through user testing. Make recommendations so that Neighborbrite can enhance usability and product value.

Discussing with the CEO of Neighborbrite, we found they had the following questions.

Are there any usability issues in generating inspiration?
How are the premium features perceived?
What kind of features do users want?
PROJECT CONSTRAINTS

Tight schedule during the holiday season.

The major difficulty of this project was time constraints. We only had 5 weeks, including Thanksgiving week. We overcome this difficulty by managing the timeline and tasks within our group and client.

TESTING OVERVIEW

We tested with the client's target user group.

Method

Moderated remote user testing

Testing Scope

Mobile website
Generating inspiration
Guidance on the premium features for free users
Customize feature as a premium user

We focused on the mobile experience because the majority of the customers access from mobile phones as indicated by the client.

Participants

Homeowner with a backyard and interested in landscaping

or

Professional Landscapers

We set the participant recruitment criteria based on the actual user group and recruited 8 participants from personal network.

SCENARIO & TASKS

Clear, natural scenario & tasks without description of actual steps.

We prepared the scenario and tasks based on our client's goal and our scopes. We made them clear but not too descriptive.

Scenario
Imagine that you have recently moved into a new home with a decent-sized yard. You are excited about the prospect of landscaping this space to make it more enjoyable and aesthetically pleasing. However, you are not sure where to start and are looking for inspiration and ideas.

Tasks
In each task, we ask users to do the following. Full information on the task is available here
1. Upload image
2. Select style & generate landscape inspiration images
3. Navigate to customize feature as a free user
4. Navigate to plant info
5. Use the customize Feature as a premium user

TESTING FLOW

Make testing seamless with pre-testing instruction & preparation

Here is the main flow of the test. We asked users to download Zoom and a sample yard image that we used for testing in advance to make the setup process seamless.

Testing Flow
TESTING MATERIALS

Make testing seamless by creating moderator's script and practice using it.

We also prepared the testing materials. I was in charge of creating the moderator’s script. In the script, I made sure to include how to answer the possible technical questions because we connect to our participants remotely. We also practiced through the pilot testing.

TESTING FINDINGS

Customize feature is difficult to use. Users need more info about the product.

We created an affinity diagram and calculated the success rates of tasks to analyze the testing result. As a result of analysis, we got the following findings. We made 4 recommendations to address the major issues.

100%
Success Rate:
Generate Inspiration
The main user flow of generating inspiration images is easy for users, with a 100% success rate.
0%
Success Rate:
Customize Inspiration
Users had difficulty in using the customize feature. Success here is defined as being able to generate customized images as participants input.
Confused by being asked to signup before understanding the product
It was difficult for users to find demo video that describes what the product offers.
Difficult to understand how to use customize feature
Users had no idea what type of elements are available in the AI customize feature.
Likely to skip checking customize page
The entry point of customize feature was difficult to find.
As a free user, difficult to understand benefits of premium feature
The explanation of benefits of premium feature was confusing and led to misunderstanding.
RECOMMENDATION1

Let users know what the product offers.

Problem
Some users get confused by being asked to log in before knowing what the website offers.

5 out of 8 participants wished they wanted to know what the website offers before signing up/logging in. The demo video that appeared on the home page helps to understand the product but users need to scroll to access it, therefore it is difficult to find.

Recommendation
Make the demo video more discoverable and delay the login process.

We recommended making the demo video more discoverable by embedding a “watch demo” button that automatically scrolls down to the video. Also, we recommended shifting the login process later in the flow so that users can know what the service offers and feel more comfortable about signing up/logging in.

RECOMMENDATION2

Give users clues on possible customization to get started.

Problem
Users got confused by AI customization feature because they have no idea what kind of request to make.

The customize feature is an AI feature where users can type any keywords they want to include/exclude in their yard inspiration and generate personalized yeard images. Free-text input boxes give users significant flexibility, but some users had difficulty in understanding what to type in.

Recommendation
Show example of the major customization elements in addition to the free text fields.

In this solution, we aimed to provide constraints of operation to guide users. While it is impossible to list all the options available for customization, we thought filters may give users examples and help them understand how to use the feature. During the client presentation, we also provided the alternative recommendation to use the autocomplete field.

RECOMMENDATION3

Encourage free users to tap "customize" button by improving UI design.

Problem
Iconography and color scheme discouraged free users to tap the “customize” button to explore detail of the feature.

Free users can check the details of what the premium customize feature offers by clicking a customize button on the style selection page. However, the “crown” icon discouraged them from tapping it. Also, the customize button is next to the dark green re-imagine button so users tend to skip checking the detail page of the customize feature.

Recommendation
Remove the crown icon and promote the customize button in the hierarchy.

In order to solve the issue above, we recommended removing the crown icon and promoting the customize button in the hierarchy by changing the color of the re-imagine button.

RECOMMENDATION4

Make the customize feature description for free users more informative.

Problem
Users misunderstood the functionality of customize feature, which made them reluctant to sign up for it.

The “Pro feature ” page explains the overview of the customized feature for free users. We discovered the participants misunderstood that only four elements are accessible to customize due to its page design. In addition, some of them stated the fixed purchase button made it clear that this feature required a payment, and discouraged further exploration.

Recommendation
Let users know that additional elements are available to customize so that they can understand the benefits of the customization plan.

We recommended adding a hint to show there are more elements available to customize. Also, we recommended the buy button be unfixed at the bottom because we found our users stopped browsing the page when the payment button was always displayed and emphasized by being fixed at the bottom.

IMPACTS

Our client was satisfied with our reports and deployed one of the recommendations within a week.

Finally, we made a presentation with the client to share our results and recommendations. Also, we had a Q&A session and answered the client’s questions.

“This is extremely useful. This is exactly what I was hoping to see.”

- Luis, the CEO of Neighborbrite

The client expressed satisfaction with our presentation and actively considered integrating our recommendations into their roadmap. Notably, the recommendation 1 (adding a watch demo button) was deployed within a week after the presentation.

REFLECTION & NEXT STEPS

Next time, I want to discuss and explore the technical aspect of the design more.

Include more buffer time for testing sessions
Discuss technical feasibility with stakeholders

This project was my first client project for the UX design realm and was a great opportunity. Through the experience of moderating testing sessions, I learned that preparation and practice are key for seamless sessions. Also, the experience of helping clients motivated me to further exploration of the UX field.

Our testing sessions took more time than we estimated because setting up mobile phones in remote sessions was complicated, and it was difficult for moderators to assist participants compared to in-person ones. To manage time better, I would like to include more buffer time for the session and narrow down testing scopes in the future.

Also, I recognized the importance of considering technical feasibility, especially for features with advanced technologies like AI. In future projects, I would like to have more discussions with stakeholders about technical feasibility and also stay updated on the latest technological trends.