To improve the experience of furnishing apartments and buying furniture in urban cities for newly graduated college students.
These students often struggle in:
- Furnishing own small urban-city apartments fast and efficient
- Planning the furniture placement and aesthetics
- Running on a limited budget for rent, furniture and other services
1. In the formative research phase, I took part in the in-store observation, task analysis, survey design and semi-structured interviews. I conducted competitive analysis independently. Then I worked with teammates to analyze our data and distilled the user requirements from our findings.
2. I actively contributed in the ideation, sketch and wireframe phases. The information architecture and general user flow were revised and finalized by me. I created the overall visual style of the prototype and designed more than half of the hifi screens in several iterations. Yuyan Duan and I completed another iteration to improve the prototype after the project was over.
3. In terms of evaluation, I mainly contributed to the feedback sessions for wireframes and prototype. I worked as facilitator and notetaker.
Sep. 2017 - Dec. 2017
Huaiwei Sun | Yuyan Duan | David Howard | Dillon Weeks
Pen | Paper | Adobe XD | Balsamiq Mockups | Sketch | Flinto | inVision
1. Understood and empathized better with our target user groups: newly graduated students who had little experience in furnishing and buying furniture.
2. Learnt about the entire process within which users made furnishing plans and bought furniture.
3. Discovered significant issues and design opportunities and selected the problem space we focused on in the following work.
Selection of Methods and Justification:
1. Field Observation - We visited an IKEA store and a Home Depot store in Atlanta. We observed the spatial layout of the store, the shelves for all the furniture or related products and how users interacted with the environment. This method could help us discover some existing pain points or space to improve in brick and mortar stores.
2. Review of User Feedback - We conducted a review online of existing user feedback for services related to furnishing provided by different companies. This method could help us explore this big topic and gather a wide range of data showing the general patterns, needs, preferences and pain points in the furnishing process.
3. Task Analysis - We examined the entire process of furnishing and furniture purchase and broke it into several steps. This could help us have an overall view of this process and zoom in on each step to do the following research and find the design opportunities.
4. Survey - We used online survey to collect data from our target users. This could allow us to gather demographic data and learn more about users’ preferences and habits at different stages in the process of furnishing and furniture purchase at a large scale
5. Semi-structured Interviews - We used semi-structured interviews to learn about users’ past experience of furnishing and furniture purchase. This method could help us gather more in-depth data, espcially those related to people's values and attitudes, which enabled us to better empathize with users.
6. Competitive Analysis - We studied the existing products on the market related to furnishing and furniture purchase. This method could inform us of the existing solutions which meet differnt user needs, such as creating the furnishing plan, seeking furniture, buying furniture, transporting furniture and so on.
During our observation in IKEA and The Home Depot, we observed the entire flow users went through and randomly picked some customers to have quick and dirty talks about their goals, attitudes and experiences. The main takeaways included:
- The Home Depot's business was not focused on our target users. They focused more on contractors, builders, interior desingers and DIY customers who have big houses rather than small apartments.
- Delivery was inconvenient, especially for people who didn't have cars. Delivery services were quite expensive.
- Returns were inevitable now because users wouldn't know how items fit into their homes until they brought them back.
- People enjoy seeing and touching furniture by themselves to be confident in their purchase.
- Though Ikea's self-pickup system is complete, some people still get confused and need help from staff.
After the offline observation, we investigated the user feedback on websites and platforms including ConsumerAffairs (website for consumer reviews), MoveLoot (Second-hand furniture platform) and CORT (Furniture Rental Service for Graduate Students).
We distilled the pain points from user's complaints and the result is listed below.
Having finished the field observation and the review of online feedback, we had a more comprehensive understanding of how users bought furniture to furnish their apartments in different ways.
We abstracted the different detailed processes into one. Then we broke it into detailed steps to zoom in on each of them to further capture user needs and existing pain points.
We used Google Forms to create our survey and distributed it on different colleges' subreddits. In the survey, we included questions collecting demographic data and participants' general opinions and preferences during the furnishing and purchase process.
Out of 86 responses we got from the survey, we observed some general preferences:
- People often collect information and make furnishing plans online.
- Most of users accept second-hand furniture as an option.
- Price, aesthetic, size and material are the 4 most important factors influencing user's choice of furniture.
- More than half of the participants transport furniture using personal vehicles.
We pursued the breadth of data in the above research. Then we turned around and started to get more in-depth data. We conducted 4 semi-structured interviews with our target users to ask about their furnishing and purchase experiences. When they were answering our questions, We kept probing whys behind their behavior based on their answers.
Some important insights we found through the interviews included:
- Measuring the dimensions of apartments was unfamiliar to users and may lead to anxiety and frustration.
- Making furnishing plans, especially creating one with a unified visual style, was hard to users if they hadn't received any design, art or related training before.
- When users started to work, their time is limited for doing research on the furnishing. So they had to finish this process in a limited amount of time to settle themselves down.
Based on users' feedback and some insights we concluded from the former research, we did competitive analysis in different fields related to furnishing and furniture purchase.
We were able to know more about the existing solutions on the market to help ourselves avoid focusing on problems or pain points which had already been well solved. For example, there were too many solutions for solving the problem of transporting furniture, whether innovative or conservative, indicating that it was not very promising for us to make more efforts for this problem.
Aggregating all the data we collected from methods above, we analyzed all the pain points, users' attitudes and habits through the entire process of furnishing and furniture purchase.
After team discussion, we determined that the most promising problem spaces were from the intersection of users' pain points and uses' valued factors. The problem spaces we would work on were: Placement of furniture and Aesthetics of the furnishing plan.
Having extracted important attributes from the demographic data collected from the survey and interviews, we created two personas representing our typical users to help us remain focus during the exploration of the problem spaces we decided on.
We also adpoted user journey map to describe our typical users' experiences of furnishing and furniture purchase, which helped us better empathize with our users. Since the journeys are similar between two kinds of our typical users, we used one map to represent the experiences of both of them.
Finally, in order to have a more specific guideline for the following design phases, we combined all the insights we had and listed user requirements and design criteria in this table.
- First, we would ensure that our designs could meet the user requirements.
- Our designs also had to work in a proper way that users enjoyed, i.e, it needed to meet the design criteria.
- The priority would help us make design decisions when we were in design dilemmas.
Keeping the user requirements in our mind, we had two brainstorming sessions and came up with more than 30 ideas.
After careful consideration of the feasibility and creativity of each idea, all four team members each picked an idea that they thought had the most potential. So four ideas were left to be further developed in the wireframe phase.
IDEA 1 - AI Planner
AI Planner stemmed from people’s needs to effortlessly furnish the apartment with a pre-prepared whole set of furniture. Thus, AI Planner gives the user quick furniture/layout recommendation based on the apartment floor plan that user draws. It allows the user to set their own budget, customize their own furnishing needs, view their apartment floor plan and layout recommendations in both a 2D and 3D views, and get spot-on layout recommendations in no time. Users could place their orders and get all the furniture inside the app. Thus, AI Planner provides quick and effortless room furnishing to users with high level of customization. The key challenges to this idea is to acquire user’s apartment floor plan as well as the ways of presenting the furniture on the app.
IDEA 2 - Furniture Remains
The idea of furniture remains is inspired by the fact that apartment residents usually have high mobility - thus, old residents have to get rid of their furniture from time to time. And what is more convenient than selling old furniture to new residents? Thus, Furniture Remains aims to build a communication bridge between the apartment old lessees and new lessees and make the apartment furniture transferred to new hands while staying exactly where they are. Such design removes the placement burden on the new lessee as it allows them to rent an apartment along with a whole set of fitting-floor-plan furniture, and also removes all delivery and transportation trouble for both parties. This idea eliminates user’s whole troublesome experience of measuring floor plan, going to the store, and purchasing furniture, as the new lessee can check out the furniture along the apartment and purchase of the whole set of furniture can be done instantly. One challenge of this app can be that the app has to collaborate with apartment, and therefore may have little user base at first.
IDEA 3 - Roomba with Measurements
The Roomba solution helps users take the measurements of rooms while doing the normal cleaning and sends them to a smartphone app. Adding a lighting and color sensor also allows the Roomba to construct a basic lighting simulation for the user when creating digital 3D models of their rooms. More importantly, we designed the Roomba to be linked with mobile applications that receive the room measurements directly from it. We think that combining digital measurements with layout recommendations and furniture purchase(Idea 1) could have huge potential. However, the feasibility of this single idea can be slim to us at this time since it's more about the development of Roomba which is out of our problem space.
IDEA 4 - 3D Planning Gamified
3D Planning Gamified allows the user to create a 3D rendering of their rooms in a more gamified context with tutorials to help teach the user furnishing strategies. They may select furnishings available in their area to place in the 3D view of their apartment to get a more accurate view of the floor plan and apartment furnishing status. The highlights of this idea is that it attempts to make the usual strenuous and tedious furnishing experience a game and thus deliver a more entertaining, gamified experience to users. By allowing the user to choose, place and rearrange furniture in their simulated view of apartment, this solution easily solves the furniture placement, aesthetics and apartment pre-planning issues and provides the user with instant purchase. The biggest challenge of this solution is that 3D planning on a mobile platform can be too small for the users to see and operate on details. Our skillsets are also inadequate to support ourselves to create a high-quality 3D gamified design in a limited amount of time.
We presented lofi wireframes of all four ideas to 7 users and asked for their opinions about these ideas at the end of each session. We then conducted a SWOT analysis to compare the four ideas.
Based on the results we got, we choosed AI Planner to continue with in the following development.
Each of us created a version of the wireframes to explore different possbilities of interaction design. Then we combined the 4 versions and developed them into the final wireframes.
Based on the wireframe, Yuyan Duan and I worked together to design the hifi prototype.
In order to have a quick test of the usability of our design, we conducted a feedback session with two users to test our prototype and share their feedback.
Feedback was used to guide the Prototype Iteration One below.
1. Find My Apartment
We Initially had users draw their own floor plan in two ways: free draw (create dots and lines to draw the contour) and fill draw(fill colored blocks to draw the room contour). According to our user feedback, although users learned how to draw rooms in the feedback session, there was a steep learning curve for both users. User also reported not being able to draw a precise floorplan on a small mobile screen. Most of all, both users mentioned that the drawing floor plan process was “long and tedious”.
Thus, we changed the “draw floor plan” to “find my apartment”, which leaves most of the flooplan acquision job to the app and thus ease the process with less time.
2. Upload Your Floor Plan
The initial “Upload Your Floor Plan”, as reported by the users, have the following issues:
- It shows different rooms in different color borders, which was not immediately obvious to the user.
- If the app recognize a room wrong, there is again the issue of difficulty drawing a room from the user’s end (especially the one with irregular shape).
- It asks the user to manually check with doors and windows, and the buttons are too small to click on a mobile screen.
According to the research we did, some researchers could improve the room recognition precision(including windows and doors) up to 95%(results come from their own datasets). Thus, we cut some work done by users and let the app do all the recognition work. We also show the rooms in different colors to be distinguished easily, and allow the user to tap each room to confirm details and change the room names. If the recognition is incorrect, users could send their floor plan picture to us for manual fix.
Users pointed out they were confused about how to rotate the furniture. Also, both users reported feeling stressful of using the app without any instruction.
Thus we simplied the action of rotation and provided a short tutorial at the beginning. We also added instructions above the layout.
Our Recommended Layout feature was designed to package all furniture in recommended plans. There was little room for changes. Both users in the feedback session mentioned that they rarely took the whole furniture package - they always have something in the plan that they like and something they hate.
Thus, we made the “Try another one” feature more obvious, in which the user could replace a certain piece of furniture with one from the list. We also included a new feature that allowed the user to add/drop the furniture they liked or disliked. In this case, the Sea Gull Lighting is added to the layouts and the item list.
After the first iteration, we decided to test the usability in a more formal and systematic way. We recruited 3 newly graduated students to help us conduct the benchmarking test and finished the SUS evaluation.
Their feedback was used for the extra Prototype Iteration Two done by Yuyan Duan and me after the project was over in the class.
1. Add a layout - Sequence
2 of 3 users reported, they thought that if creating floor plan was an necessary process, we should make it the first thing that users saw after logging into the app and guide them through the process.
Thus, we let "Add a layout" step appear immediately after the users logged in to avoid future trouble.
2. Display of Furnishing Plans
Users expressed confusion about how we displayed completed furnishing plans. It was impossible for users to create multiple plans for the same room to compare.
Thus, we rearranged the layout and sorted the completed furnishing plans into groups based on the room plan they belonged to.
3. Generated Plans
In the recommended plan page, 2 users did not realize that the screen was actually a full 3D view that they could swipe left and right to get a whole view of the room. Also, a user reported that the white texts were hard to see.
Thus, we added a dark shadow under the texts to make it salient and more readable. We also included an instruction on the page to remind users to try the 3D View.
4. Style Preference
We initially let users pick up to 10 favorite photos. However, although all users reported liking such feature, 2 out of 3 reported that some people might not have the time or patience to choose up to 10.
Thus, we made the user choose 5-10 photos for the app to work out their style preferences. Once the user hit 5 pictures, a "Generate" button would automatically appear on the top right of the screen. Users could tap it to skip the rest of the small quiz and get the recommendation immediately or finish 10 and get results automatically.
5. Add to Cart/Save Plans
All three of users reported that they didn’t want to go to "Cart" after they added the items to "Cart". They were also wondering how to save the layouts for later comparison.
Thus, we added one option and had two now - "save" and "add to cart". If users added items to cart, the cart icon on the top right would change correspondingly. The user can access cart any time later.
6. 2D Display of Plans
We also improved the 2D layout representation, especially how we visualized windows and doors. Originally, we designed the door and windows to have a red border. But red usually means danger and causes anxiety. Thus, we changed the doors and windows to appear less intrusive but still enough obvious for the users to see.
All the participants in the evaluation test mentioned some extent of confusion of how to use this app.
In the final design, we decided to include a quick tutorial to inform users the basic features and how to use them.
Feel free to play with it and share your feedback with us.