Fitsearch
Launching an e-commerce search engine in 6 months
While many of TrueToForm's solutions are B2B, Fitsearch is TrueToForm's consumer solution. Fitsearch is a search engine that pairs with TrueToForm's mobile scan app to find clothing that fits, by collecting garment design files across apparel merchant sites to return highly accurate fit recommendations for shoppers.
Timeline
From explorations to launch in 6 months while working on multiple projects at the same time.
Impact
Fitsearch has signed 33 affiliate contracts since launching and gained 500+ immediate signups following a feature in the Hustle.
Role
I was one of three designers on the team, contributing to:
Early-stage discovery and user research
Design sprints
User flows
Low fidelity design explorations
Instagram and Tiktok content creation
Rolling usability testing for Fitsearch
Team
CEO & Co-Founder
CTO & Co-Founder
Developer
Lead product designer
Product designer
33
Sign affiliate contracts since going live
Chances of increased traffic, reduced return rates, and customer trust have gained the interest of brands and merchants
500+
Immediate signups after being featured in The Hustle
Interest in Fitsearch landed us a feature in the Hustle: Using AI for the perfect fit
Measures of success for Beta & V1
Validate if shoppers would be willing to scan their avatar
Test that Fitsearch could get accurate garment specs
Test that shoppers can scan and connect their avatars to Fitsearch
As a starting point, I conducted discovery research to learn what specific challenges shoppers struggled with when shopping for clothing that fits. I began my research by combing through dozens of reviews for popular clothing brands on trustpilot.com. I found this method helpful for finding honest and unfiltered feedback from shoppers.
I noticed several themes after reviewing shoppers reviews and conducting white paper research online that places increased burdens on both shoppers and retailers.
Confusing size charts and increased burdens on shoppers
Fit and size continue to be the primary reason for online returns, accounting for 42% of returns. There are no universal size charts used by clothing brands which makes it difficult to shop online. In consequence, some shoppers hesitate to shop from new brands or decide to risk ordering the wrong size.
Bracketing
Some shoppers resort to 'bracketing' which refers to buying multiple sizes of one item at one time. This places a burden on both shoppers to return items that don't fit which can drive traffic away from retailers who don't offer free shipping and returns. Retailers also bear the burden for shipping, inspection, and repair costs.
Lack of diversity and inclusion for different body types
Many shoppers find that many existing size charts do not accommodate for the full range of body types. 67% of American women wear a size 14 or above, and most retailers don't carry these sizes.
In addition to reviewing user painpoints online, I also conducted white paper research on our core target audience - Gen-Z and Millennial shoppers. I researched economic, cultural, and technological events that shaped their motivations and spending habits to make sure I was framing for problems experienced by these groups.
The convenience seeker
Shoppers who prioritize convenience only shop online with websites that offer generous return policies, as they often struggle to find the right clothing size. Otherwise, they tend to prefer thrifting or shopping in person to avoid the hassle of paying for return shipping or dealing with time-consuming, complicated return processes.
Social media motivated
Many Gen-Z shoppers and some Millinneal shoppers rely on social media platforms for inspiration when shopping for clothing. Some shoppers struggle to fit within the sizing charts offered by many brands and thus will turn to social media outlets that appeal to their specific body type such as petite or plus size.
Brand loyal
Many shoppers, especially millennials, may be reluctant to try new brands and tend to stay loyal to brands that they are familiar with. However, bodies are not static and can change over time or due to events like pregnancy, which can pose a new challenge when shopping with brands they frequent.
I collaborated with two of my design colleagues to map out all the user problems and potential blockers identified during our discovery research. Together, we used the these painpoints to guide our how might we questions to reframe these challenges into opportunity areas.
I brainstormed solutions in the group session and afterwards conducted lightning demos to draw more solutions and visual inspiration from a range of web products.
Building trust
How Might Wes
How might we instill trust in our avatar scan process?
How might we increase customer trust and confidence in our fit predictions?
The TrueToForm avatar would be the key to powering fit recommendations but without the willingness of shoppers to scan the search engine would lose it's competitive edge. Would shoppers be willing to go through the effort of downloading the scan app and scanning their body? And what would encourage them to do so?
From my lightning demo research, I found that many websites build trust through social proof by sharing customer testimonials, product reviews, or influencer endorsements.
Educating shoppers
How Might Wes
How might we educate new users on how to scan and connect to the search engine?
For first time visitors of Fitsearch, shoppers would need to understand how to connect their avatar to view their fit predictions. Furthermore, since we would be asking shoppers to use a mobile app outside of Fitsearch we needed to educate users on why they were doing so and how it was tied into their fit predictions.
For lighting demos, I shared snapshots from websites such as Honey, Stich Fix and Short Story, where I found helpful step by step instructions on how to use the services.
Value of scanning an avatar
How Might Wes
How might we communicate the value of making a personalized avatar?
How might we communicate that our fit predictions sometimes may be inaccurate if retailers provide inaccurate garment specs?
For most shoppers, this would be their first time encountering scan app technology. Downloading a new app outside of Fitsearch and learning to scan their body would be a barrier to entry. How could we convince users to scan and create a personalized avatar?
Communicating fit
How Might Wes
How might we communicate how something will fit (loose or tight)?
How might we communicate how something will fit while considering styles of clothing (boyfriend jeans vs. skinny jeans).
What further complicates clothing fit are the nuances in style. For instance, shoppers often search for specific styles of jeans such as boyfriend jeans that tend to have a baggier fit or skinny jeans that have a tighter fit. How would we provide an accurate fit recommendations for the same person for these two styles of jeans?
From conducting lightning demos on existing fit predictor tools, this inspired the fit survey option which provides a way for users to quickly generate a recommendation without taking the effort to generate an avatar.
I developed user flows to determine which features and screens needed to be designed for implementation.
I participated in a design sprint by brainstorming low-fidelity design solutions for Fitsearch. While the lead designer was responsible for the design system and high-fidelity designs, she reviewed my initial concepts, provided feedback, and considered my explorations into the vision for Fitsearch. She encouraged us to explore pie in the sky designs even though we were working towards launching a beta version first. The lead designer asked me to brainstorm for the following flows:
Avatar scan flow
Fit survey flow
Search & browsing flow
Fit survey & scanning an avatar flow
Since shoppers frequently browse on their phones, I started by exploring design solutions from a mobile-first breakpoint. The essential features for the landing page included:
Option to fill out a fit survey
Option to scan an avatar
A way to run a search
Fit survey & scanning an avatar: Recommended design
I considered integrating a dropdown menu into the search bar. Since shoppers are already familiar with search bars and tend to gravitate toward them, adding a dropdown component could encourage them to associate searching with adding an avatar. The advantage of this design is that it saves space on the landing page, leaving more room for instructions on how to use Fitsearch.
Given that this may be the first time someone encounters the term "avatar," I chose to visually convey the concept using an avatar icon. To enhance clarity, the avatar icon is paired with a clock icon during the "in progress" state and a checkmark once the avatar has successfully uploaded to Fitsearch.
Fit survey & scanning an avatar: alternative exploration 1
I considered using a popup solution because it saves space on the landing page. However, the downside is that popups can be a disruptive design choice.
Fit survey & scanning an avatar: alternative exploration 2
I considered an option where the scan avatar and fit survey features would be displayed directly on the landing page. The advantage of this approach is that both components are immediately visible to shoppers. However, this design quickly clutters the landing page and pushes important educational content below the fold, reducing its visibility to users.
Searching & browsing
The must haves for the search and browsing flow included:
Option to alter the search
Image of garment (pulled from the retailer selling the garment)
Fit score - the CEO and CTO indicated that Fitsearch would recommend items with a 98% Fit score
Price of garment
Name of retailer
Searching and browsing: recommended design
The exploration I presented to my design team included all the must have features for beta. With shopping online, I thought it would be important to offer a sort option since shoppers often like to sort by price.
searching and browsing: pie in the sky exploration
Since we were in the early stages of designing, the lead designer on the team encouraged me and a design colleague to explore pie in the sky designs. While I did not recommend this for beta, I shared an exploration that showed a preview of the fit score to the users.
I liked that on some retailer sites, I was able to preview some product details for a garment without leaving the browsing page. I explored a version where a module could give the shopper a preview of their fit evaluation. From my research, many retailers include a like or bookmark feature so that shoppers can save items that they like.
I interviewed users and conducted rolling usability tests on Fitsearch when the beta was released to a select number of users and with subsequent feature updates. The usabilty testing was helpful in capturing usability issues and bugs in real time and gathering qualitative and quantitive feedback from participants.
In our usability tests, I wanted to validate two objectives that we were using to measure success:
Are shoppers willing to scan their avatars?
Are shoppers able to successfully scan and connect their avatar to Fitsearch?
Is the retailer providing accurate garment specs? (TrueToForm staff would test internally)
In the usability studies, participants were free to choose between creating a fit survey and scanning an avatar. Most participants chose to begin with a fit survey because it was quick and easy and so that they could begin searches right away. However, many users expressed curiosity about creating a body scan after encountering the fit evaluation.
For beta, we conducted a first round of usability testing with 14 participants. Of the 14 participants, a majority were not able to connect to their avatar to Fitsearch. Some participants were not able to successfully complete their scans and generate an avatar while others were not able to successfully connect their avatar to Fitsearch.
As part of TrueToForm's small, scrappy startup team, our entire team and I validated garment specs by trying on Fitsearch's recommended garments, measuring them, and comparing the results to retailer-provided specs. I also shared qualitative feedback on garment fit.
Avatar connection issues from usability testing
During usability testing for beta, I noticed that many shoppers lost connection to their scanned avatars which prevents them from shopping with their avatars on Fitsearch and incurs a loss of time and effort. As a team, we recognized we needed to develop a stop gap solution to resolve the connection issues in the event the connection is lost so that there is a way for the shopper to connect an avatar to Fitsearch.
Reasons for failure
User closes the dialog box in Fitsearch
Technical bugs and scan app crashing
User needs to update app permissions (camera) during scan process
User restarts scan
Scan submission stalls
User accidentally hits “create new avatar”
User force kills the app
My coworker and I collaborated to brainstorm a stopgap solution that would ensure shoppers could consistently connect their avatars to Fitsearch. Drawing inspiration from streaming platforms like Netflix and HBO, we explored the use of backup authentication codes as an alternative when QR codes fail.
100%
100% of avatar fit connection were resolved as a result of the new stop gap solution implementations
Being part of this amazing project was an incredible experience. I learned so much from contributing to the development of an end-to-end product. One of the key takeaways was recognizing how valuable and essential it is to engage with users throughout the design process. My skills in conducting interviews and usability testing improved significantly as a result.
After implementing updates based on multiple rounds of usability testing, the product was officially launched to the public by removing the password-protected homepage. Since going live, the project has secured 33 affiliate contracts, which has been exciting and rewarding to witness. Although there is still much work to be done, Fitsearch's next steps are to expand into additional clothing categories and strive for a better product-market fit.