While this was our second of five projects total throughout the program, it was our first group project and it lasted about two weeks. Our assignment? To integrate a new feature into an existing app. My group was assigned the app for the online high-end retail business Gilt. We had three key objectives from our “stakeholders” in developing the new feature(s):
- Broaden user base by offering social engagement
- Foster social connections between users and brands
- Increase user engagement in order to drive sales
These are known as the “business needs” or “business goals.” Now it was up to us to see how our users’ needs, goals, behaviors, and pain points would measure up to our business goals, and design ways in which we could satisfy both with an intuitive user experience.
We were also given some “aspects to consider” on our project brief, which helped to guide our research:
- Consider the relationships between brands, users, products, and all combinations (sharing on and off platform, shopping history, browsing habits, seasonal content)
- What are the ways social media users engage in sharing, “liking”, and saving on other platforms?
- What social aspects of shopping in-store can be imitated or enhanced by these new features?
Creating a Scanner Survey
To begin researching and defining our users, we started with a survey on Google Forms that garnered over 60 responses. The survey was intended to gauge people’s online shopping habits, and it included a total of 10 questions. We came up with these questions via techniques called mind mapping (the sticky notes and whiteboard writings below) and empathy map (the infographic below), in which we strategized and categorized what was most important to ask our participants:
We then asked a select number of eligible survey participants if they would be willing to be interviewed more in-depth about their experiences with shopping and online shopping specifically, as well as social media habits and experiences. Below is our Interview Script that acted as a guide during our five user interviews:
Affinity Mapping and Data Synthesis
We conducted a total of five user interviews using this script and synthesized our findings into an Affinity Map, compiling our user responses into categories:
Based on our synthesis and analysis of the data, we extracted the following key learnings which mainly came from the Trust and Reviews categories:
- Users could “see right through” brand’s celebrity endorsements and social media marketing tactics (i.e.: sponsored Instagram posts)
- Users do not trust that the pictures and/or description they see online is an accurate representation of the product they’ll receive in the mail
- Users rely heavily on product reviews found online since they are “from real people” and are honest/truthful
- Users will show friends their prospective purchases before buying “to get their opinions on the items first”
Creating User Personas and Defining the Problem
After learning that users prefer user-generated content over content that comes straight from the brands or sponsors, we developed the following two personas that truly represented our users. Creating these personas and defining our users’ behaviors, needs, goals, and especially pain points played a crucial role in developing our designs.
In looking at our two personas more closely, we discovered that the main commonality between them was a lack of trust in brands.
This helped us to define our ultimate problem statement. We iterated quite a few times on this before finally deciding what was most important (specifically speaking) and how to convey it.
Our ultimate problem statement is as follows:
People are more likely to interact with a brand if they trust it. Both our user personas Alexandra and Susan like to socialize about their purchases, potential purchases, and style thoughts. How can we create a user experience that will instill trust in users, thereby increasing engagement and ultimately driving sales?
This problem statement addresses our users’ chief pain point to trust brands more, while also referencing the business goals/needs to increase user engagement and drive sales.
Now that we had our problem statement, it was time to start ideating solutions and sketch our ideas! Because of the stakeholder requests in the brief, we already knew that we should 1.) create some kind of style profile, 2.) add style recommendations, and 3.) introduce a way for users to engage socially within the app. But most importantly for the business needs, we needed these features to work (together, preferably) to drive sales.
We came up with the following list of potential features to include in the existing Gilt app so we could start sketching them…
…but we knew this was way too many! We also saw opportunities where some features we listed could be combined to work together in the Gilt app. It was time for some *feature prioritization* so we headed to our trusted MoSCoW Method. The MoSCoW Method is a prioritization technique that places items in the following categories: “Must have,” “Should have,” “Could have,” and “Won’t have.” My group and I wrote down our features on individual sticky notes (we ❤ sticky notes!) and organized them according to the MoSCoW Method.
Originally, our MoSCoW matrix looked like this:
We then prioritized further, which you can see in the below final MoSCoW Matrices (one with sticky notes and one hand-written):
After hours spent on feature prioritization, we ultimately landed on the four main features needed in order to earn user trust and generate sales:
- A personalized comprehensive style profile
- Style recommendations that are provided to the user based on their style profile
- “Favorites” feature where users can mark their favorite recommended styles (and we later decided users can create “collections” of favorites; originally, we called them “closets.”)
- An open chat forum
So, how exactly would these features achieve our users’ #1 problem of lack of trust in brands?
The personalized style profile and accompanying recommendations would work together to make users feel known, instilling their trust in Gilt and giving them a reason to engage in the app. And the style profile would have to be comprehensive enough to make the user feel Gilt can accurately recommend items to them based on the information they provide.
Where the style profile and recommendations allow Gilt to personalize the experience for the user, the Favorites feature allows the user to personalize their experience themselves by creating “collections” of their favorited items, to then look back on and potentially purchase.
The open chat forum is intended to further instill trust in Gilt, providing users with a platform to discuss products, styles, trends, etc. with other users, instead of relying solely on the company’s own content. This forum would increase user-to-user engagement, which is an additional way Gilt can earn users’ trust.
Additionally, how do these features satisfy the chief business needs to increase sales and increase user engagement?
We thought the recommended and favorited items needed to link to a purchasable product page, so that users could directly buy items from their “Recommended” list and “Favorites” collections.
We also wanted to find a way to add profitability into the forum, which we called “The Talk,” while still ensuring that our users consistently felt like they could go to Gilt as a trusted style source. Below is a picture of Chriss and myself ideating how we the products mentioned in the forum would to link to purchasable product pages, as well, in order to further drive sales. This stemmed off of a feature we discussed during the MoSCoW Method (can be seen above) called Product Mentions.
Prototyping and Usability Testing
We eventually designed a final paper prototype and tested with two users. Below are some sketches we used to test:
After initial testing with two users, asking them to create their style profile and locate “The Talk” threads, we knew we needed to expand on our ideas a bit to make everything seem seamless to the user. We then created a mid-fidelity prototype, which we tested with eight users, bringing us to a total of 10 user tests overall.
We conducted these tests using our usability script with five task scenarios and associated user goals (below) but first began with asking them to just browse around the app to get a feel for it and tell us their thoughts. They said it was “very intuitive,” “easy to use,” and had a lot of insightful, constructive feedback that we incorporated into the reiterations of our designs, including our final high-fidelity design.
Measuring KPIs and Metrics
- The Time on Task (time it takes to complete the task) went down 25% from first paper iteration to last mid-fidelity iteration.
- Our tasks success rate during our first round of testing was at 90% and increased to a full 100% by our last round or usability testing, bring our error rate to 0%!
- We also asked users throughout and at the end of their test if the interface was intuitive (10/10, and ask them about memorability (9/10) and learnability (8/10), as well.
You can see these new features built into the existing Gilt app via our high-fidelity prototype here:
- We’ll measure the following KPIs:
- Did user retention rates increase?
- Do users purchase their “Favorite” items?
- Are users engaging in “The Talk”?
- Develop a fully functioning app with these features
- Consider additional features for the future, i.e. real person stylist, direct message system among users
- Continue usability testing to measure overall effectiveness